Advances in Agronomy, Volume 82

VOLUME 82 Advisory Board John S. Boyer University of Delaware Paul Bertsch University of Georgia Ronald Phillips U...

Author: Donald L. Sparks

199 downloads 1211 Views 8MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

VOLUME

82

Advisory Board John S. Boyer University of Delaware

Paul Bertsch University of Georgia

Ronald Phillips University of Minnesota

Kate M. Scow University of California, Davis

Larry P. Wilding Texas A&M University

Emeritus Advisory Board Members Kenneth J. Frey Iowa State University

Eugene J. Kamprath North Carolina State University

Martin Alexander Cornell University

Prepared in cooperation with the American Society of Agronomy Monographs Committee Diane E. Stott, Chair Lisa K. Al-Almoodi David D. Baltensperger Warren A. Dick Jerry L. Hatﬁeld John L. Kovar

David M. Kral Jennifer W. MacAdam Matthew J. Morra Gary A. Pederson John E. Recheigl

Diane H. Rickerl Wayne F. Robarge Richard Shibles Jeffrey Volenec Richard E. Zartman

Edited by

Donald L. Sparks Department of Plant and Soil Sciences University of Delaware Newark, Delaware

Amsterdam · Boston · Heidelberg · London · New York · Oxford · Paris · San Diego San Francisco · Singapore · Sydney · Tokyo

ELSEVIER Inc. 360 Park Avenue South New York, NY 10010-1710, USA q 2004 Elsevier Inc. All rights reserved. This work is protected under copyright by Elsevier, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-proﬁt educational classroom use. Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone: (+44) 1865 843830, fax: (+44) 1865 853333, e-mail: [email protected]. You may also complete your request on-line via the Elsevier homepage (http://www.elsevier.com), by selecting ‘Customer Support’ and then ‘Obtaining Permissions’. In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (978) 7508400, fax: (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 207 631 5555; fax: (+44) 207 631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of Elsevier is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier Science & Technology Rights Department, at the phone, fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent veriﬁcation of diagnoses and drug dosages should be made. First edition 2004 Library of Congress Cataloging in Publication Data A catalog record from the Library of Congress has been applied for. British Library Cataloging in Publication Data A catalogue record from the British Library has been applied for. Academic Press An Imprint of Elsevier 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.academicpress.com ISBN:

0 12 000780 0

Typeset by Alden Press, Oxford Printed in Great Britain by MPG Books Ltd, Bodmin, Cornwall 1 The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). W

Contents CONTRIBUTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xiii xv

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES IN AQUEOUS SOLUTION Treavor A. Kendall and Steven K. Lower I. Introduction — Forces in Nature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. Fundamental Forces at the Interface of Biological Particles and Inorganic Surfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The van der Waals Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. The Electrostatic Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. The Solvation Force. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. The Steric Force and Bridging Polymers . . . . . . . . . . . . . . . . . . . . . . . . III. Force Curve Theory and Collecting Force Data. . . . . . . . . . . . . . . . . . . . . A. Force – Distance Curves, Capturing a Potential Force Versus Separation Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Hysteresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Tip Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Spring Constant Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Artifacts in Force Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Data Processing and Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. Advanced Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. Relating Bond Chemistry and Energies to Force Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I. Relevance of Dynamic Force Spectroscopy to Biological – Inorganic Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Forces at the Biomolecule – Mineral Interface . . . . . . . . . . . . . . . . . . . . . . A. Ligand Linkage Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Siderophores and Oxide Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Forces at the Bacterium – Mineral Interface . . . . . . . . . . . . . . . . . . . . . . . . A. Force Microscopy Technique Using Whole Cells . . . . . . . . . . . . . . . . B. Forces Between Escherichia coli and Muscovite . . . . . . . . . . . . . . . . . C. Forces Between Shewanella oneidensis and Goethite or Diaspore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

2 4 4 5 7 8 11 12 15 16 16 17 18 19 19 25 27 28 30 35 35 39 41 44 45 46

vi

CONTENTS

RICE FUNCTIONAL GENOMICS : LARGE -SCALE GENE DISCOVERY AND APPLICATIONS TO CROP IMPROVEMENT Hei Leung and Gynleung An I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. Rice as a Model Genetic System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Diversity of Oryza . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Grass as a Single Genetic System . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. Rice Genome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Genome Sequencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Salient Features of the Rice Genome . . . . . . . . . . . . . . . . . . . . . . . . . C. Comparison Between Indica and Japonica Rice . . . . . . . . . . . . . . . . . IV. Key Ingredients for Gene Discovery in Rice . . . . . . . . . . . . . . . . . . . . . . . A. Genetic Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. High-throughput Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Biological Evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Bioinformatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Forward and Reverse Genetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Insertional Mutants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Chemical- and Irradiation-induced Mutants . . . . . . . . . . . . . . . . . . . . . C. Natural Genetic Variation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Functional Validation of Rice Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Gene Expression Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Gene Silencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Heterologous Bioassays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Gene Replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Allelic Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Applications to Crop Improvement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Candidate Gene Approach and Allele Mining . . . . . . . . . . . . . . . . . . . B. Pathways and Genetic Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Cross-species Inference of Gene Function . . . . . . . . . . . . . . . . . . . . . . VIII. International Collaboration and the Role of Developing Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

56 59 59 62 64 65 66 68 69 70 70 71 72 73 73 83 88 89 89 92 92 94 95 95 96 97 97 98 100 101 102

THE NATURE , PROPERTIES AND MANAGEMENT OF VOLCANIC SOILS R. A. Dahlgren, M. Saigusa and F. C. Ugolini I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. Volcanic Soil Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

114 116

CONTENTS III. Soil Genesis in Volcanic Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Allophanic and Nonallophanic Andisols . . . . . . . . . . . . . . . . . . . . . . B. General Trends in Soil Development on Volcanic Materials. . . . . . . . IV. Soil Classiﬁcation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. World Reference Base for Soil Resources — Andosols . . . . . . . . . . . B. Soil Taxonomy — Andisols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Chemical Weathering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Mineralogical Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Common Colloidal Constituents of Volcanic Soils. . . . . . . . . . . . . . . B. Formation and Transformation of Colloids in Volcanic Soils . . . . . . . VII. Selected Chemical Characteristics of Volcanic Soils . . . . . . . . . . . . . . . . A. Organic Matter Accumulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Aluminum Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII. Productivity and Management of Volcanic Soils . . . . . . . . . . . . . . . . . . . A. Charge Characteristics and Chemical Fertility . . . . . . . . . . . . . . . . . . B. Physical Properties and Fertility . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Soil Management and Conservation of Volcanic Soils . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii 118 119 122 123 124 125 127 134 135 141 145 145 147 150 150 163 166 167

MICROARRAY TECHNOLOGY AND APPLICATIONS IN ENVIRONMENTAL MICROBIOLOGY Jizhong Zhou and Dorothea K. Thompson I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. Microarray Types and Advantages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Types of Microarrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Advantages of Microarrays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. Microarray Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Microarray Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Surface Modiﬁcation for the Attachment of Nucleic Acids. . . . . . . . . C. Arraying Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Critical Issues for Microarray Fabrication . . . . . . . . . . . . . . . . . . . . . IV. Microarray Hybridization and Detection . . . . . . . . . . . . . . . . . . . . . . . . . A. Probe Design and Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Target Labeling and Quality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Hybridization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Critical Issues in Hybridization and Detection . . . . . . . . . . . . . . . . . . V. Microarray Image Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Assessment of Spot Quality and Background Subtraction. . . . . . . . . . VI. Microarray Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Data Normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

184 187 187 188 190 190 191 197 201 205 205 206 210 210 211 214 214 215 219 219

viii

CONTENTS

B. Data Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Methods for Identifying Differentially Expressed Genes . . . . . . . . . . D. Microarray Data Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Using Microarrays to Monitor Genomic Expression . . . . . . . . . . . . . . . . A. General Approaches to Revealing Differences in Gene Expression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Experimental Design for Microarray-based Monitoring of Gene Expression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Microarray-based Functional Analysis of Environmental Microorganisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII. Application of Microarrays to Environmental Studies . . . . . . . . . . . . . . . A. Functional Gene Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Phylogenetic Oligonucleotide Arrays . . . . . . . . . . . . . . . . . . . . . . . . C. Community Genome Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Whole-genome Open Reading Frame Arrays for Revealing Genome Differences and Relatedness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Other Types of Microarrays for Microbial Detection and Characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

225 225 227 231 231 234 236 242 243 250 253 256 258 258 259

THE AGRONOMY AND ECONOMY OF BLACK PEPPER (PIPER NIGRUM L.) — THE “KING OF SPICES ” K. P. Prabhakaran Nair I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. The Pepper Plant — Its Botany and Chemistry . . . . . . . . . . . . . . . . . . . . A. Pepper Botany . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Pepper Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. Pepper Agronomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The Pepper Soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Nutrition of Black Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Evolution of Pepper Manuring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Response of Pepper to Mineral Nutrients . . . . . . . . . . . . . . . . . . . . . IV. The Role of “The Nutrient Buffer Power Concept” in Pepper Nutrition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The Buffer Power and Effect on Nutrient Availability . . . . . . . . . . . B. Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Measuring the Nutrient Buffer Power and its Importance in Affecting Nutrient Concentration on Root Surfaces . . . . . . . . . . . . . . . . . . . . . D. Background Informaton on the Importance of Measuring Zn Buffer Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Quantifying Zn Buffer Power of Pepper Growing Soils . . . . . . . . . . .

273 279 279 285 288 289 290 291 291 296 296 296 298 301 303

CONTENTS V. Establishing a Pepper Plantation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The Indian Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. The Indonesian Experience. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Pepper Pests and Their Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. The Processing of Black Pepper On-Farm. . . . . . . . . . . . . . . . . . . . . . . . A. Sun Drying of Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Solar Drying of Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Garbling, Cleaning, and Fractionation . . . . . . . . . . . . . . . . . . . . . . . . D. Packaging and Storing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII. An Account of Indonesian Pepper Processing . . . . . . . . . . . . . . . . . . . . . IX. Industrial Processing of Black Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . . A. White Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Cryoground Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Pepper Oil and Oleoresin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . X. The Future of the Global Pepper Economy . . . . . . . . . . . . . . . . . . . . . . . A. The Supply Side of the Pepper Economy . . . . . . . . . . . . . . . . . . . . . B. The Demand Side of the Pepper Economy . . . . . . . . . . . . . . . . . . . . C. Prices and World Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. The Pepper Price Outlook by 2020 . . . . . . . . . . . . . . . . . . . . . . . . . . E. The Pepper Supply Outlook by 2020 . . . . . . . . . . . . . . . . . . . . . . . . F. The Pepper Demand Outlook by 2020 . . . . . . . . . . . . . . . . . . . . . . . G. Country-wise Economic Growth Impacting Production and Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XI. Pepper Economy in India. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Pepper Production Scenario in India . . . . . . . . . . . . . . . . . . . . . . . . . B. Bush Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Economics of Pepper Production in the State of Kerala . . . . . . . . . . . D. Marketing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Pepper Futures Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XII. Pepper Pharmacopoeia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Antiinﬂammatory and Central Nervous System (CNS) Depressant Activity of Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Effect on Hepatic Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Carcinogenic and Mutagenic Effects of Black Pepper . . . . . . . . . . . . D. Pepper as an Antioxidant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Pepper as an Antimicrobial Agent . . . . . . . . . . . . . . . . . . . . . . . . . . F. The Pharmacological Effect of Pepper on Human Health . . . . . . . . . . G. Clinical Applications of Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. Toxicological Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I. The Insecticidal Activity of Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . XIII. Consumer Products Out of Black Pepper . . . . . . . . . . . . . . . . . . . . . . . . XIV. Value Addition in Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XV. Conclusions and a Peep Into Pepper’s Future . . . . . . . . . . . . . . . . . . . . . Acknowledgment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix 306 306 310 311 322 324 324 325 327 328 330 333 335 335 339 339 343 345 345 346 346 346 347 349 351 352 352 353 355 357 358 359 360 361 362 363 363 364 365 365 369 372 373

x

CONTENTS

SOIL MINERAL – ORGANIC MATTER –MICROORGANISM INTERACTIONS : FUNDAMENTALS AND IMPACTS P. M. Huang I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. Mechanisms of Binding of Nonhumic Organics and Humic Substances by Soil Mineral Colloids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Nature of Mineral Colloid Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . B. Binding of Nonhumic Organics . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Binding of Humic Substances. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. Inﬂuence of Organic Substances on the Formation, Transformation, and Surface Properties of Metal Oxides . . . . . . . . . . . . . . . . . . . . . . . . . A. Aluminum Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Iron Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Role of Soil Minerals in Abiotic Catalysis of the Formation of Humic Substances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Polyphenol Pathway. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Maillard Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Interactions of Soil Mineral Colloids with Microorganisms . . . . . . . . . . . A. Surface Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Inﬂuence on Microbial Activity and Survival . . . . . . . . . . . . . . . . . . VI. Interactions of Soil Colloids with Enzymes. . . . . . . . . . . . . . . . . . . . . . . A. Mineral Colloid – and Humic– enzyme Complexes . . . . . . . . . . . . . . B. Effects on Enzymatic Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Microbial Mediation of Soil Mineral Weathering and Transformation . . . A. Mineral Weathering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Fine-grained Mineral Development. . . . . . . . . . . . . . . . . . . . . . . . . . VIII. Interactions of Soil Mineral Colloids with Organic Substances and Microorganisms in Relation to Soil Structure Stability . . . . . . . . . . . . . . A. Organo-mineral Complexes in Relation to Soil Structure . . . . . . . . . . B. Dynamics of Aggregate Turnover . . . . . . . . . . . . . . . . . . . . . . . . . . . IX. Inﬂuence of Mineral Colloids on Biogeochemical Cycling of C, N, P, and S in Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Role of Mineral Colloids in Soil Organic Matter Storage and Turnover B. Decomposition and Stabilization of C, N, P, and S in Relation to Primary and Secondary Particles. . . . . . . . . . . . . . . . . . . . . . . . . . . . X. Effects of Interactions between Microorganisms and Soil Colloids on the Transformation of Organic Pollutants . . . . . . . . . . . . . . . . . . . . . . A. Catalytic Transformations of Organic Pollutants . . . . . . . . . . . . . . . . B. Binding of Organic Pollutants, Enzymes, and Microorganisms on Mineral and Humic Surfaces and the Effects on Pollutant Bioavailability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XI. Impact of Interactions of Physicochemical, Biochemical, and Biological Processes on Metal Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Redox Reactions of Metals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

393 394 394 395 397 400 400 404 406 406 410 410 410 413 415 415 417 418 418 419 421 421 423 426 426 429 431 431

436 440 440

CONTENTS B. C. D. E. F.

Complexation Reactions of Metals . . . . . . . . . . . . . . . . . . . . . . . . . . Adsorption –Desorption Reactions of Metals . . . . . . . . . . . . . . . . . . . Precipitation – Dissolution Reactions of Metals . . . . . . . . . . . . . . . . . Metal Sorption and Uptake by Microorganisms and Biomineralization Metal Transformation by Microbial Excretions and Mycorrhizal Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XII. Foreseeable Impacts of Soil Mineral – Organic Matter – Microorganism Interactions on the Ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Global Ion Cycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Global Climate Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Biodiversity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Biological Productivity and Human Nutrition . . . . . . . . . . . . . . . . . . E. Geomedicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Ecotoxicology and Human Health. . . . . . . . . . . . . . . . . . . . . . . . . . . G. Biotechnology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. Risk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I. Risk Management, Remediation, and Restoration . . . . . . . . . . . . . . . XIII. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi 441 442 444 445 447 447 448 449 451 451 452 453 454 456 457 459 459 460

LOW EXTERNAL INPUT TECHNOLOGIES FOR LIVELIHOOD IMPROVEMENT IN SUBSISTENCE AGRICULTURE Anil Graves, Robin Matthews and Kevin Waldie I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. The Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Intercropping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Alley Cropping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Cover Crops and Green Manures . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Biomass Transfer Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Compost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Animal Manure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. Improved Fallows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. Generic Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Soil Fertility Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Socio-economic Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Integrated Nutrient Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . B. A Systems Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

473 476 476 479 484 490 493 496 499 504 504 516 526 526 530 534 539 541 541

xii

CONTENTS

AMMONIA EMISSION FROM MINERAL FERTILIZERS AND FERTILIZED CROPS Sven G. Sommer, Jan K. Schjoerring and O. T. Denmead I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. Mineral Fertilizer Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. Ammonia Volatilization from Mineral Fertilizers . . . . . . . . . . . . . . . . . . A. Production and Transport of NH3 in the Soil– Fertilizer – Atmosphere Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Temporal NH3 Loss Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Hydrolysis of Urea. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Soil Hþ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Soil CEC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Solid Phase Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. Climate and Inﬁltration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. Microbial Processes (Nitriﬁcation/Immobilization) . . . . . . . . . . . . . . IV. Ammonia Emission from Crop Foliage. . . . . . . . . . . . . . . . . . . . . . . . . . A. Transport of NH3 Between Leaves and the Atmosphere. . . . . . . . . . . B. Magnitude of NH3 Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Physiological Processes Involved in NH3 Emission from Crops . . . . . V. Management Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Techniques for Reduction of NH3 Emission . . . . . . . . . . . . . . . . . . . B. Fertilizer Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Flooded Fields (Rice Paddies) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Injection of Anhydrous Ammonia. . . . . . . . . . . . . . . . . . . . . . . . . . . E. Crop NH3 Emissions as Affected by Fertilizer Application . . . . . . . . F. Ammonia Emission from Decomposing Plant Material . . . . . . . . . . . G. Absorption by Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Measurement Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Tracer Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Enclosures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Micrometeorological Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Gradient Diffusion Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Eddy Correlation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Relaxed Eddy Accumulation or Conditional Sampling. . . . . . . . . . . . G. Lagrangian Dispersion Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Conclusions and Perspectives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

558 560 561 561 565 568 570 574 575 576 578 581 581 582 584 586 587 589 591 594 594 596 598 598 598 599 601 603 605 605 606 607 609 611 611

INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

623

Contributors Numbers in parentheses indicate the pages on which the authors’ contributions begin.

G. AN (55), Department of Life Science, Pohang University of Science and Technology, San 31 Hyoja-dong, Nam-gu, Pohang, Korea 790-784 R. A. DAHLGREN (113), Land, Air and Water Resources, University of California, Davis, California 95616, USA O. T. DENMEAD (557), CSIRO Land and Water, GPO Box 1666, Canberra ACT 2601, Australia A. GRAVES (473), Institute of Water and Environment, Cranﬁeld University, Silsoe, Bedfordshire MK45 4DT, UK P. M. HUANG (391), Department of Soil Science, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK, S7N 5A8 Canada T. A. KENDALL (1), Division of Engineering and Applied Sciences, Harvard University, 40 Oxford St., Cambridge, MA 02138, USA H. LEUNG (55), International Rice Research Institute, DAPO Box 7777, Metro Manila, Philippines S. K. LOWER (1), Department of Geological Sciences & School of Natural Resources, The Ohio State University, 275 Mendenhall Laboratory, 125 South Oval Mall, Columbus, Ohio 43210, USA R. MATTHEWS (473), Institute of Water and Environment, Cranﬁeld University, Silsoe, Bedfordshire MK45 4DT, UK K. P. PRABHAKARAN NAIR (271), Akshaya, East Hill, Calicut 673 005, Kerala State, India M. SAIGUSA (113), Experimental Farm of Tohoku University, Kawatabi, Naruko, Tamatsukuri, Miyagi 989-6711, Japan J. K. SCHJOERRING (557), Plant Nutrition Laboratory, Department of Agricultural Sciences, The Royal Veterinary and Agricultural University, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmark S. G. SOMMER (557), Department of Agricultural Engineering, Danish Institute of Agricultural Sciences, Research Centre Bygholm, PO Box 536, DK-8700 Horsens, Denmark D. K. THOMPSON (183), Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA F. C. UGOLINI (113), Dipartimento di Scienza del Suolo e Nutrizione della Pianta, Universita` Degli Studi, Piazzale delle Cascine 15, 50144 Firenze, Italy K. WALDIE (473), International and Rural Development Department, University of Reading, Reading RG6 6AH, UK J. ZHOU (183), Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA

xiii

This Page Intentionally Left Blank

Preface Volume 82 continues the tradition and excellence of “Advances in Agronomy” by containing eight state-of-the-art reviews on topics of interest in the plant and soil sciences. Three of the reviews present cutting-edge molecular scale techniques and approaches that directly impact food production, crop improvement, and environmental quality and sustainability. Chapter 1 is an excellent review on the use of force microscopy to investigate reactions at the mineral/microbe interface. Chapter 2 is a deﬁnitive review on rice functional genomics including large-scale gene discovery and applications to crop improvement. Chapter 3 presents a comprehensive overview of volcanic soils including their genesis, chemical, mineralogical, and physical properties and the productivity and management of volcanic soils. Chapter 4 is a seminal review on microarray technology and applications to environmental microbiology. Microarray types, fabrication, hybridization and detection, image processing, data analysis, the use of microarrays to monitor genomic expression, and applications of microarrays to environmental studies are covered. Chapter 5 is a deﬁnitive overview on the agronomy and economy of black pepper. Topics that are discussed include: the botany and chemistry of black pepper, pepper agronomy, production and management, and economic aspects of global black pepper production. Chapter 6 is an outstanding overview of soil mineral-organic matter-microorganism interactions. Topics that are discussed include: mechanisms of binding of nonhumic organics and humic substances by soil mineral colloids, the inﬂuence of organic substances on the surface chemistry and transformations of metal oxides, interactions of minerals with microorganisms and enzymes, the role of microbes on mineral weathering, and impacts of mineral-humic-microbe interactions on nutrient cycling and transport of pollutants and ecosystems. Chapter 7 reviews low external input technologies for livelihood improvement in subsistence agriculture. The technologies are discussed along with economic and nutrient management considerations. Chapter 8 is a comprehensive review on ammonia emission from mineral fertilizers and fertilized crops. Topics that are covered include: ammonia volatilization from mineral fertilizers, ammonia emission from crop foliage, and management strategies and techniques. Many thanks to all the authors for their excellent reviews. DONALD L. SPARKS University of Delaware

xv

This Page Intentionally Left Blank

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES IN AQUEOUS SOLUTION Treavor A. Kendall1 and Steven K. Lower2 1

Harvard University, Division of Engineering and Applied Sciences, 40 Oxford St., Cambridge, MA 02138, USA 2 Department of Geological Sciences & School of Natural Resources, The Ohio State University, 275 Mendenhall Laboratory, 125 South Oval Mall, Columbus, Ohio 43210, USA, Email: [email protected]

I. Introduction — Forces in Nature II. Fundamental Forces at the Interface of Biological Particles and Inorganic Surfaces A. The van der Waals Force B. The Electrostatic Force C. The Solvation Force D. The Steric Force and Bridging Polymers III. Force Curve Theory and Collecting Force Data A. Force– distance Curves, Capturing a Potential Force Versus Separation Plot B. Hysteresis C. Tip Shape D. Spring Constant Determination E. Artifacts in Force Measurements F. Data Processing and Statistics G. Advanced Algorithms H. Relating Bond Chemistry and Energies to Force Measurements I. Relevance of Dynamic Force Spectroscopy to Biological – Inorganic Interface IV. Forces at the Biomolecule– Mineral Interface A. Ligand Linkage Schemes B. Siderophores and Oxide Surfaces V. Forces at the Bacterium– Mineral Interface A. Force Microscopy Technique Using Whole Cells B. Forces Between Escherichia coli and Muscovite C. Forces Between Shewanella oneidensis and Goethite or Diaspore VI. Future Work Acknowledgments References

1 Advances in Agronomy, Volume 82 Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved. 0065-2113/03 $35.00

2

T. A. KENDALL AND S. K. LOWER

At the most fundamental level, intermolecular forces (e.g., van der Waals, electrostatic, solvation, steric) control interactions between biological molecules and mineral surfaces. These are forces with magnitudes of piconewtons to nanonewtons, which operate in a space that is on the order of nanometers. We have used force microscopy to quantitatively probe forces, energies, and distances between crystal surfaces and living microbial cells or biological molecules in their native state. The systems we have studied include those involving: Escherichia coli, Shewanella oneidensis, siderophores, muscovite, goethite, and/or diaspore, in aqueous solutions of varying composition. Direct force measurements at the organic – inorganic interface have been interpreted with theoretical models describing interfacial forces, adhesion, and molecular dynamic calculations. A new perspective on bacterium – mineral interactions is emerging from these studies. We have discovered a world that operates under a very different set of principles than macroscopic bodies. A world where the intermolecular force, rather than gravitational attraction, is the preeminent force controlling the evolution of processes at the bacterium – mineral interface. q 2004 Academic Press.

I. INTRODUCTION — FORCES IN NATURE The bacterium – mineral interface is ubiquitous near the surface of the Earth. As many as 97% of the , 1030 prokaryotes on Earth live in close proximity to minerals in soil, marine, and terrestrial subsurface environments (Whitman et al., 1998). As we will show in this manuscript, the fundamental forces at this interface are very small, seemingly insigniﬁcant. This review will provide evidence that forces on the order of nanonewtons (1029 N) to piconewtons (10212 N) dominate the properties/processes at bacterium –mineral and biomolecule – mineral interfaces. For comparison, there is , 0.2 nN of gravitational attraction between a person (50 kg) and the paper (5 g) upon which these words are written. Despite their small magnitude, these forces are at the heart of all interactions between biologically produced polymers and mineral surfaces in nature. It is now well established that there are four fundamental forces in nature: the strong and weak nuclear forces, the gravitational interaction, and electromagnetic forces, which are the source of all intermolecular forces (Israelachvili, 1992). Because the ﬁrst two (i.e., nuclear forces) have a range of action that is less than 1025 nm (Israelachvili, 1992), we need not consider these for interactions between biological molecules, microbial cells, and/or mineral surfaces. The question then becomes, under what conditions do gravitational forces or electromagnetic forces (more speciﬁcally, intermolecular forces) dominate bacteria – mineral or biomolecule – mineral interactions?

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

3

In nature, living organisms exist in communities that are in contact with one another, in contact with mineral surfaces, and they are also in contact with the surface of the Earth (i.e., the upper crust). For simplicity, let us deﬁne a particular species of organism as a spherical particle (having a density of water) with a unique size or radius. Each species may interact with one another and/or the Earth. In both instances, there is a force of gravitational attraction at each interface. Figure 1 reveals that the gravitational attraction is much greater between the Earth and a particle of a given size (e.g., , 4 £ 1025 N for a 1 mm particle) relative to the gravitational attraction between two particles of the same given size (e.g., , 3 £ 10216 N between two 1 mm particles). Also shown on this ﬁgure is a theoretical prediction for another type of attractive force, the so-called van der Waals force. This intermolecular force was determined using Eq. (1) (see below) to describe the attraction between two similar objects of equal size in contact with one another. For example, two identical 1 mm (radius) particles are

Figure 1 Log–log plot of the theoretical forces describing (1) gravitational attraction between a particle and the Earth (solid “Earth-particle” line), (2) gravitational attraction between two particles of the same size (dashed “particle-particle” line), and (3) van der Waals attraction between two particles of the same size (dashed “vdw” lines). In all instances the particles are assumed to be in “contact” with the Earth (for 1) or another particle (for 2 and 3). For gravitational attraction, mass was determined by assuming each particle was a solid homogeneous sphere with a density of 1 g cm23, and contact was deﬁned as the radius of the Earth (~6.4 £ 106 m radius; “Earth-particle” interaction) or the sum of the radii of two interacting particles (“particle-particle” interaction). The shaded region outlines the boundaries of the expected van der Waals force using values for Hamaker constant of 10220 to 10221 J, which is appropriate for biological and inorganic phases (Israelachvili, 1992; Leckband and Israelachvili, 2001; Vigeant et al., 2002), and deﬁning “contact” as an effective separation between particles of ~0.2 (for one hydration layer) to 2 nm, according to Israelachvili (1992) and Leckband and Israelachvili (2001). Only the magnitudes of the forces are shown. By convention, attractive forces (shown here) are negative. For reference, the three diamond symbols represent gravitational forces between the Earth (~1024 kg) and each of three bodies (from left to right): a bacterium (10215 kg), a human (50 kg), or the moon (1022 kg).

4

T. A. KENDALL AND S. K. LOWER

expected to have an attractive, adhesion force at contact (due solely to the van der Waals force) equal to , 3 £ 1025 N (Hamaker constant ¼ 10220 J; effective separation ¼ 0.165 nm, i.e., the “universal” cut-off separation, Israelachvili, 1992). This force magnitude is approximately the same as the gravitational force between the Earth’s surface and one of these 1 mm particles. While it is debatable whether the van der Waals force applies in the same manner to both a particle of the size of an atom and an object of the size of the moon, the predictions shown in Fig. 1 for objects smaller than , 1 cm are in agreement with others (e.g., Israelachvili, 1992). Consequently, the force of gravity may dominate the interactions between macroscopic bodies (e.g., plants and animals), but intermolecular forces (e.g., van der Waals and others, see below) are the prevailing forces with which microscopic bacteria must contend. This is particularly true when one considers that the van der Waals force is signiﬁcantly weaker and shorter range than other intermolecular forces, such as electrostatic and hydrophobic interactions as discussed below.

II. FUNDAMENTAL FORCES AT THE INTERFACE OF BIOLOGICAL PARTICLES AND INORGANIC SURFACES “All intermolecular forces are essentially electrostatic in origin” (page 11, Israelachvili, 1992). In theory, classical electrostatics could be used to calculate intermolecular forces if one could determine the spatial distribution of the electron cloud by solving the Schro¨dinger equation (Israelachvili, 1992). Unfortunately this is challenging for even simple atomic interactions in vacuum, never mind molecular or organism scale interactions between different functional groups on bacteria and minerals in water. For this reason, it is useful to classify four types of intermolecular forces that are expected to dominate the bacterium – mineral and biomolecule –mineral interfaces. These include the van der Waals force, electrostatic forces, solvation interactions, and steric or entropic forces (Israelachvili and McGuiggan, 1988). The reader is referred to a number of excellent reviews on these types of forces (e.g., Israelachvili and McGuiggan, 1988; Butt et al., 1995; Leckband and Israelachvili, 2001). This review will touch on all four types of intermolecular forces, although the van der Waals and electrostatic forces will be explored in more detail.

A. THE

VAN DER

WAALS FORCE

The van der Waals force, like the force of gravity, acts between all particles (Israelachvili, 1992). It is quantum mechanical in origin and arises because of

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

5

the time dependent ﬂuctuations in the electric dipole moment of a particle as it comes into contact with other particles nearby. Even nonpolar particles, which have a time averaged dipole moment of zero, have instantaneous dipoles due to the movement of electrons relative to protons in a nucleus. Dipoles generate an electric ﬁeld that polarizes adjacent particles and gives rise to an instantaneous force between neighboring particles. Two terms describe the van der Waals force: the ﬁrst polarization potential, which represents the energy necessary to ionize an atom (i.e., a dipole moment due to interactions between electrons and protons within a single particle); and the so-called dispersion term, which describes the dipole induced interactions between two or more atoms (Israelachvili, 1992). Because the dispersion term dominates the van der Waals force, it is sometimes referred to as (London) dispersion forces (Butt et al., 1995). The van der Waals force has an inverse power law dependence on the separation between two particles. For atoms and small molecules the van der Waals force is , D 27, where D is the separation distance between particles. It can be attractive or repulsive (e.g., it is always attractive between two similar particles immersed in a third liquid) and is described in terms of the Hamaker constant (Ha), which depends upon the refractive indices and dielectric constants of the interacting particles and intervening media (see Israelachvili, 1992). Hamaker constants are in the order of 10220 to 10221 J, for biological cells or molecules interacting with themselves or minerals across an aqueous solution (Ducker et al., 1991; Butt et al., 1995; Ong et al., 1999; Bhattacharjee et al., 2000; Leckband and Israelachvili, 2001; Vigeant et al., 2002). For simple geometries, the forces between atoms or molecules can be assumed to be additive (Israelachvili, 1992; Butt et al., 1995) such that equations can be derived for larger particles (e.g., organic and inorganic surfaces). Two commonly encountered geometric conﬁgurations include interactions between two spheres or a sphere and a ﬂat surface, both of which are given by Israelachvili (1992), Butt et al. (1995) and Leckband and Israelachvili (2001): FðDÞ ¼

2Ha Rx 6D2

ð1Þ

where Ha is the Hamaker constant (J), D is the separation distance (m) between the two spheres or a sphere and a plane, and Rx (m) equals the radius of the sphere for the sphere– plane conﬁguration, or it is equal to (R1R2/(R1þ R2)) for the interaction between two spheres of radius R1 and R2. A positive Hamaker constant indicates attraction (negative force sign).

B. THE ELECTROSTATIC FORCE The electrostatic force arises through a variety of mechanisms leading to the development of surface charge (e.g., see Sposito, 1989). Water, which has a high

6

T. A. KENDALL AND S. K. LOWER

dielectric constant, causes the dissociation of surface functional groups. These functional groups display protonation/deprotonation reactions that are dependent upon pH. For example, in water, silanol groups on a silica surface undergo the following reaction: . Si OH ¼ . Si O2 þ Hþ. Similar acid – base reactions take place on carboxylic groups, amine groups, and other reactive moieties on biological molecules and inorganic surfaces. Hence, many inorganic and biological surfaces develop a charge that is dependent upon pH. Other factors such as the adsorption of charged ions and presence of permanent structural charge (e.g., for clays) are additional contributors to surface charge. The overall charge on a surface is balanced by the dissolved counterions in solution, which are attracted to the surface by its electric ﬁeld and dispersed such that they (i.e., the counterions) increase entropy (Butt et al., 1995). This creates the so-called electric double-layer around surfaces immersed in aqueous solution (Stumm, 1992). When two charged surfaces approach one another, the electric double-layers are perturbed resulting in an electrostatic interaction. This interaction may be attractive (if surfaces are of opposite charge) or repulsive (if surfaces are similarly charged). The electrostatic force varies exponentially with the distance between particles. It depends strongly upon the surface charge densities of the interacting particles and the ionic strength of the intervening solution. Similar to the van der Waals force (see above) equations can be derived to describe the electrostatic force for various geometric conﬁgurations. The model for electrostatic forces between two spheres or a sphere and ﬂat surface is (Butt et al., 1995; Muller and Engel, 1997; Leckband and Israelachvili, 2001): FðDÞ ¼

4ps1 s2 Rx 2kD e 110 k

ð2Þ

where s is the surface charge density (C m22) of particles 1 and 2, 1 is the dielectric constant of water (78.54 at 298 K), 10 is the permittivity of free space (8.854 £ 10212 C2 J21 m21), Rx and D are deﬁned as above. The Debye length (1/k) describes the thickness of the diffuse double-layer of counterions that surrounds charged particles in solution. The Debye length depends upon the valence and concentration (c, mol L21) of the electrolyte. For monovalent electrolytes (e.g., NaCl) at a temperature of 298 K, the Debye length (in nm) ¼ 0.304/ (c)1/2; for 1:2 or 2:1 electrolytes (e.g., CaCl2) it is 0.174/(c)1/2; for 2:2 electrolytes it is 0.152/(c)1/2 (Muller and Engel, 1997). In many instances, it is easier to determine a particle’s surface potential as opposed to surface charge. The Graham equation can be used to relate these two parameters according to (Stumm, 1992), qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ z cF ð3Þ s ¼ 8RT110 c £ 103 £ sinh 2RT where R is the gas constant (8.314 J mol21 K21), T is the temperature (K), z is the valence of ions in solution, c is the surface potential (V), and F is the Faraday constant (96,490 C mol21). A potential measured across an interface

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

7

contains contributions from at least two layers, the so-called Stern layer and the “diffuse” layer (see Stumm, 1992). Techniques such as streaming potential and electrophoresis are commonly used to determine a particle’s zeta potential, which is used as a proxy for surface potential. However, the zeta potential probably represents only the “diffuse” double-layer, which is lower than the true surface potential (Stumm, 1992). Leckband and Israelachvili (2001) describe the differences for surfaces that are assumed to have a constant surface charge versus those that are assumed to have a constant surface potential. Interactions at constant surface charge are expected to occur when surface ionizable groups are fully dissociated and remain as such for all separations (D). This may be true when the pH of a solution is much greater than the pK value(s) of a particular protonation/deprotonation reaction(s). In instances where surface functional groups are not fully ionized but in equilibrium with solution ions, interactions at constant potential are expected to occur. In this latter case, as two surfaces come together (i.e., very small D) the intervening concentration of solution ions increases locally such that some solution ions bind to the surface thereby reducing that surface’s density of charged sites (Leckband and Israelachvili, 2001). For many instances, this distinction inﬂuences the interaction only at small separations where these two conditions deﬁne the boundaries of the expected electrostatic force.

C. THE SOLVATION FORCE The origin, theory, and force – distance relationships of the remaining two force classes — solvation and steric — are indeﬁnite compared to the forces discussed above. Much work remains to be done before solvation and steric forces can be appreciated to the same extent as the van der Waals and electrostatic forces. However, it is well established that the models developed for the van der Waals and electrostatic forces, which treat the intervening solution as a continuum, break down when two particles or surfaces are within a few nanometers (Butt et al., 1995; Leckband and Israelachvili, 2001). At such close separations, solvation forces may dominate because the solvent (e.g., water) takes on a more ordered structure. Steric forces may also come into play for surfaces with polymers (e.g., biological cells or particles). Our discussion of solvation and steric forces will be more qualitative however, because general force laws (such as those described above) are relatively sparse for these latter two force classes. Solvation forces (also called hydration or structural forces when the solvent is water) seem to be the result of interactions of solvent molecules with themselves (e.g., in a conﬁned space between two surfaces) or interactions between solvent molecules and a surface (e.g., the orientation of water molecules at the interface of a strongly hydrophilic surface). As two surfaces approach one another the

8

T. A. KENDALL AND S. K. LOWER

intervening liquid ceases to behave as a structureless media resulting in a force that can be attractive, repulsive, or oscillatory (Butt et al., 1995). These forces can be further subdivided into those that result from solvent –solvent, solvent – surface, and surface – surface interactions (Israelachvili and McGuiggan, 1988). For two rigid crystalline surfaces at short-range (, 2 nm for water), water molecules interact with themselves such that they take on a semi-discrete layering or structure, which causes the “structural” forces between the interacting surfaces to oscillate between attraction and repulsion with a periodicity equal to the molecular dimension of water (Leckband and Israelachvili, 2001). Between surfaces with polymers, water cannot form well-deﬁned layers because headgroups on lipids, for example, are “rough” on the scale of a water molecule (Israelachvili, 1992), and macromolecules in surfaces are thermally mobile (Beveridge, 1999). Consequently, any repulsion is smeared out and takes on a monotonic component (Israelachvili, 1992). For strongly hydrophilic surfaces in aqueous solution, there is a strong solvent –surface interaction that leads to the formation of hydration shells. These ordered water molecules within the “shell” generate an electric ﬁeld that impinges upon another surface as two particles approach to within a few nanometers of one another (Israelachvili and McGuiggan, 1988). For example, water molecules may associate with two, adjacent hydrophillic surfaces such that the water’s hydrogens are oriented towards each surface (attracted via hydrogen bonds) and the water oxygens are exposed to the solution. This confers a negative character (from the lone pairs of the water’s oxygens) to each surface, thereby generating a repulsive force. Conversely, the dipoles may complement one another forming an attractive force if water molecules are staggered on the two surfaces. This hydration force may extend outwards more than the oscillatory force discussed previously (Leckband and Israelachvili, 2001). Finally, for nonpolar surfaces that cannot bind to water molecules — the so-called hydrophobic surfaces (deﬁned as those surfaces having a contact angle of 75– 1158 with water) — there is often a strong attractive force that extends to separations of tens of nanometers or greater (Leckband and Israelachvili, 2001). Hydrophobic forces can be signiﬁcantly greater than the van der Waals force and may play an important role in interactions involving hydrophobic molecules and/or surfaces (Israelachvili and McGuiggan, 1988; Israelachvili, 1992).

D. THE STERIC FORCE

AND

BRIDGING POLYMERS

The steric force affects surfaces that have ﬂexible polymers extending out into solution (e.g., polysaccharides on biological cells). As two surfaces approach one another, the polymer chains become conﬁned such that they are not free to move at random. This entropic conﬁnement results in a repulsive force whose length scale is approximately equal to the radius of gyration of the polymer

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

9

(Butt et al., 1995), where the radius of gyration is proportional to the number of monomer segments raised to some power between 0.33 (for poor solvents) and 0.6 (for good solvents) (Leckband and Israelachvili, 2001). Approximations derived for the interaction between two ﬂat surfaces reveal that this force depends on the surface coverage of the polymer and may take an exponential form (Israelachvili, 1992; Leckband and Israelachvili, 2001). At close separation, the magnitude of the steric force can be similar to that of the electrostatic force (Leckband and Israelachvili, 2001). In some instances, surface-bound polymers may form an attractive interaction at close separation as the polymer forms a “bridge” between two particles or surfaces (Leckband and Israelachvili, 2001). The resulting adhesive bond may be very long range (i.e., extend well beyond the radius of gyration of the polymer) and resist separation when the surfaces are pulled apart (Jeppesen et al., 2001). While there is no general description for attractive bridging forces by polymers, the linkage of surfaces via a polymeric tether has been described by the so-called freely jointed chain (see e.g., Leckband and Israelachvili, 2001), or worm-like chain models (see e.g., Flory, 1989; Bustamante et al., 1994). In the case of the latter, the polymer is viewed as an elastic element and the force (F) needed to stretch the tethered polymer to a length x is: FðxÞ ¼ ðkB T=bÞ½0:25ð1 2 x=LÞ22 2 0:25 þ x=L

ð4Þ

where kB is the Boltzmann’s constant, T is the temperature, b is the persistence length (i.e., length of the stiff segment or monomer of the chain), and L is the contour length (i.e., length of the completely stretched chain). Polymer bridging is a phenomenon that crosses between the disciplines of colloidal science — which has historically tended to investigate intermolecular forces that dominate the interface between two rigid surfaces that are approaching one another — and adhesion science — which is interested in describing the contact between two surfaces and the forces necessary to pull them apart. While attractive intermolecular and intersurface forces (i.e., the four force classes discussed above) are responsible for adhesion events, real particles (e.g., bacteria and minerals) that make contact will also adhere to one another due to elastic or ﬂuid-like deformation, which is an intrinsic and natural part of contact. There is a wealth of information on adhesion processes and theories including the Johnson – Kendall –Roberts (Johnson et al., 1971) and Derjaguin – Muller – Toporov (Derjaguin et al., 1975) theories, which relate the force required to pull two surfaces apart (i.e., the “pull off” force) to the surface energy, surface tension, or work of adhesion. Sufﬁce it to say that surface energy (or tension or work) is determined from intermolecular forces between surfaces. For particles or surfaces that are incapable of forming hydrogen bonding (e.g., nonmetallic compounds), the surface energy can be related directly to the van der Waals force, where surface tension < Ha/2.1 £ 10221 (Israelachvili, 1992). The surface

10

T. A. KENDALL AND S. K. LOWER

energies of more polar surfaces, which tend to be larger, are dependent upon van der Waals interactions, as well as an additional electrostatic-like term that relates surface energy to Lewis acid/base reactions (van Oss, 1993). These four forces — van der Waals, electrostatic, solvation (hydration and hydrophobic), and steric — operate concurrently at the interfaces between microorganisms, biological molecules, and/or mineral surfaces (see Table 1). In some instances, one force may dominate at all separations. In other instances, there is a delicate balance such that each force dominates at its own length scale. These four force classes are often invoked to describe interactions as two surfaces approach one another. Two particles that are pulled apart may experience the same sign, magnitude, and range of forces that existed upon approach. However, there is often a notable hysteresis between the forces measured upon approach versus those that are observed upon retraction for soft biological particles and surfaces. This is due to the formation of adhesive bonds (e.g., see discussion of polymer bridges and adhesion, above) once contact has been established between surfaces. This review will provide examples that illustrate the various forces and force models discussed above as they pertain to interactions between biological and inorganic particles. Further, we will discuss the differences between those forces measured as surfaces

Table I Summary of Physical Forces of Interaction Between Particles and/or Surfacesp Type of interaction van der Waals Electrostatic

Solvation

Steric

Description Force between all particles due to polarization; usually attractive; short-range Force between charged particles; attractive (for particles of opposite sign) or repulsive (for particles of similar sign); depends upon ionic strength of solution; short to long range Structural or hydration forces are typically repulsive due to sorbed water layers; short-range Hydrophobic force is attractive between nonpolar surfaces; short to long range Typically a short-range, repulsive force associated with polymers; may be longer range, attractive force for “bridging” polymers.

p This review has followed the force characterization of Israelachvili (1988). Other force classes, such as hydrogen bonding or thermal ﬂuctuations, may dominate when two particles or surfaces are very close. However, these other classes can often be described as a subset of electrostatic (for hydrogen bonding) or steric (for thermal ﬂuctuations) interactions, according to Leckband and Israelachvili (2001). So-called speciﬁc interactions (e.g., ligand–receptor interactions) are typically a result of unique combinations of these four “non-speciﬁc” physical forces (Israelachvili, 1992).

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

11

come together relative to surfaces that are pulled apart. As a ﬁnal point to this section, it should be noted that a force of interaction is related to energy (E) according to F ¼ 2 dE/dD.

III. FORCE CURVE THEORY AND COLLECTING FORCE DATA Force measurements attempt to capture interactions representing the electrostatic interplay between single molecules and atoms that are bound to a solid surface or exist as components of a solvated environment. Given the extraordinarily small dimensional (nanometer to angstrom) scale over which they operate, many challenges exist in capturing molecular level forces. This section reviews how force microscopy (or atomic force microscopy, AFM; also know as scanning probe microscopy) addresses these challenges, describes its operation and assesses how accurately the interactions are captured. Highlighted are some of the basic assumptions associated with force microscopy, while noting some of its advantages and limitations. An AFM force probe consists of a tip attached to a ﬂexible cantilever, which is modeled mechanically as a single harmonic oscillator. Forces exerted on the tip are registered as a spring-like deﬂection in the cantilever, which may be recorded with various detection systems, including electron tunneling (Binnig et al., 1986), interferomotery (Erlandsson et al., 1988; Rugar et al., 1989) and capacitance (Goddenhenrich et al., 1990). The following summarizes an optical lever collection system (Meyer and Amer, 1990) that is most commonly found in commercially available AFMs, including the widely used Veeco/Digital Instrument system. Here the deﬂection is typically recorded as a change in voltage resulting from the displacement of a laser spot that is reﬂected off the top of the cantilever and into a photodiode. Voltage (V) is translated into cantilever deﬂection (nm) using a detector sensitivity value (V nm21) that is equal to the slope of the line when the tip and sample are in contact (see region of contact in Fig. 2). Provided the sample stiffness is signiﬁcantly higher than the cantilever (which is the case when probing mineral surfaces), there should be a 1:1 correlation between piezo movement and cantilever deﬂection once the (V to nm) conversion is made. Small deviations from an absolute slope equal to one may be an indication of detector drift, and can be corrected by dividing the deﬂection values by the slope (H. Skulason, personal communication). A slope less than one may also be an indication of a sample compliance that is less than the cantilever (which may be the case when making measurements on a cell), in which case alternative sensitivity determinations, such as the photodiode shift voltage method may be employed (D’Costa and Hoh, 1995; Lower et al., 2001b).

12

T. A. KENDALL AND S. K. LOWER

Figure 2 A typical force versus piezo movement plot showing three general regions — contact, interaction, and no contact. For clarity a single trace is shown (e.g. an approach curve); however, force plots with both approach and retraction traces are also common. In the region of no contact the tip and sample are separated at distances large enough that no interaction occurs. Hysteresis between the approach and retraction curve in the region of no contact may be a function of solution viscosity, or inelastic deformation of the cantilever. As the piezo advances the sample closer, the tip begins to “feel” the surface. In the example plot we see an initial repulsion followed by an attraction recorded as a sharp jump to contact that generates a minimum in the curve. Once in contact, the slope trace is typically constant as the cantilever is moving with the piezo. Information from this region may be used to determine detector sensitivity or elastic properties of the sample or tip.

Hooke’s law, F ¼ 2 ksd, then allows conversion of cantilever deﬂection, d, into force, F, using the spring constant of the cantilever, ks. Note that sign convention dictates that negative forces reﬂect attractive interactions and positive forces are repulsive.

A. FORCE – DISTANCE CURVES, CAPTURING A POTENTIAL FORCE VERSUS SEPARATION PLOT Figure 2 shows a typical plot of force versus piezo movement. Note the x-axis represents relative piezo movement or an indexing of a sample’s position relative to the cantilever (tip). It does not reﬂect tip-sample separation (discussed below). Three main components of the plot are identiﬁed: the regions of no contact, interaction and contact. Several sub-features are contained within each region including oscillations, subtle slope changes, linear and non-linear extensions, jumps to and from contact (Ducker et al., 1992; Cappella and Dietler, 1999; Gergely et al., 2001), which, in addition to providing reference points to register the force curve to an origin (discussed below), contain valuable information on the interaction between the tip and the surface, the nature of the intervening solution, tribology, adhesion, and elastic properties of the system.

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

13

The focus is now turned to the region of interaction, which is of primary interest when studying intermolecular forces at cell or biomolecule –mineral interfaces. Here a wealth of information on the charge character of a mineral surface or biomolecule; the nature and contour length of a polymer extending from a bacterium; DLVO forces (see below) and Debye lengths associated with a colloidal particle or cell; and the energy landscape and activation barriers of a bond are found. But ﬁrst, to draw both qualitative and quantitative conclusions from forces of interaction, it is imperative to have an understanding of the mechanical constraints of what is recorded in this region using force microscopy using an AFM. To illustrate this we show a simple, short-range interaction potential for atomic scale particles (note, Part II concentrated mainly on larger particles and/or surfaces) described by the Lennard – Jones equation: EðDÞ ¼ 2A=D6 þ B=D12

ð5Þ

Energy, E, has an inverse power law dependence on distance, D, with the 2 1/D 6 term representing the attractive component of the van der Waals force. The absolute value of this term is maximized at a distance De where the ﬂuctuations in charge density coincide to result in a potential well. At separations less than De, the potential rises rapidly with distance, 1/D 12, as the interaction is repulsive in nature due to electronic overlap and nuclear interaction (Israelachvili, 1992; Cygan, 2001). Force microscopy (or AFM), however, does not record energy values directly, but instead measures force. To compare the Lennard– Jones potential with an AFM data set, we take its derivative, such that graphing the relationship dE=dD ¼ FðDÞ ¼ 26A=D7 þ 12B=D13

ð6Þ

produces a theoretical force –separation distance curve similar to the one in Fig. 3a. To further facilitate comparison with the theoretical curve, an origin is deﬁned for the force microscopy data set as follows. A force equal to zero can be deﬁned as the average force value within the region of no contact, while the point at which the tip and sample come into (for approach) and out (for retraction) of contact can be deﬁned as the zero point on the x-axis. Determining the point of contact is clear when a distinct attractive or adhesive component (e.g., a jump to contact) is present, but ambiguous when such features are absent. In the latter case, the intersection between the slope of the region of no contact and constant compliance can be used as a guide (Cappella and Dietler, 1999). In a ﬁnal important step, the xaxis in the force microscopy data set is adjusted to reﬂect tip sample separation instead of piezo movement, by adding the cantilever deﬂection values to the piezo movement distances (Ducker et al., 1992; Butt et al., 1995). Here the selection of the sign convention for the forces becomes intuitive. Addition of positive repulsive deﬂections to the piezo movement results in larger tip-sample separation, while adding negative attractive deﬂections result in a decrease in the separation. Unlike the Lennard – Jones curve, note that the values in Fig. 3b, left of the point of contact

14

T. A. KENDALL AND S. K. LOWER

are essentially meaningless in terms of interaction force because the tip and sample are in direct contact. The end result is a force versus tip-sample separation plot with a region of interaction that can be compared to the theoretical curve (see Fig. 3a and b; also see Section V, below). Two main differences exist between the force microscopy data and the potential: (1) the slope of the attractive component of each curve, and (2) the hysteresis that exists between the approach and retraction forces in the force microscopy plot. With the force microscope, it is not uncommon to record

Figure 3 (a) Differentiated Lennard– Jones potential provided as an example interaction to be captured with force microscopy or AFM. During AFM operation the forces associated with the potential are recorded as deﬂections in the cantilever. If the force gradient (tangent to the solid trace) exceeds the spring constant, ks, the cantilever becomes mechanically unstable and will jump along a slope equal to kg (dashed line). (b) Force– tip sample separation curve showing jumps to and from contact along slope ¼ kg. Unlike Figure 2, the x-axis represents separation distance between the tip and the sample. Here, both the approach (open circles) and the retraction (closed circles) traces are shown. Note that the hysteresis between the two traces is absent in the Lennard–Jones curve where the solid line represents both approach and retraction forces. Points to the left of zero separation (i.e., lowest most point on the approach or retraction curves) represent movement of the piezo while the tip and sample are in contact. (c) Increasing the spring constant (e.g., using a stiffer cantilever) from ksA to ksB will capture more of the potential (region A2 –B2), however force resolution is lost and smaller magnitude forces will go undetected.

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

15

an attractive force as a characteristic jump to contact on approach. These “jumps” represent mechanical instability in the cantilever due to a force gradient that exceeds its spring constant, ks. Clearly, interaction information is lost as the cantilever encounters a force gradient (tangent to the theoretical curve) at point A1 that is greater than its stiffness and consequently jumps to point A2 along a slope equal to ksA (see Fig. 3c). More of the attractive potential can be sampled with a stiffer cantilever (e.g., ksB), however, force resolution is lost, and the region along the theoretical curve between B2 and A1/B1 remains unsampled. A similar situation may be encountered upon retraction, which, in part, contributes to the hysteresis observed in the force data. Speciﬁcally, cantilevers with smaller spring constants generate larger amounts of hysteresis. However, hysteresis between approach and retraction curves is also due to the formation of adhesive bonds once surfaces are in contact. This is common for soft samples such as biological cells (see below). For some investigations excessive hysteresis is undesirable and several techniques have been developed to reduce it thereby recovering the “lost” information (i.e., region A2 – B2). These methods employ an opposing force that is external to the system in an attempt to increase the effective stiffness of the cantilever, while retaining force resolution. Electrostatic force (Joyce and Houston, 1991), magnetic feedback (Jarvis et al., 1996; Yamamoto et al., 1997; Jarvis et al., 1998; Ashby et al., 2000) and radiation pressure from a laser (Aoki et al., 1997; Tokunaga et al., 1997) have all been used to supply the steadying force to the cantilever.

B. HYSTERESIS Certainly other sources besides the instability of the cantilever contribute to approach –retraction hysteresis. In the theoretical Lennard –Jones relationship given as a potential example, no adhesive reaction between the tip and sample is modeled and the retraction curve retraces the approach curve (Fig. 3a). However, this is not an appropriate model for soft biological cells, which have biopolymers, designed over millions of years of evolutionary selection, for the express purpose of adhesion. When making force microscopy measurements, the tip comes into contact with the surface allowing for reaction and deformation between the two. The bonds and coordinations that result can then be explored and characterized using the associated adhesion forces and approach-retraction hysteresis (Burnham et al., 1990; Cappella et al., 1997). In some systems, the number of bonds that form (and, thus the level of hysteresis) is correlated with the amount of pressure that is applied on the sample by the tip (Weisenhorn et al., 1992). Speciﬁcally, increased pressure leads to sample and tip deformation resulting in increased contact area (Israelachvili, 1992; Cappella et al., 1997), and in the case of functionalized tips (e.g., those coated with self-assembling organic monolayers), a possible rearrangement of functional groups terminating the monolayer.

16

T. A. KENDALL AND S. K. LOWER

Both phenomena facilitate additional bonding, larger hysteresis and higher adhesion values, as documented by several workers (Weisenhorn et al., 1992; Hutter and Bechhoefer, 1993; Toikka et al., 1996; Ashby et al., 2000). Therefore, the amount of indentation must be carefully documented to facilitate the comparison of adhesion data from one study to another. One way of controlling the amount of indentation using commercially available AFMs is by adjusting the scan start position, setpoint or the trigger settings. Varying these parameters can be especially useful when probing many biological systems, where pressure and/ or contact time may be a natural mechanism of inducing adhesion (e.g., see Leckband and Israelachvili, 2001; Lower et al., 2001a).

C. TIP SHAPE Tip shape is a critical AFM parameter that can dictate the force values and contact geometry between the sample and force probe (Hartmann, 1991; Butt et al., 1995). Constraining this value is essential if experimental force traces are going to be compared theoretical models such as “DLVO” (see Section V below). Yet, tip shape can be difﬁcult to determine, in part due to the surface roughness, irregularities and asperities that are associated with traditional silicon or silicon nitride tips (Cappella et al., 1997). Moreover, tip shape can change over time as continued use promotes wear (Cappella et al., 1997). Solutions to this problem include careful, periodic characterization of the tip with electron microscopy, better constraint of tip geometry by attaching a spherical colloidal probe (Ducker et al., 1991; Butt et al., 1995), or, as described in more recent work, by attaching a carbon nanotube (Wong et al., 1998a; Hafner et al., 1999; Cheung et al., 2000).

D. SPRING CONSTANT DETERMINATION If a quantitative analysis of absolute force values is desired, determination of the spring constant (ks) is critical and nominal values provided by the manufacturer generally cannot be relied upon (Lower et al., 2001b). Many factors affect the spring constant including primary characteristics such as cantilever dimensions, geometry and substrate material; as well as, additional modiﬁcations common in force spectroscopy such as gold coating, the addition of organic monolayers, the attachment of colloidal spheres or cells, and even ion adsorption (Sader et al., 1995; Craig and Neto, 2001; Cherian and Thundat, 2002). As a result, a large body of literature detailing several methods of directly determining ks exists. A procedure commonly used because of its simplicity, non-destructive nature, and applicability to common cantilever geometries (e.g., V-shaped, rectangular) is provided by Cleveland et al. (1993). This method

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

17

derives the spring constant by measuring changes in the cantilever’s resonance frequency after small masses (e.g. W microspheres) are loaded onto the end of the tip. The Cleveland method is further optimized when corrected for errors introduced by off-end loading of the mass (Sader et al., 1995). More recent methods measure hydrodynamic drag of the cantilever through a ﬂuid of known viscosity to determine ks for bare rectangular cantilevers (Sader, 1998; Sader et al., 1999; Maeda and Senden, 2000). This concept has also been applied to determine ks for cantilevers activated with a colloidal probe (i.e., a silica or polystyrene microsphere) (Craig and Neto, 2001). In the latter method, it is useful to directly measure ks for a cantilever with an attached sphere because it accounts for changes in the spring constant due to the shifts in the point of load associated with the position of the colloid sphere and the change in the stiffness associated with the adhesive used for microsphere attachment. Other methods measure ks using thermal oscillations and the equipartition theorem (Hutter and Bechhoefer, 1993; Butt and Jaschke, 1995), a ﬁnite element analysis of the static deﬂection of a cantilever for which the Young’s modulus is known (Sader and White, 1993); unloaded resonant frequency of a cantilever of known mass (which is commonly not the case) (Sader et al., 1995), radiation pressure from an acoustic transducer (Degertekin et al., 2001); microscopic and macroscopic reference cantilevers of known stiffness (Rabinovich and Yoon, 1994; Torii et al., 1996; Jericho, 2002); and the change in resonant frequency due to gold coating (Gibson et al., 2001).

E. ARTIFACTS

IN

FORCE MEASUREMENTS

Several artifacts can arise during force measurements with the AFM. The inverse path effect represented as an upward, hysteretic shift in the retraction trace in the region of contact arises from nonlinearities of the piezoelectric actuator that positions the sample (or tip) (Cappella et al., 1997; Cappella and Dietler, 1999; Heinz and Hoh, 1999). A shift in the contact portion of the retraction trace such that it is parallel with the extension trace reﬂects friction as the tip plows or slides along the surface (Heinz and Hoh, 1999). A sinusoidal oscillation in the region of no contact may also be present, representing the interference of stray laser light bouncing off the sample and interfering with the laser light reﬂected off the top of the cantilever (Weisenhorn et al., 1992; Cappella et al., 1997). This oscillation can be distinguished from other artifacts, such as noise due to mechanical vibrations, because its wavelength should roughly be equal to , l/2n, where l is the wavelength of the laser source and n is the index of refraction of the ﬂuid between the tip and sample (Weisenhorn et al., 1992; Craig and Neto, 2001). Thermal oscillation in the region of contact can be recognized by deﬂection ﬂuctuations whose standard deviation is roughly equal to (kskBT)0.5; ks, the spring constant; kB, Boltzmann’s constant and T, temperature (Gergely et al., 2001). Operational artifacts may include a large slope that is

18

T. A. KENDALL AND S. K. LOWER

present in the region of no contact, common when making measurements in a ﬂuid cell. The origin of this slope is unclear but can often be remedied by eliminating air bubbles in the system or insuring a ﬂat, even orientation of the gasket used to seal the system. Large plateaus at the extremities of both traces often represent a saturation of the detector, requiring an adjustment of the scale of the plot, the deﬂection limit of the detector or the starting position of the scan.

F. DATA PROCESSING

AND

STATISTICS

With standard AFMs, the one-click ease with which a single force curve is collected allows hundreds of curves to be recorded at a sample point in only a few minutes. Considering the fact that a typical curve can contain 2048 data points, a single experiment can produce an enormous volume of data. Further, the variability between force curves collected at a single location can often be quite high. This raises several questions regarding data processing and interpretation that are often neglected. What is the most efﬁcient way to process these data? What is the minimum number of curves necessary to characterize each sample point or a particular interaction? What level of error and variability is associated with the force measurements? How is force data distributed about its mean? What measurable parameters or features of a force curve are the most important in characterizing the interactions (e.g., adhesion force)? What is the best way to identify trends or correlations in these parameters? Clearly, answering these questions requires statistical techniques, tests and models that determine appropriate, signiﬁcant average values of force curve parameters and facilitate the identiﬁcation of meaningful force curve features. This process begins by collecting summary statistics for each data set, including calculation of means, standard deviations, error values (e.g. conﬁdence limits) and by plotting histograms for multiple parameters derived from the curves, including adhesion force and distance of jump to contact. When comparing parameters from the curves, statistical tests (e.g. ANOVA, t-tests; correlative tests) may be performed using a standard statistical and data processing package (e.g., Igo Pro, Wavemetrics, Inc.). Simple regression models may also be employed to determine important variables that contribute to the shape of a force curve. To this end, a routine has been written (Kendall and Hochella, 2003) that rapidly processes force curve data to produce plots of force versus piezo movement and force versus tip sample separation using the procedures discussed above. Automated parameter determination, statistical calculations, whole force curve averaging, autocorrelation calculations (for identifying quantized force values) and histogram generation are incorporated in this customized routine written using Igor Pro’s internal programming environment. The simple parameter extraction module quickly and consistently identiﬁes features and selects the values using basic, objective criteria such as

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

19

maxima thresholds in the differentiated force data and tolerance limits for speciﬁc changes in slope (see Fig. 4).

G. ADVANCED ALGORITHMS These criteria are appropriate when the bond ruptures and snaps to contact are large and/or distinct. However, other more advanced algorithms (Baumgartner et al., 2000; Kasas et al., 2000; Gergely et al., 2001) are required when features are small, numerous, less distinct (e.g., multiple ligand –receptor interactions) and/or have the potential to be masked by vibrational and thermal noise. For example Kasas et al. (2000) employ a fuzzy logic algorithm that enables automated discrimination of speciﬁc, signiﬁcant adhesions in a retraction curve that might otherwise be overlooked. The routine assigns a grade to each potential rupture event, ranking it somewhere between non-speciﬁc (0) and speciﬁc (1). Assignment of the grade relies on a priori knowledge of the interaction event morphology, and uses criteria such as the angle between the jump and the background trace, or whether or not the jump is U-shaped or V-shaped. This means that the procedure is operationally deﬁned and ﬁrst has to be “taught” what the features of interest look like in order to calibrate it to the system/features being studied. Gergely et al. (2001) present an algorithm that identiﬁes ruptures based on a comparison of the minima with neighboring peaks. Selection is controlled by adjusting an appropriate noise level, m, such that the difference between a feature and its nearest neighbors must be greater than 2m times the standard deviation of the force values. Additional smoothing of the force curve is also achieved by ﬁtting a second order polynomial to a designated amount, p, of consecutive points. Using this routine, forces measured between human ﬁbrinogen interacting with a silica surface were processed. By monitoring histograms of inter-rupture distances selected with successively more rigorous (higher) m values, the authors were able to detect a signiﬁcant peak at 20 – 25 nm, a value that corresponds nicely to the known spacing between two domains in the protein.

H. RELATING BOND CHEMISTRY AND ENERGIES TO FORCE MEASUREMENTS Force microscopy measurements intuitively have the potential to describe energies, E(D), associated with an interaction at a small separation, D, by R integrating force over distance, E(D) ¼ F dD. As discussed above, however, differences in spring constants can produce variable hysteresis and, therefore, can lead to drastically different energy values. Without ﬁne control of the effective spring constant, it is difﬁcult to accurately capture a potential in a quantitative fashion, which is critical for single molecular work. Moreover, if reaction occurs

20

T. A. KENDALL AND S. K. LOWER

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

21

upon contact (and provided single interactions can be identiﬁed in the force spectra); simple prediction of bond/interaction energies based on rupture forces is non-trivial. Speciﬁcally, it might be postulated that the maximum gradient in the potential, [dE/dD ]max, is equal to the adhesion or rupture force from the retraction trace; however, in a seminar paper, Evans and Ritchie (1997) showed that such a simple correlation is not valid for single molecule interactions, and more sophisticated theory is required for quantitative comment on the absolute energetics of a bond using force data. Before continuing, it must be noted that these ﬁndings do not preclude valuable quantitative and qualitative comparison of force measurements and bond ruptures to energy parameters. Indeed, early force experiments with various ligand receptors (e.g., biotin, iminobiotin, avidin, streptavidin combinations) revealed a correlation between the rupture forces and enthalpy values associated with each complex (Moy et al., 1994a). This information together with the lack of correlation between the rupture forces and total free energy suggested the unbinding was adiabatic and that any entropic contributions to the system (e.g., solvation forces) occurred outside of the binding pocket, and were not recorded with the AFM. Other studies followed, relating thermodynamic parameters to interaction forces (Chilkoti et al., 1995), as well as many force experiments that employed “elementary” or averaged rupture forces to compare two or more systems in a relative fashion (Florin et al., 1994; Frisbie et al., 1994; Dammer et al., 1996; Noy et al., 1997; Ito et al., 1999; Schmitt et al., 2000; Fiorini et al., 2001; Lower et al., 2001a; Kreller et al., 2002; Kendall and Hochella, 2003;). The true value of these studies is their relative quantitative and qualitative comparisons of force data. These characterize the nature of forces at an interface, demonstrate surface and molecule recognition, and deﬁne relative afﬁnity between two molecules or between a molecule and a surface. However, Evans demonstrated that these rupture forces, as absolute values, represent one point in a continuum of bond strengths (Merkel et al., 1999); and that the detachment force recorded with the AFM (and other force measuring techniques) is not a singular fundamental

Figure 4 Screen shot of one module (Sensitivity Tweaks) in the force curve processing routine AFM 4.4 written in Igor Pro, 4.04, WaveMetrics, Inc. (Kendall and Hochella, 2003; some of the base code was provided by H. Skulason). It is designed primarily for handling force data produced Digital Instrument’s Nanoscope IIIa MultiMode system. The Sensitivity Tweaks module is designed to rapidly review and assess how well the normalizaiton routine automatically registers and normalizes force data to an origin. The normalization procedure includes the identiﬁcation of a baseline in the region of no contact, calculating the detector sensitivity from the region of constant compliance and detecting when the tip and the sample are in contact. The latter is determined using peaks in the differentiated force wave that are selected based on threshold/sensitivity settings shown in the panel in the lower right. If initial normalization is unsatisfactory, these settings may be optimized and an autonormalization may be run again; or features can be identiﬁed manually.

22

T. A. KENDALL AND S. K. LOWER

property of the molecular interaction being probed (Evans and Ritchie, 1997; Evans, 1998; Merkel et al., 1999). Instead, apparent bond strength as estimated by rupture force is a function of the loading rate (Evans and Ritchie, 1997). This relationship represents a reﬁnement of a model proposed by Bell (1978) that predicts an exponential ampliﬁcation of dissociation kinetics in the presence of an external force (Merkel et al., 1999). Dissociation of relatively weak associations can be conceptualized as a particle moving out of a potential well (bond), over single (simple interaction) or multiple (complex interaction) activation barriers representing transition states (Fig. 5). Under a zero force condition the particle will migrate out of the well, through the transition states, and ultimately to complete dissociation on a time scale that is dictated by thermal agitation (kBT). A constant external force on the bond, however, expedites the thermally mediated kinetics and decreases the lifetime of the bond by lowering the activation barriers in the energy landscape along a projection that is proportional to the amount of force (Fig. 5) (Evans and Ritchie, 1997). Under a dynamic load (e.g. a retracting cantilever) where the force, F, increases over time, t, as loading rate, Rf ¼ dF/dt ¼ ksvc (ks is the spring constant of the system and vc is the velocity of the cantilever), inner activation barriers are revealed as outer activation barriers are progressively lowered by the accumulating force. This

Figure 5 The effect of an external force on the energy landscape of a bond. (Modiﬁed from Evans and Ritchie, 1997; Merkel et al., 1999). The minimum of the traces represents a bond or potential well found along a reaction coordinate, x. Two activation barriers (local maxima) exist representing transition states that a system must go through during dissociation of the bond. External force, F, is represented as a mechanical potential, 2(F cos q)x oriented at an angle u to the reaction coordinate. Increasing force lowers outer activation barriers to reveal the inner maxima. Eventually all barriers are lowered allowing free diffusion from the initial minimum/bond (Evans and Ritchie, 1997).

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

23

phenomenon leads to an intriguing positive correlation between the rupture force, Fr, and loge(Rf), that is best conceptualized in terms of thermally mediated nature of the bond rupture kinetics. At small loading rates, activation barriers are lowered at a rate slow enough for thermal contributions from the medium to be effective in helping the molecule diffuse out of the well and over the barrier before higher forces are reached (Gergely et al., 2002). Thus, with an effective thermal contribution, a lower rupture force is recorded at the time of dissociation. Under large loading rates (e.g., those typical of many AFM experiments), activation barriers are lowered fast enough such that dissociation proceeds with minimal thermal contribution, resulting in a higher rupture force (Gergely et al., 2002). At ultrafast loading rates on the time scale typical of molecular dynamic (MD) simulations (1012 s), the entire bonding potential is compromised quick enough that only frictional drag is recorded as the molecule traverses a completely “stretched” and coarsened energy landscape (Evans and Ritchie, 1997). Here, loading rates commonly exceed the time scale of unencumbered, diffusive passage of a molecule from its bonded state, leaving the complex kinetically trapped as force continues to rise. This was observed during molecular dynamic simulations of the biotin – streptavidin complex (Grubmuller et al., 1996). For the biotin system, the time for diffusive passage, tD (e.g., the lifetime of the bond) is estimated to be 500 ps under a constant force of 280 pN (Evans and Ritchie, 1997). All activation barriers are lowered at this force such that the initial minimum (e.g., the original potential well representing the bond) is exposed allowing direct diffusion out of the well. However, the ultrafast molecular dynamic simulation loading rate (1.3 £ 1012 pN s21) exceeds 280 pN/ tD, therefore, leaving the complex kinetically trapped in the bound state as rupture force rises well above 280 pN (Evans and Ritchie, 1997). With these observations and extensive theory development, Evans recognized that measurements of rupture forces over a large range of loading rates effectively probes the lifetime of the interaction under different levels of force while mapping out energy barrier position and heights in a technique now known as dynamic force spectroscopy (Evans, 1998). A dynamic force spectra is constructed by plotting the most probable rupture force of a single interaction, Fr, versus loge[Rf], where Rf values span several orders of magnitude. Regions of constant slope deﬁned as fb ¼ kBT/xb represent activation barriers at a distance of xb from the potential well. Barrier heights Eb can be derived from the intercept of the slopes at zero force deﬁned as (Evans and Ritchie, 1997): loge ðRof Þ ¼ 2Eb =kB T þ loge ð fb =tD Þ

ð7Þ

where, again tD is the time of diffusive passage, and thus 1/tD represents an attempt frequency. The attempt frequency is generally not known, but can be estimated from the damping phenomenon (Evans, 1998). Activation barrier positions derived from experimental dynamic force spectra of the biotin –avidin

24

T. A. KENDALL AND S. K. LOWER

interaction compared well with barriers predicted by molecular dynamic simulations (Evans and Ritchie, 1997; Izrailev et al., 1997), further emphasizing the value of coupling force measurements with computer simulations. However, it must be noted that bridging the gap between the orders of magnitude in the loading rate of experimental systems versus loading rates used in computer simulations is not straightforward and extrapolation of calculated rupture forces to experimentally determined forces must be done with caution (Grubmuller et al., 1996; Izrailev et al., 1997; Wong et al., 1999). While this was attempted by Grubmuller (1996), Izrailev (1997) and Evans (1997, 1998) indicate that the real value of molecular dynamic simulations, in this context, is their potential to provide clues as to which structural determinants of the interactions contribute to the activation barriers, thereby providing a qualitative mechanism to account for dynamic force spectra features. For example, Izrailev et al. (1997) used a collection of molecular dynamic simulation packages to demonstrate a ground state avidin – biotin complex that was stabilized by hydrogen bonds between the biotin head group and polar amino acids (e.g., Tyr33) within the binding pocket. With the application of an external force, two intermediate states stabilized by H-bonds with amino acids at different positions in and near the pocket (e.g., Ser16 and Arg114) are revealed as the ligand is removed — an observation that was consistent with dynamic force spectroscopy experiments. Lo et al. (2002) also used a variation of the Bell model to explore the relationship between rupture force and ambient temperature in the biotin – avidin system. The experiments were conducted on an AFM with a constant, millisecond time scale loading rate that was slow enough, compared to the nano to picosecond time scale of molecular dynamic simulations, to neglect any frictional energy loss due to viscous drag. The slow loading rates also allowed the key assumption that thermal equilibrium is achieved at any moment during the unbinding process. This validated the use of a Maxwell –Boltzmann energy distribution to describe the thermal energy being supplied to the complex. The end result is a relationship (Lo et al., 2002): t Fi2 ¼ 2DEC kbond 2 2kB Tkbond loge R ð8Þ tD that can be ﬁtted to an experimental AFM data set of adhesion forces (Fi), to derive bond stiffness (kbond), and critical binding energy (DEC). Both the derived values reﬂect a summation of the different types of forces that make up the biotin – avidin interaction (e.g., H-bonds, van der Waals and polar interactions). Other variables include thermal energy (kBT), and the ratio of the rupture time (tR, determined from the AFM data) and time required for the ligand to diffuse out of the binding pocket (tD, estimated independently). Critical binding energy, DEC, may be related to a dissociation energy (De) by deﬁning a potential to describe the interaction — in this case a Morse potential was used. Their De value based on force measurements, 28.4 kcal mol21, compared favorably with

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

25

the enthalpy change of the dissociation determined by independent means (DH ¼ 23.4 kcal/mol) (Swamy, 1995; Lo et al., 2002). Moreover, the enthalpy value was combined with bond stiffness in additional calculations to determine a critical displacement magnitude (0.1 nm) that was close to inner barrier position, xb, value determined with dynamic force spectroscopy (xb ¼ 0.12 nm). Both temperature dependent and load dependent dynamic force spectroscopy rely on force measurements of the interaction of a single ligand – receptor pair. This is accomplished by reducing the density of the sites that are present and available for bonding (similar to protocols outlined by Florin et al. (1994), Rief et al. (1997a, b), Marszalek et al. (1998) and Grandbois et al. (1999)), such that 1 in 7 –10 touches results in attachment (Evans, 1998). Governed by Poisson statistics (Evans and Ipsen, 1991; Williams et al., 1996), 90– 95% of the attachments are predicted to be single bonds. In addition to the biotin –avidin linkage, Poisson distributions are common in force measurements associated with several other systems (Han et al., 1995; Williams et al., 1996; Wenzler et al., 1997; Lo et al., 1999; Stevens et al., 1999). The probability, P(n), of an attachment representing n linkages follows a deﬁned, Poisson distribution, making it possible to extract the n ¼ 1 case from a large number of rupture force measurements (Lo et al., 2002). Feedback mechanisms are also employed to control impingement on the sample thereby insuring each approach and retraction cycle has the same magnitude and history of contact force (Evans, 2001). This is especially important when making measurements with biomolecules secured to monolayers that can easily deform on contact to produce various contact areas and conﬁgurations, and ultimately different numbers of attachment (Evans, 2001). Although successful dynamic force spectroscopy experiments have been carried out on a single force measuring instrument (AFM) with a range of loading rates from 100 –5000 pN s21 (Yuan, 2000), due to the exponential relationship between kinetic rates and barrier energies (discussed above), dynamic force spectroscopy is optimized when collecting force measurements over a range of loading rates that are different by orders of magnitude. This can require the use of several force measuring techniques including laser/optical tweezers for slow loading rates 1– 10 pN s21, a biomembrane force probe (BFP) for intermediate rates (10 – 1000 pN s21) and AFM for fast loading rates (104 –106 pN s21) (Evans and Ritchie, 1997).

I. RELEVANCE OF DYNAMIC FORCE SPECTROSCOPY TO BIOLOGICAL – INORGANIC INTERFACE Techniques to extract energy information from force spectroscopy were developed primarily using the biotin –(strept)avidin system (Moy et al., 1994a; Evans and Ritchie, 1997) due to its relevance to biological systems, non-covalent nature, high afﬁnity and extensive history of experimentation and study. Since

26

T. A. KENDALL AND S. K. LOWER

then, dynamic force spectroscopy has been applied to other systems, primarily non-covalent and biological in nature. These include DNA base pair interactions (Strunz et al., 1999), unfolding of muscle protein domains (Rief et al., 1997a), antibody – antigen interactions (Schwesinger et al., 2000) and lipid anchoring in membranes (Evans and Ludwig, 2000) to name a few. Grandbois et al. (1999) also made an attempt at using the dynamic force spectroscopy concept to measure the strength of a single covalent bond. Although they produced values for only one loading rate due to the difﬁculty of collecting individual covalent interactions (H. Gaub, personal communication). Their calculation of rupture force probabilities based on dynamic force spectroscopy methods allowed them to identify the covalent attachment being terminated as a Si –C bond. Applying dynamic force spectroscopy concepts to AFM data collected on environmental systems has great potential to provide new insight on the interaction energetics and bond chemistry associated with biogeochemical interfaces. This is, in part, because experimental data collected at the molecular level to describe surface reactions between single biomolecules and mineral surfaces is lacking. Traditionally conﬁned to computer simulation (Cygan, 2001), dynamic force spectroscopy now affords a unique, direct examination of energy landscapes associated with some of the non-covalent mechanisms (e.g., H-bonding) assumed to initiate sorption reactions between minerals and ligands (Holmen and Casey, 1996), the possible ionic or covalent binding of a metal in a mineral surface associated with dissolution (Stumm, 1992), reversible and non-reversible adhesion states of colloids or cells to a surface (Absolom et al., 1983; Ryan and Elimelech, 1996; Ryan and Gschwend, 1994), mineral and or metal recognition of a mineral structure by membrane bound proteins (Lower et al., 2001a). Perhaps, the true advantage of using dynamic force spectroscopy is realized when used in a comparative framework, for example, dynamic force spectra of cell or biomolecule– mineral interaction before and after structural and functional changes in either (1) the cell surface (e.g., via altered gene expression due to imposed environmental conditions) or the biomolecule (e.g., via point mutations in proteins, functional group substitutions/inactivation in ligands) or (2) the mineral via metal substitution or by comparing isostructural mineral equivalents or different crystal growth faces. Changes in slopes of the dynamic force spectra resulting from structural modiﬁcations can provide clues as to which proteins, functional groups or even crystallographic constraints contribute to surface complex stability or speciﬁc activation barriers to binding or detachment. Concomitant correlation of force values with independently determined thermodynamic parameters can also provide insight as to whether a surface attachment or detachment or metal extraction is enthalpy driven or entropically dominated. And as seen above, a common theme when using dynamic force spectroscopy is to supplement and validate characteristics of a particular energy landscape with mechanistic information derived from computer simulations.

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

27

The role of these simulations is anticipated to be just as important when applying these techniques to biogeochemical systems.

IV. FORCES AT THE BIOMOLECULE–MINERAL INTERFACE Organic ligands produced by microorganisms such as low molecular weight organic acids and siderophores have the potential to greatly impact the geochemistry and ecology of soil environments (Stone, 1997; Hersman, 2000). These ligands interact with mineral surfaces to form a critical interface that has implications on biological nutrient/metals acquisition, control of metal toxicity or even ecological competition (Bossier et al., 1988; Stone, 1997; Neubauer et al., 2000; Brantley et al., 2001; Kraemer et al., 1999, 2002). Ligands enter into sorption and desorption reactions with minerals that enhance dissolution or surface passivation, mediate contaminant mobility, or alter the charge character of the mineral surface (Barker et al., 1997; Stumm, 1992). As a result this interface has been studied extensively with bulk experiments and sorption studies (e.g., Kummert and Stumm, 1980; Ludwig et al., 1995; Yao and Yeh, 1996), and with surface sensitive techniques such as XPS (Kalinowski et al., 2000), and Fourier Transform Infrared (FTIR) Spectroscopy (Hansen et al., 1995; Holmen et al., 1997). Key to sorption and desorption reactions between ligands and minerals, however, are the forces that bring the ligand into and out of contact with the surface. Such forces are dependent on the charge character, structure and reactivity of the ligand, the mineral surface and the intervening solution. Characterization of these forces using force microscopy holds great potential to complement information from the existing methodologies listed above in addition to providing new insight on how ligands interact and coordinate with mineral bound metals. In pioneering work, activation of an AFM tip with a speciﬁed chemistry was carried out to examine the biotin –avidin interaction (Florin et al., 1994; Lee et al., 1994b). Florin et al. (1994) sorbed biotin (an organic ligand) to an AFM tip, and probed a surface coated with the protein receptor avidin. Force measurements of this high afﬁnity ligand – receptor system showed a positive correlation between the elementary quantized adhesion forces detected with an autocorrelation analysis and the thermodynamic binding afﬁnities. Speciﬁcally, biotin adhesion to the avidin substrate measured 160 pN, while iminobiotin, which contains a nitrogen substitution in place of an oxygen and has a lower binding afﬁnity, exhibited adhesions closer to 85 pN. Several force spectroscopy studies of the biotin system followed (Moy et al., 1994a, b; Chilkoti et al., 1995; Wong et al., 1999; Lo et al., 2002), along with other force investigations of biomolecules, including examination of: interactions

28

T. A. KENDALL AND S. K. LOWER

between antibodies and antigens (Dammer et al., 1996; Hinterdorfer et al., 1996; Schwesinger et al., 2000); enzyme activity (Fiorini et al., 2001); proteoglycans (Dammer et al., 1995); observations on the stretching of polysaccharides (Rief et al., 1997b; Marszalek et al., 1998) and muscle proteins (Rief et al., 1997a); and the hybridization of oligonucleotides (Lee et al., 1994a; Mazzola et al., 1999). Simple functional groups have also been covalently attached to AFM tips in order to explore more fundamental interactions, such as the forces between methyl, carboxyl or methyl –carboxyl pairs (Frisbie et al., 1994; Noy et al., 1997). Speciﬁcally, this technique, termed chemical force microscopy (CFM), was used to identify the nature of the interacting force (H bond, van der Waals, electrostatic), characterize surface energies and charge distributions, and generate force maps that showed the spatial arrangement of simple functional groups or hydrophobic regions on a monolayer or surface, sometimes with nanometer resolution (Noy et al., 1997). Collectively, these studies provide a foundation, which allows the application of force spectroscopy to additional, more complex, natural systems, such as the ligand/biomolecule– mineral interface that is characteristic of soil environments. Indeed the same forces (e.g., H-bonding, hydrophobic/hydrophilic forces, the van der Waals force, steric forces, non-speciﬁc and speciﬁc interactions) that allow molecular recognition between biomolecules are also present in ligand mineral interaction (Israelachvili, 1992; Stumm, 1992). However, to our knowledge, only two studies, one of which is summarized below, have probed ligand interaction with a mineral surface using force microscopy (Kendall and Hochella, 2003; Kreller et al., 2002). A discussion of this burgeoning application begins with a description of protocols enabling linkage of a ligand to an AFM tip.

A. LIGAND LINKAGE SCHEMES Devising a suitable linkage scheme to attach the ligand of interest to the AFM probe can present a signiﬁcant challenge. Each scheme should be appropriately tailored to the relevant experimental goal; however, the following summarizes general considerations. Successful linkage will provide a strong (e.g., covalent or stronger than the interaction of interest), reproducible bond between the ligand and the tip while avoiding non-speciﬁc interactions associated with the cantilever material, tip or linker molecule (Wagner, 1998; Fiorini et al., 2001). Simple ligands such as carboxylate and phosphate groups are commonly linked as terminations of alkylthiol monolayers that coat the tip (Noy et al., 1997; Kreller et al., 2002). The ampiphilic molecules of the monolayer not only provides an anchor for the ligand but also serves as a spacer, providing separation between the ligand and the tip material thereby reducing non-speciﬁc interactions. Larger ligands and proteins that contain either a free amino or carboxyl group may be attached using an active ester technique commonly used

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

29

to couple two proteins (Cheung et al., 2000; Fiorini et al., 2001; Hinterdorfer et al., 2002; Kendall and Hochella, 2003). In the presence of a carboxyl group, 1-Ethyl-3-(3-Dimethylaminopropyl) carbodiimide (EDC) together with N-hydroyxsuccinimide (NHS) will form a stable, hydrolysis resistant, active succinimidyl ester that readily forms a peptide bond with an available amino group (Grabarek and Gergely, 1990). Note that the position of the amino and carboxyl groups can vary with one being supplied as a self-assembling monolayer (SAM) terminal group on the tip and the other contributed by the molecule to be attached, or vice versa. Other linkage protocols employ polyethylene glycol (PEG) as a crosslinking spacer that is terminated with various functional groups such as pyridildithiopropionate (PDP). PDP coordinates with thiol groups and nitrilotriacetic acid (NTA) which, in combination with various divalent metals, binds to consecutive histidine residues (Kienberger et al., 2000; Schmitt et al., 2000; Hinterdorfer et al., 2002). One advantage in using PEG-NTA-Me2þ-His linkage system is that selection of the divalent metal (Cu2þ, Co2þ, Ni2þ) permits control of the binding force, and, to a certain extent, the probability of the linkage. In addition, the NTA-Me2þ-His bond is easily reversible, such that it can be terminated with the use of EDTA, and then regenerated with the reintroduction of the free metal (Schmitt et al., 2000). Other workers propose attaching ligands or molecules via carbon nanotubes that extend from the AFM tip (Wong et al., 1998a, b, Hafner et al., 2001). This provides ideal spacing between the molecule and the tip, but more importantly, drastically increases the resolution of the force spectroscopy (and imaging) due to the nanotube’s extremely small radius of curvature compared to a traditional Si3N4 tip. Because nanotubes can only be functionalized at the end termination of the carbon lattice this also places an important constraint on the orientation and localization of the molecules being linked. As a result, the probability of capturing a single molecule interaction is increased, especially when working with lower molecular weight molecules. It is important that the linkage must not directly interfere with the activity of the ligand (Fiorini et al., 2001), and thus, electron donor functional groups should be protected during the linkage reaction. Kendall and Hochella (2003) accomplished this by inserting a metal (Al3þ) into the ligand (azotobactin) structure to occupy and protect the chelating groups, while carrying out the linkage reaction. Once attached to the tip, the azotobactin was reactivated by removing the Al with high concentrations of a competing ligand (EDTA), a process that was monitored in a test solution with UV – vis spectroscopy (Fig. 6). Unfortunately, inherent to ﬁxing a molecule to a surface is a reduction in the degree of freedom afforded to the molecule’s conformation. This can result in an alteration or loss of chelation or ligand activity and should be considered. To this end, control activations are often run in parallel to tip activations, where monolayers, linker molecules and the biomolecule of interest are reacted with a ﬂat, Si3N4 or SiOH substrate (Fiorini et al., 2001; Hinterdorfer et al., 2002). Similar in composition to the tip, the ﬂat test substrates serve as a proxy for

30

T. A. KENDALL AND S. K. LOWER

Figure 6 UV –vis spectra showing the transition of Al into and out of the azotobactin (Azb) structure; corrected for dilution. Upon the addition of Al to the system a characteristic shoulder appears in the spectra. This shoulder could be eliminated with high concentrations of EDTA. A similar process was employed to protect and then regenerate the azotobactin chelating groups during linkage of the siderophore to a hydrazide terminated AFM tip (see Kendall and Hochella, 2003).

tip that are readily probed with AFM imaging, ﬂuorescence and confocal microscopy, surface plasmon resonance (SPR) and various enzyme and ligand assays in an effort to assess the success of the linkage reaction; estimate coverage, density and footprint area of the monolayer-biomolecule construct; and evaluate activity retention in the immobilized biomolecule (Fiorini et al., 2001).

B. SIDEROPHORES

AND

OXIDE SURFACES

Kendall and Hochella (2003) collected force signatures between a ligand (siderophore) and two mineral bound metals (Fe(III) and Al(III)) in an attempt to examine the mechanism of siderophore-mediated dissolution of oxide surfaces. Siderophores are ligands produced by microorganisms to assimilate the essential nutrient ferric iron, in spite of its extreme insolubility in near surface, circumneutral environments. The aqueous chemistry of siderophores has long been studied (Winkelmann, 1991), and it is recognized that their effectiveness in acquiring iron, can, in part, be attributed to a thermodynamic binding afﬁnity for Fe(III) (aq) that has a magnitude above that for other metals, including Al(III). Only recently, however, it was recognized that, in addition to the formation of stable, aqueous iron complexes, siderophores can release iron from minerals (Seaman et al., 1992; Watteau and Berthelin, 1994; Hersman et al., 1995; Holmen and Casey, 1996; Liermann et al., 2000; Maurice et al., 2000). The mechanism of this release, however, is not clearly deﬁned. Force microscopy of the pyoverdin type siderophore azotobactin interacting with iron and aluminum oxide surfaces showed a unique relationship between

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

31

ligand – metal afﬁnity and adhesion forces (Kendall and Hochella, 2003). Average adhesion forces between azotobactin and goethite (a-FeOOH) at pH 7 were 2 –3 times the value between azotobactin and goethite’s isostructural Al-equivalent, diaspore (a-AlOOH) (Fig. 7a). A similar force relationship was also observed between the trihydroxamate siderophore deferoxamine (DFO) and each oxide surface (Fig. 7b). Control experiments where each mineral surface was probed with a SAM coated tip lacking the azotobactin molecule produced force signatures that were almost identical, indicating the distinction in the force signature between diaspore and goethite could be attributed to the presence of the azotobactin on the tip. Force measurements collected under various solution conditions (e.g., pH, ionic strength and soluble iron concentrations) and at different sample locations on the mineral extended the characterization of the ligand –mineral interaction and helped identify the source of discrepancy in adhesion values associated with each oxide. As a ﬁrst guess, it could be hypothesized that the forces of interaction

Figure 7 Force spectra showing the interaction of two siderophores (a) azotobactin and (b) deferoxamine (DFO) with goethite (FeOOH) and diaspore (AlOOH) surfaces. Note the large increase in the adhesion force between each siderophore and goethite and versus the adhesion value for diaspore.

32

T. A. KENDALL AND S. K. LOWER

are dominated by an electrostatic component; and that the difference in the adhesion values between diaspore and goethite, simply reﬂects variation in the charge character of each mineral. Although point-of-zero-charge (pzc) literature values for goethite (pzc 7 –9) and diaspore (pzc 7 –8) are similar, suggesting both should be neutral or slightly positively charged (Cornell and Schwertmann, 1996; Kosmulski, 2001), it is possible that our model system deviates from pristine charge conditions, such that the goethite is positive and the diaspore is slightly negative. With a net negative charge predicted for azotobactin at pH 7 (pKa hydroxycarboxylate ¼ 4 –5; Telford and Raymond, 1996), this could result in a larger adhesion force for goethite presumably due to a stronger electrostatic interaction. Measurements at lower pH and different sample locations, however, suggest otherwise. At pH 3.5, far from the pzc value of each mineral, and where the azotobactin is anticipated to be neutral, the same 2 –3-fold increase in adhesion values is observed. Moreover, the azotobactin-goethite/azotobactindiaspore force relationship remained intact when comparing adhesion distributions representative of different sample locations on each mineral surface, where anomalous charge distributions and changes in microtopography are expected. Overall, similar to observations made with the biotin ligand system (Moy et al., 1994a; Chilkoti et al., 1995; Izrailev et al., 1997), adhesion values upon retraction appear to be relatively independent of protonation equilibria, and may reﬂect a speciﬁc interaction between the siderophore oxygens and the metal contained in each mineral. The discrepancy in adhesion for goethite versus diaspore can then be explained by differences in the electronic character of each metal (e.g., Fe(III) versus Al(III)), where the more electronegative ferric iron will behave as a harder acid with a higher afﬁnity for the oxygens. In additional experiments with goethite only, this surface afﬁnity was readily disrupted with the addition of soluble iron (Fig. 8). Here, increased [FeCl3(6H2O] (pH 3.5) led to a saturation of the ligand as the soluble iron out competed the mineral iron for the siderophore oxygens, resulting in lowered adhesion values. This does not discount an electrostatic component to the azotobactin –oxide interaction. Indeed, decrease in the jump to contact distance with increasing ionic strength thought to reﬂect a collapse in the double-layer associated with the mineral surface (Noy et al., 1997; Lower et al., 2000), conﬁrms the effect of charge, especially upon approach. Instead, force evidence suggests a balance between electrostatics dominating the approach and more speciﬁc interactions directing surface adhesion; a scheme that is embodied in the following observation — goethite force signatures at pH 7 often show a long range, electrostatic repulsion on approach that was equal to or signiﬁcantly lower in the diaspore signatures; yet, the goethite adhesion force averaged 3.81 nN compared to 1.38 nN for diaspore. Force data also provided information on which azotobactin functional groups might be important in the interaction. Distinctive plateaus were commonly observed after retracting the tip ~6 – 7 nm from the surface (Fig. 9). These features

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

Figure 8 Plot showing decrease in azotobactin– geothite adhesion forces with increasing concentrations of added soluble iron (FeCl3 6H2O). Measurements were collected at pH 3.5 to minimize the precipitation of solid iron phases upon the addition of the iron chloride. Taken from Kendall and Hochella (2003).

33

34 T. A. KENDALL AND S. K. LOWER Figure 9 Plateau feature common in many retraction curves while probing oxide surfaces with an azotobactin activated AFM. It is suggested that this feature may represents the extension of the azotobactin and linker molecule during separation from the mineral surface as shown in the inset (not to scale). Also shown in the inset is the geometry of the linkage of the siderophore to the tip. Modiﬁed from Kendall and Hochella (2003).

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

35

are thought to represent the energy absorption associated with the combined extension and stretching of azotobactin’s polypeptide chain and the molecule used to link the ligand to the tip. Using an approximation of 0.38 nm amino acid, a quick calculation shows that azotobactin’s fully outstretched length of , 3.8 nm, together with an additional 3 nm from the linker molecule gives a value that is close to the 6 – 7 nm observed in the force signatures. This distance, then, requires that azotobactin’s terminal homoserine group serves as an anchor to the surface, providing a strong, persistent link in the interaction. This coincides with other reports that, in aqueous systems, the adjacent hydroxamate group initiates chelation (Telford and Raymond, 1996; Albrecht-Gary and Crumbliss, 1998). Additionally, considering its terminal position on the molecule, it is feasible that the homoserine group is a dominant component during surface interaction. Finally, these force microscopy results give cause to reassess the role of large ligands, such as azotobactin, in dissolving minerals. Instead of serving as an Fe shuttle between smaller ligands that interact with the surface and the cell, the force evidence demonstrating azotobactin’s strong surface afﬁnity presents a distinct possibility of the relatively large molecule entering into a strong, stable complex with the mineral. As seen above, the force data also allow comment on the nature of the association with the surface. Steric constraints imposed by ligand size, structure and conformation, together with the limited access to an iron atom contained on a mineral surface, would certainly preclude the hexadentate coordination characteristic of the siderophore-aqueous complex (Holmen and Casey, 1996; Hersman, 2000; Cocozza et al., 2002). Instead, plateau features in the retraction curves suggest a strong coordination formed by a single oxygen pair that terminates the azotobactin molecule as one possibility. Recent, ongoing MD simulations, in collaboration with U. Becker (University of Michigan), conﬁrm this possibility, as well as the extended dimensions of the azotobactin-linker construct. Interestingly enough, however, simulations reveal that the spacing between the two, chelating hydroxamate oxygens is sufﬁcient to allow individual coordination with neighboring irons in the goethite structure (Fig. 10). Siderophore– oxide interaction continues to be examined with molecular dynamic simulation as well as dynamic force spectroscopy.

V.

FORCES AT THE BACTERIUM–MINERAL INTERFACE

A. FORCE MICROSCOPY TECHNIQUE USING WHOLE CELLS The fundamental forces between a bacterium and mineral surface are central to understanding the intricacies of interfacial phenomena such as bacterial

36

T. A. KENDALL AND S. K. LOWER

Figure 10 Molecular model of azotobactin (with linker molecule) interacting with a goethite surface. Simulations were completed using Cerius2, Accelrys, Inc. Arrows point to terminal hydroxamate group oxygens interacting and coordinating with irons (balls) in the lattice. Note the spacing of the siderophore oxygens allow for “bonds” (i.e., Fe–O distances ,2.1 Angstroms) with neighboring irons. With this coordination, the cross-distance between a siderophore oxygen and an iron diagonally across is over 3 Angstroms.

adhesion to minerals and dispersal in the environment (van Loosdrccht et al., 1989; Fletcher, 1996), mineral growth and dissolution (Myers and Nealson, 1988; Hiebert and Bennett, 1992; Schultze-Lam et al., 1992; Roden and Zachara, 1996; Fortin et al., 1997), bioﬁlm formation (Lawrence et al., 1991; Davies et al., 1998),

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

37

and bacterial afﬁnity for or recognition of speciﬁc mineral surfaces (Ohmura et al., 1993; Fleminger and Shabtai, 1995; Bhosle et al., 1998; Dziurla et al., 1998; Edwards et al., 1998). A myriad of physicochemical interactions occur at biological – mineral interfaces in nature, due to (1) the mosaic of spatially discrete macromolecular cell wall structures on bacteria, (2) the dynamic nature of these structures, and (3) the diversity of mineral surface functionality, topography, and crystallography (Lower et al., 2000). As discussed above in section II, these interactions are expected to be governed by the cumulative effects of intermolecular forces (Israelachvili and McGuiggan, 1988; Israelachvili, 1992; Kendall, 1994; Butt et al., 1995; Fletcher, 1996; Gay and Leibler, 1999). However, acquiring even an elementary appreciation of these forces presents a daunting challenge, primarily due to the minute scale at which these interfaces must be probed, and the difﬁculty in developing a technique that preserves the natural intricacies of the bacterial surface (Lower et al., 2000). Measurement of fundamental forces between whole bacterial cells and inorganic phases can be conducted in one of two ways with force microscopy. The ﬁrst involves “ﬁxing” cells to a solid substrate (e.g., a glass slide) and probing these cells with a force-sensing cantilever. The simplest setup makes use of the sharp tip that is integrated into most force microscopy cantilevers (see above). In many instances, however, this is not ideal because these tips are not well constrained with respect to their geometry and/or area of contact. As shown in section II, this greatly inﬂuences force measurements thereby making it difﬁcult to compare measured data to theoretical force models, and impedes the comparison of data collected with different tips. To overcome the limitations imposed by using a sharp tip, Ducker et al. (1991) devised a simple yet ingenious solution. They created a “colloidal tip” by attaching a glass bead to the end of a force-sensing cantilever. This bead was then used to probe a ﬂat silicon surface (Ducker et al., 1991, 1992), although such a “colloidal tip” could also be used to probe microbial cells on a glass surface. A number of companies, such as Duke Scientiﬁc Corporation, Polysciences Incorporated, and Bangs Laboratories Incorporated, sell spheres ranging in size from nanometers to micrometers. A major drawback to this scenario, however, is that it limits the inorganic phases that can be utilized to those materials commonly used to make tips (e.g., silicon and silicon nitride) or beads (e.g., plastic and glass). With the exception of silica (e.g., glass beads), minerals or other inorganic phases cannot be attached to a force-sensing cantilever. Therefore, interactions between bacteria and minerals much employ another technique. That is, the cells must be linked to the forcesensing cantilever, which is then used to probe a particular face on a mineral crystal or other surface. The ﬁrst cell to be linked to a force-sensing cantilever was a large mammalian cell (Antonik et al., 1997). This cell was not actually “attached”, rather it was induced to grow on the cantilever. The researchers conducting this experiment were not interested in measuring forces, which was fortunate because cells grew

38

T. A. KENDALL AND S. K. LOWER

on both the top and bottom surfaces of the cantilever. Hence, the cell growth would have affected not only the spring constant of the lever, but it would also alter the optical lever detection system. Nonetheless, this opened the door to a number of other protocols of linking cells to a force microscope cantilever. It is a difﬁcult challenge to link microbial cells on the order of 1 mm to the end of a cantilever. Early attempts relied on the attachment of cells that had been chemically ﬁxed or treated with harsh chemicals (e.g., gluteraldehyde) (Razatos et al., 1998a, b). While these investigations produced some very intriguing force measurements, this type of linkage protocol is often undesirable because the cells are killed in the process. Further, chemicals such as gluteraldehyde are known to denature proteins and other macromolecules. Another method was developed that allowed the force-sensing cantilever to support bacterial cells in a living, native, fully functional state — thereby creating “biologically active force probes” (Lower et al., 2000, 2001b). A polycationic linker molecule (e.g., aminopropyltriethoxysilane or polylysine) can be used to link living bacteria to a small glass bead that is then attached to the cantilever, or the bacteria can be linked directly onto the cantilever itself (Lower et al., 2001b). Polycationic linkers work well because many bacteria are negatively charged over a wide range of pH conditions. Hydrophobic molecules (e.g., octadecyltrichlorosilane) are also attractive linkers because many microorganisms have hydrophobic surfaces. Techniques similar to afﬁnity chromatography (e.g., see Pleuddemann, 1991; Egger et al., 1992; Rezania et al., 1999) may be employed to design tailor-made linker molecules (e.g., ligand –receptor or antibody – antigen) that work on virtually any bacterial species. The use of polycationic linkers, or similar molecules, preserves the natural conformation, structure, and function of the macromolecules on the microbial surface. When live cells are used (i.e., a biologically active force probe), force measurements may be collected under different physiological or environmental conditions in real time (Lower et al., 2000, 2001a, b). Finally, for larger microbial cells such as yeast or fungal cells, the “colloidal tip” technique (see above) can be used to glue a single cell to the end of a cantilever (Bowen et al., 1998b). Using these techniques, a number of groups have used force microscopy to measure intermolecular forces at the bacterium – mineral interface (Ong et al., 1999; Bowen et al., 1998a; Razatos et al., 1998a, b; Bowen et al., 2000a, b; Camesano and Logan, 2000; Lower et al., 2000, 2001a, b). In our laboratories, we have used biological force microscopy (Lower et al., 2000) to measure intermolecular forces between living bacteria (e.g., E. coli, Burkholderia cepecia, and S. oneidensis) and inorganic phases (e.g., muscovite, goethite, diaspore, graphite, and glass) in solutions of varying ionic strength, pH, and oxygen concentration (Lower et al., 2000, 2001a, b). Below we will examine the force– distance relationships at the E. coli – muscovite and S. oneidensis– goethite interfaces. We will concentrate on the forces measured upon approach of a bacterium towards a mineral in the case of the former. For the latter system

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

39

(i.e., S. oneidensis – goethite), we will explore forces measured as the two surfaces are pulled apart or retracted from one another.

B. FORCES

BETWEEN

ESCHERICHIA

COLI AND

MUSCOVITE

Figure 11 shows the interaction between E. coli and the (001) surface of muscovite as the sodium chloride (NaCl) solution was exchanged ﬁve times between low (, 1025 M) and high (~102l M) ionic strength. While both approach and retraction forces were measured, shown in Fig. 11 are only the forces detected upon approach of the mineral towards living cells on a biologically active force probe. At low ionic strength, repulsive (positive sign) forces were detected at a separation of approximately 100 run. This repulsive interaction increased exponentially (see below) to a maximum value of , 30 – 35 nN at contact. At high ionic strength, the magnitude of repulsion was signiﬁcantly less as was the range of separation over which force interactions took place. The two surfaces did not “feel” one another until they were within 15– 20 nm of separation. As with the measurements at low ionic strength, an exponential force appears to dominate at high ionic strength. It is important to note that these data shown in Fig. 11 span the entire range of measurements for literally hundreds of force – distance curves taken as a solution was exchanged

Figure 11 Force–distance relationship between the basal plane surface of muscovite and E. coli in solutions of high (open symbols across lower portion of ﬁgure) or low (closed symbols across upper portion of ﬁgure) ionic strength. Shown for each solution condition are ﬁve data curves that span the entire range of measurements for literally hundreds of force-distance curves. The lines correspond to the DLVO model prediction at high (dotted) or low (solid) ionic strength. Repulsive forces have a positive sign; whereas attractive forces have a negative sign. Only those forces measured upon approach of the mineral towards the bacteria are shown. See text for discussion.

40

T. A. KENDALL AND S. K. LOWER

several times between high and low ionic strength. Hence, the measurements are reproducible. Results can be interpreted with the so-called DLVO theory (Derjaguin and Landau, 1941; Verwey and Overbeek, 1948). This theory describes forces (F) as a function of the distance (D) (e.g., between a bacterium, treated as a sphere and a mineral, treated as a ﬂat plane) as the sum of the electrostatic and van der Waals forces (Ducker et al., 1991; Israelachvili, 1992; Butt et al., 1995; Muller and Engel, 1997): FDLVO ðDÞ ¼ Felectrostatic ðDÞ 2 Fvdw ðDÞ ¼

4psbacteria smineral R 2kD HR e 2 a2 110 k 6D

where s is the surface charge density (C m22), R is the radius of a cell (or in this case the radius of the bacteria coated bead attached to the cantilever), 1 is the dielectric constant of water (78.54 at 298 K), 10 is the permittivity of free space (8.854 £ 10212 C2 J21 m2l), k is the inverse Debye length (Debye length , 1 nm at 102l M and , 100 nm at 1025 M; see above), and Ha is the Hamaker constant. For the model results plotted in Fig. 11, Hamaker’s constant was 10221 J (Vigeant et al., 2002); surface charge density of the bacterium was estimated using Eq. (3) as 2 0.001 or 2 0.04 C m22 at low or high IS, respectively (surface potential measurements were taken from Camesano and Logan (2000), Ong et al. (1999) and Vigeant et al. (2002), and the surface charge density of the mineral was estimated using Eq. (3) as 2 0.004 or 2 0.2 C m22 at low or high ionic strength, respectively (surface potential measurements were taken from Ducker et al. (1992) and Ong et al. (1999)). Figure 11 compares the measured forces with those predicted by DLVO theory. Ionic strength (approx. 102l versus 1025 M) appears to have a strong effect on the interactions between E. coli and muscovite. This is because higher salt concentrations cause the electrostatic double-layer to become thinner (i.e., surfaces cannot “feel” one another until they get very close). Further, the increased concentration of counter ions at high ionic strength effectively screens the negative charges on both surfaces, thereby resulting in smaller magnitude forces of repulsion. These particular measurements are fairly consistent with DLVO theory. However, there are some important discrepancies. For example, at low ionic strength the attractive van der Waals force is expected to dominate the interaction at separations less than 5 nm. However, measurements reveal that E. coli and muscovite do not exhibit attraction even at the closest approach. Indeed, adhesion forces were not detected when E. coli and muscovite were forced together and subsequently pulled apart at low ionic strength (Lower et al., 2000). This suggests that electrostatic and/or other repulsive forces (e.g., solvation interactions) dominate this particular interaction. Many other force measurements conducted in our laboratories, suggest that electrostatic and van der Waals forces are not the only intermolecular forces at the bacterium – mineral interface (S. Lower, unpublished results). Others have

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

41

attempted to invoke extended-DLVO models to explain deviations from purely van der Waals and electrostatic forces and ﬁt model predictions to measurements (see e.g., Ong et al., 1999; Camesano and Logan, 2000). While these investigations may be valid, one needs to remember that force models are sensitive to the geometric shapes of interacting particles as well as the roughness of surfaces and contact area (Israelachvili, 1992) all of which are difﬁcult to rigorously deﬁne or control for minerals and cells with biopolymers. Further, DLVO was developed to describe the phenomena between inanimate particles rather than living cells that have exquisite control over the expression of surface macromolecules. Seemingly simple concepts such as “contact” become difﬁcult to deﬁne for cells having polymers of varying length, which extend for some distance beyond the cell wall. Further, living cells and/or surfaces with polymers are expected to show a time dependent adhesion (measured by Lower et al., 2001a) as biopolymers diffuse into the cell wall and reorient with respect to another surface (Beveridge, 1999; Leckband and Israelachvili, 2001). The true value may not be in whether a model perfectly ﬁts data, but the most deﬁnite answer comes when the measurements contradict the theory, thereby disproving a particular construct and suggesting that other forces are responsible for a particular bacteria –mineral interaction. As stated by Oreskes et al. (1994) scientiﬁc investigations are at their best when one combines experimental measurements and model predictions to challenge existing formulations. Hence, there is a great need to test such models by comparing theories to precise force measurements using “model” microorganisms and minerals. Only then will we be able to understand how all of the various intermolecular forces (e.g., electrostatic, van der Waals, hydration, hydrophobic, and steric interactions) govern interactions at the bacterium –mineral interface.

C. FORCES

BETWEEN

SHEWANELLA ONEIDENSIS OR DIASPORE

AND

GOETHITE

The forces required to pull the mineral and bacteria apart (i.e., retraction data) are not shown in Fig. 11. In fact, a very strong adhesion force was detected between E. coli and muscovite, but only at high ionic strength (Lower et al., 2000). Aside from being a notable example of a situation that DLVO theory often cannot explain, retraction data provide an immense amount of information about the adhesive strength and structural properties of speciﬁc biopolymers on a cell’s surface. Recently, we interpreted these retraction data for studies of bacterial adhesion and electron transfer reactions between S. oneidensis (a dissimilatoryiron-reducing-bacteria) and the minerals goethite (FeOOH) or diaspore (AIOOH) under aerobic and anaerobic solution conditions (Lower et al., 2001a; Lower et al., 2002). S. oneidensis is capable of using either oxygen or ferric iron in the crystal structure of iron oxyhydroxides as a terminal electron

42

T. A. KENDALL AND S. K. LOWER

acceptor (Nealson and Saffarini, 1994). We used force microscopy to determine whether a microorganism could discriminate between two very similar minerals (diaspore and goethite) under anaerobic conditions, when electron transfer is expected to occur between S. oneidensis and iron containing minerals. A mineral crystal, mounted on a piezoelectric scanner, approached live bacteria on a biologically active force probe at a rate that was comparable to the natural velocity of motile bacteria. Once contact was established, the two surfaces were pulled apart resulting in retraction data. Figure 12 illustrates the retraction proﬁle for S. oneidensis and goethite versus diaspore under anaerobic or aerobic conditions (Lower et al., 2001a). The intricate details of these curves and the entire data set provide valuable information about intermolecular forces and structures at the bacterium –mineral interface. Initially the entire retraction data were characterized by integrating the force with respect to distance. This provided quantitative energy values associated with adhesion. The retraction curves were further analyzed by the worm-like chain model (see above) to establish a correlation between speciﬁc bridging polymers and unique signatures in the retraction curves. Energy values determined from retraction curves similar to those shown in Lower et al. (2001a) revealed that S. oneidensis had a higher afﬁnity for diaspore (versus goethite) under aerobic conditions (Lower et al., 2001a). However, under anaerobic conditions the bacteria exhibited a signiﬁcant increase in afﬁnity for goethite (see Lower et al., 2001a); whereas the adhesion energy for diaspore was indifferent to oxygen concentrations (Lower et al., 2001a). The attractive energy between S. oneidensis and goethite was 30 aJ (aJ ¼ 10218 J) and 130 aJ under aerobic and anaerobic conditions, respectively. Further, the energetic afﬁnity between goethite and S. oneidensis also increased with contact time under

Figure 12 Measured force-distance (or force-extension) relationship between living cells of S. oneidensis and diaspore (AlOOH) under anaerobic conditions (solid circles). Interactions with goethite (FeOOH) under aerobic (“X” symbols), or anaerobic conditions (open squares). Also shown is the theoretical prediction for the unravelling of a 150 kDa protein that may tether the bacteria to the surface of goethite (see text). Attractive forces, shown here, have a negative sign. Shown are only those forces measured as the mineral is pulled away from the bacteria (retraction data). Modiﬁed from Lower et al. (2001a).

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

43

anaerobic conditions. This provided quantitative evidence suggesting that this microorganism recognized a particular inorganic surface such that it localized and/or produced biopolymers to mediate contact with goethite under anaerobic conditions. This idea was further tested by using a “bridging polymer” model to interrogate the intricate details (e.g., the saw-tooth pattern) in the retraction data. As shown above, for the worm-like chain model, linear polymers such as proteins are expected to unravel according to Eq. (4). According to this model, one needs only the persistence length and the contour length to describe the force as a function of the distance a polymer is extended. This model describes a physical process similar to that which is recorded in the retraction data of force microscopy. A protein, for example, in the cell wall of a bacterium makes contact with a mineral. The mineral is then pulled away from the bacterium causing the protein to unravel until it is completely extended at which point the protein is either ripped from the cell wall or it breaks free of the mineral surface and recoils into the cell surface. The outer surface proteins of Shewanella are well characterized. Shewanella is known to have proteins on its outer membrane that mediate contact with iron oxyhydroxides (Caccavo, 1999; Caccavo et al., 1997; Das and Caccavo, 2000). Several of these proteins are putative iron reductases, which are expected to make physical contact with goethite such that they can transfer electrons across the organic – inorganic interface (Arnold et al., 1990; Myers and Nealson, 1990; Myers and Myers, 1992, 1993, 1997; DiChristina and Delong, 1994; Nealson and Saffarini, 1994; Roden and Zachara, 1996; Myers and Myers, 1998, 2000, 2001). In fact, four putative iron reductase proteins have been characterized according to their mass and/or genetic sequence (Myers and Myers, 1997, 1998, 2000, 2001, 2002). The worm-like chain model was used to predict the way in which each of these four proteins (ranging in size from , 50 to 150 kDa; (Myers and Myers, 1997; Myers and Myers, 1998) would unfold. The molecular mass of each protein was used to estimate its overall length according to the following conversion, , 110 Da per amino acid residue (Voet and Voet, 1995). The persistence length of each amino acid, deﬁned as the distance between two adjacent Ca, is equal to 0.38 nm (Karlsson et al., 1996; Muller et al., 1999; Myers and Myers, 2001). As shown in Fig. 12, the saw-tooth pattern at approximately 500 nm corresponds to the force-extension proﬁle of one of the four putative iron reductase proteins. This proﬁle was reproducible suggesting that the cell wall protein was not ripped from the bacterium, but rather retained its native conformation after multiple extensions. This unique signature was present only for goethite under anaerobic conditions where it was detected in , 80% of the retraction curves, but only when the bacterium was given some period of time to make contact with the surface of goethite (Lower et al., 2001a). This suggests that the bacterium required time to “recognize” the mineral surface and subsequently create and/or mobilize a speciﬁc protein to the area of contact with goethite.

44

T. A. KENDALL AND S. K. LOWER

VI. FUTURE WORK On Earth, literally millions of different species of prokaryotes may interact with any of the thousands of different minerals. An interface is formed at the junction of a bacterium and a mineral surface that is complex, dynamic, and by its very nature, nanoscale in size. This is because bacteria are living cells that have mastered the art of synthesizing fully functional structures (e.g., lipids, proteins, polysaccharides) and utilizing properties that exist only at the nanometer scale. The study of the interface between minerals and microorganisms requires a unique fusion of geomicrobiology and nanoscale science. What are the fundamental forces that control the binding of a silanol group on a mineral surface to a carboxylic group in a bacterium’s cell wall? How does the density and distribution of functional groups on a crystal face inﬂuence the way microorganisms sense mineral surfaces? Do bacteria express speciﬁc outer surface proteins to interact with certain minerals? How do bacteria modulate forces of interaction between themselves and minerals (or other bacteria) to either enhance or inhibit adhesion and subsequent bioﬁlm formation? Researchers must be able to thoroughly explore both sides of the interface (i.e., the bacterium and the mineral) and the fundamental nanoscale forces in the intervening region to discover phenomena that exist only in the nanospace between a microorganism (or microbially produced polymers) and a mineral surface. As mentioned earlier, application of force microscopy to the biogeosciences is in its infancy, and there exists many other uses and unexplored possibilities of force experiments with ligands, microorganisms, and minerals. Structural elements within a particular biomolecule/ligand may contribute to its ability to bind to a surface or promote dissolution, or chelate dissolved or mineral bound metals (Stumm, 1992; Ludwig et al., 1995; Nubel et al., 1996). One can envision collecting a force signature for a large ligand interacting with a mineral, followed by collection of spectra associated with several, individual cognate functional groups associated with the ligand. Comparison of the whole ligand, baseline spectrum with the individual component spectra could reveal which functional groups are dominating the interaction with the mineral. Or, a similar process could be achieved by making force measurements after successive chemical modiﬁcation of the original ligand structure. Such modiﬁcation might include inactivation of a speciﬁc functional group with a residue-speciﬁc reactive reagent (Voet and Voet, 1995), or an amino acid substitution resulting from alteration of the genes associated with the biosynthesis of the molecule. Again changes in the force signature with each modiﬁcation might help determine the critical moieties contributing to the interaction. Force maps (Noy et al., 1997) are also possible using ligand activated tips. Here, the contrast in the map may be supplied by the differential adhesion between the ligand and various metals that are associated with a surface.

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

45

For example, a map constructed with a siderophore activated tip might show large adhesions in areas of high concentrations of trivalent metals such as Fe(III) or Al(III) and lower adhesions for divalent metals such as Cu(II), Zn(II) or Fe(II). Given the spatial resolution of the AFM, such images could be useful for identifying contaminant distribution on a surface or pinpointing impurity concentrations on a mineral growth face, both on a nanometer scale. CFM is traditionally carried out in a ﬂuid cell (Digital Instruments) that allows direct observation of ligand –surface interaction under environmentally relevant conditions with pico- to nanonewton force resolution and a spatial resolution of tens of nanometers down to potentially the atomic level. Changes in the forces of interaction with solution composition provide important information about the structure and charge character of the ligand and mineral surface, and the nature of the interaction between the two. While the effect of solution composition (e.g. pH) on ligand sorption can be monitored with force measurements using a force titration (Kreller et al., 2002). The sensitivity of this technique also allows small changes in mineral solubility and associated metal concentrations, pH, and ionic strength to be detected (Kendall and Hochella, 2003). Given the spatial resolution mentioned above, this opens up the possibility of using this technique to detect localized solution micro- or even nanoenvironments associated with a surface. Finally, force investigations with living microorganisms are rich with possibilities. For example, one could measure forces of adhesion using wildtype stains versus mutants that are incapable of producing speciﬁc cell wall macromolecules. These data may result in unique force signatures characteristic of particular biomolecules. Force measurements could also be coupled to other techniques such as confocal scanning laser microscopy. This provides the potential to collect force measurements concurrent with ﬂuorescence observations of the distribution and localization of cell wall macromolecules.

ACKNOWLEDGMENTS SKL acknowledges the support of the National Science Foundation, the Department of Energy, the American Chemical Society, and the General Research Board of the University of Maryland. SKL would also like to thank J. Tak for support. Funding was provided to TAK by a GAAN Fellowship (U.S. Dept. of Education), the NSF’s Nanoscale Science and Engineering (NSE) Program (EAR 01-03053), and the Department of Energy’s OBES Geosciences Program (DE-FG02-99ER 15002). TAK acknowledges the support of Michael F. Hochella, Jr. (Virginia Tech).

46

T. A. KENDALL AND S. K. LOWER

REFERENCES Absolom, D. R., Lamberti, F. V., Policova, Z., Zingg, W., van Oss, C. J., and Neumann, A. W. (1983). Surface thermodynamics of bacterial adhesion. Appl. Environ. Microbiol. 46, 90–97. Albrecht-Gary, A.-M., and Crumbliss, A. L. (1998). Coordination chemistry of siderophores: Thermodynamics and kinetics of iron chelation and release. In “Iron transport and storage in microorganisms, plants, and animals” (H. Sigel, Ed.), Vol. 35, pp. 239–327. Marcel Dekker, New York. Antonik, M. D., D’Costa, N. P., and Hoh, J. H. (1997). A biosensor based on micromechanical interrogation of living cells. IEEE Engng. Med. Biol. 16, 66 –72. Aoki, T., Hiroshima, M., Kitamura, K., Tokunaga, M., and Yanagida, T. (1997). Non-contact scanning probe microscopy with sub-piconewton force sensitivity. Ultramicroscopy 70, 45– 55. Arnold, R. G., Hoffmann, M. R., DiChristina, T. J., and Picardal, F. W. (1990). Regulation of dissimilatory Fe(III) reductase activity in Shewanella putrefaciens. Appl. Environ. Microbiol. 56, 2811–2817. Ashby, P. D., Chen, L., and Lieber, C. M. (2000). Probing intermolecular forces and potentials with magnetic feedback chemical force microscopy. J. Am. Chem. Soc. 122, 9467–9472. Barker, W. W., Welch, S. A., and Banﬁeld, J. F. (1997). Biogeochemical weathering of silicate minerals. In “Geomicrobiology: Interactions Between Microbes and Minerals” (K. H. Nealson, Ed.), Vol. 35, p. 448. Mineralogical Society of America, Washington DC. Baumgartner, W., Hinterdorfer, P., and Schindler, H. (2000). Data analysis of interaction forces measured with the atomic force microscope. Ultramicroscopy 82, 85 –95. Bell, G. I. (1978). Models for speciﬁc adhesion of cell to cells. Science 200, 618 –627. Beveridge, T. J. (1999). Structures of Gram-negative cell walls and their derived membrane vesicles. J. Bacteriol. 181, 4725–4733. Bhattacharjee, S., Chen, J. Y., and Elimelech, M. (2000). DLVO interaction energy between spheroidal particles and a ﬂat surface. Colloid. Surf. A: Physicochem. Engng. Aspects 165, 143–156. Bhosle, N., Suci, P. A., Baty, A. M., Weiner, R. M., and Geesey, G. G. (1998). Inﬂuence of divalent cations and pH on adsorption of a bacterial polysaccharide adhesion. J. Colloid. Interf. Sci. 205, 89–96. Binnig, G., Quate, C. F., and Gerber, C. (1986). Atomic force microscope. Phys. Rev. Lett. 56, 930–933. Bossier, P., Hofte, M., and Verstraete, W. (1988). Ecological Signiﬁcance of Siderophores in Soil. In “Advances in Microbial Ecology” (K. C. Marshall, Ed.), Vol. 10, pp. 385–414. Plenum Press, New York. Bowen, R. W., Hilal, N., Lovitt, R. W., and Wright, C. W. (1998a). Direct measurement of interactions between adsorbed protein layers using an atomic force microscope. J. Colloid Interf. Sci. 197, 348–352. Bowen, R. W., Hilal, N., Lovitt, R. W., and Wright, C. W. (1998b). Direct measurement of the force of adhesion of a single biological cell using an atomic force microscopy. Colloid Surf. A: Physicochem. Engng. Aspects 136, 231–234. Bowen, W. R., Lovitt, R. W., and Wright, C. J. (2000a). Direct quantiﬁcation of Aspergillus niger spore adhesion in liquid using an atomic force microscope. J. Colloid Interf. Sci. 228, 428– 433. Bowen, W. R., Lovitt, R. W., and Wright, C. J. (2000b). Direct quantiﬁcation of Aspergillus niger spore adhesion to mica in air using an atomic force microscope. Colloid Surf. A-Physicochem. Engng. Aspects 173, 205–210. Brantley, S. L., Liermann, L., and Bau, M. (2001). Uptake of trace metals and rare earth elements from hornblende by a soil bacterium. Geomicrobiol. J. 18, 37–61.

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

47

Burnham, N. A., Dominguez, D. D., Mowery, R. L., and Colton, R. J. (1990). Probing the surface forces of monolayer ﬁlms with an atomic-force microscope. Phys. Rev. Lett. 64, 1931–1934. Bustamante, C., Marko, J. F., Siggia, E. D., and Smith, S. (1994). Entropic elasticity of lambda-phage DNA. Science 265, 1599–1600. Butt, H. J., and Jaschke, M. (1995). Calculation of Thermal Noise in Atomic-Force Microscopy. Nanotechnology 6, 1– 7. Butt, H. J., Jaschke, M., and Ducker, W. (1995). Measuring surface forces in aqueous electrolyte solution with the atomic force microscope. Bioelectrochem. Bioenerg. 38, 191 –201. Caccavo, F. (1999). Protein-mediated adhesion of the dissimilatory Fe(III)-reducing bacterium Shewanella alga BrY to hydrous ferric oxide. Appl. Environ. Microbiol. 65, 5017–5022. Caccavo, F. Jr., Schamberger, P. C., Keiding, K., and Nielsen, P. H. (1997). Role of hydrophobicity in adhesion of the dissimilatory Fe(III)-reducing bacterium Shewanella alga to amorphous Fe(III) oxide. Appl. Environ. Microbiol. 62, 3837–3843. Camesano, T. A., and Logan, B. E. (2000). Probing bacterial electrosteric interactions using atomic force microscopy. Environ. Sci. Technol. 34, 3354–3362. Cappella, B., Baschieri, P., Frediani, C., Miccoli, P., and Ascoli, C. (1997). Force–distance curves by AFM: A powerful technique for studying surface interactions. IEEE Engng. Med Biol. Mag. 16, 58 –65. Cappella, B., and Dietler, G. (1999). Force–distance curves by atomic force microscopy. Surf. Sci. Rep. 34, 1–104. Cherian, S., and Thundat, T. (2002). Determination of adsorption-induced variation in the spring constant of a microcantilever. Appl. Phys. Lett. 80, 2219–2221. Cheung, C. L., Hafner, J. H., Chen, L. W., and Lieber, C. M. (2000). Direct growth of single-walled carbon nanotube scanning probe microscopy tips and application to ultrahigh-resolution chemical force microscopy and structural imaging. Abstr. Pap. Am. Chem. Soc. 219, 182. Chilkoti, A., Boland, T., Ratner, B. D., and Stayton, P. S. (1995). The relationship between ligandbinding thermodynamics and protein–ligand interaction forces measured by atomic force microscopy. Biophys. J. 69, 2125–2130. Cleveland, J. P., Manne, S., Bocek, D., and Hansma, P. K. (1993). A non-destructive method for determining the spring constant of cantilevers for scanning force microscopy. Rev. Sci. Instrum. 64, 403–405. Cocozza, C., Tsao, C. C. G., Cheah, S. F., Kraemer, S. M., Raymond, K. N., Miano, T. M., and Sposito, G. (2002). Temperature dependence of goethite dissolution promoted by trihydroxamate siderophores. Geochimica et Cosmochimica Acta 66, 431– 438. Cornell, R. M., and Schwertmann, U. (1996). “The Iron Oxides: Structure, Properties, Reactions, Occurrence and Uses”, VCH. Craig, V. S. J., and Neto, C. (2001). In Situ calibration of colloid probe cantilevers in force microscopy: hydrodynamic drag on a sphere approaching a wall. Langmuir 17, 6018–6022. Cygan, R. T. (2001). Molecular Modeling in Mineralogy and Geochemistry. In “Molecular Modeling Theory: Applications in the Geosciences” (J. D. Kubicki, Ed.), Vol. 42, p. 531. Mineralogical Society of America Geochemical Society, Washington DC. Dammer, U., Hegner, M., Anselmetti, D., Wagner, P., Dreier, M., Huber, W., and Guentherodt, H. J. (1996). Speciﬁc antigen/antibody interactions measured by force microscopy. Biophys. J. 70, 2437–2441. Dammer, U., Popescu, O., Wagner, P., Anselmetti, D., Guntherodt, H. J., and Misevic, G. N. (1995). Binding Strength Between Cell-Adhesion Proteoglycans Measured By Atomic-Force Microscopy. Science 267, 1173–1175. Das, A., and Caccavo, F. (2000). Dissimilatory Fe(III) oxide reduction by Shewnnella alga BrY requires adhesion. Curr. Microbiol. 40, 344 –347.

48

T. A. KENDALL AND S. K. LOWER

Davies, D. G., Parsek, M. R., Pearson, J. P., Iglewski, B. H., Costerton, J. W., and Greenberg, E. P. (1998). The involvement of cell-to-cell signals in the development of a bacterial bioﬁlm. Science 280, 295 –298. D’Costa, N. P., and Hoh, J. H. (1995). Calibration of optical lever sensitivity for atomic force microscopy. Rev. Sci. Instrum. 66, 5096– 5097. Degertekin, F. L., Hadimioglu, B., Sulchek, T., and Quate, C. F. (2001). Actuation and characterization of atomic force microscope cantilevers in ﬂuids by acoustic radiation pressure. Appl. Phys. Lett. 78, 1628–1630. Derjaguin, B. V., and Landau, L. (1941). Acta Physiochim. URSS 4, 633. Derjaguin, B. V., Muller, V. M., and Toporov, Y. P. (1975). J. Colloid Interf. Sci. 53. DiChristina, T. J., and Delong, E. F. (1994). Isolation of anaerobic respiratory mutants of Shewanella putrefaciens and genetic analysis of mutants deﬁcient in anaerobic growth on Fe3þ. J. Bacteriol. 176, 1468–1474. Ducker, W. A., Senden, T. J., and Pashley, R. M. (1991). Direct measurement of colloidal forces using an atomic force microscope. Nature 353, 239–241. Ducker, W. A., Senden, T. J., and Pashley, R. M. (1992). Measurement of forces in liquids using a force microscope. Langmuir 8, 1831–1836. Dziurla, M.-A., Achouak, W., Lam, B.-T., Heulin, T., and Berthelin, J. (1998). Enzyme-linked immunoﬁltration assay to estimate attachment of Thiobacilli to pyrite. Appl. Environ. Microbiol. 64, 2937–2942. Edwards, K. J., Schrenk, M. O., Hamers, R., and Banﬁeld, J. F. (1998). Microbial oxidation of pyrite: Experiments using microorganisms from an extreme acidic environment. Am. Mineral. 83, 1444–1453. Egger, M., Heyn, S.-P., and Gaub, H. E. (1992). Synthetic lipid-anchored receptors based on the binding site of a monoclonal antibody. Biochim. Biophys. Acta 1104, 45–54. Erlandsson, R., McClelland, G. M., Mate, C. M., and Chiang, S. (1988). Atomic force microscopy using optical interferometry. J. Vacuum Sci. Technol., A: Vacuum, Surf. Films 6, 266–270. Evans, E. (1999). Energy landscapes of biomolecular adhesion and receptor anchoring at interfaces explored with dynamic force spectroscopy. Faraday Discussions 111, 1 –16. Evans, E. (2001). Probing the relation between force-lifetime and chemistry in single molecular bonds. Annu. Rev. Biophy. Biomol. Struct. 30, 105–128. Evans, E., and Ipsen, J. (1991). Entropy-driven expansion of electric double layer repulsion between highly ﬂexible membranes. Electrochim. Acta 36, 1735– 1741. Evans, E., and Ludwig, F. (2000). Dynamic strengths of molecular anchoring and material cohesion in ﬂuid biomembranes. J. Phys.: Condensed Matter 12, A315–A320. Evans, E., and Ritchie, K. (1997). Dynamic strength of molecular adhesion bonds. Biophysical Journal 72, 1541– 1555. Fiorini, M., McKendry, R., Cooper, M. A., Rayment, T., and Abell, C. (2001). Chemical force microscopy with active enzymes. Biophys. J. 80, 2471–2476. Fleminger, G., and Shabtai, Y. (1995). Direct and rapid analysis of the adhesion of bacteria to solid surfaces: Interaction of ﬂuorescently labeled Rhodococcus strain GIN-1 (NCIMB 40340) cells with titanium-rich particles. Appl. Environ. Microbiol. 61, 4357–4361. Fletcher, M., (Ed.) (1996). “Bacterial Adhesion: Molecular and Ecological Diversity”. Wiley-Liss, New York. Florin, E. L., Moy, V. T., and Gaub, H. E. (1994). Adhesion Forces Between Individual Ligand– Receptor Pairs. Science 264, 415 –417. Flory, P. J. (1989). “Statistical Mechanics of Chain Molecules”. Hanser Publishers, Munich. Fortin, D., Ferris, F. G., and Beveridge, T. J. (1997). Surface-mediated mineral development by bacteria. In “Geomicrobiology: Interactions between Microbes and Minerals” (K. H. Nealson, Ed.), Vol. 35, pp. 161 –180. Mineralogical Society of America, Washington DC.

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

49

Frisbie, C. D., Rozsnyai, L. F., Noy, A., Wrighton, M. S., and Lieber, C. M. (1994). Functional-group imaging by chemical force microscopy. Science 265, 2071–2074. Gay, C., and Leibler, L. (1999). On stickiness. Phys. Today 52, 48– 52. Gergely, C., Hemmerle, J., Schaaf, P., Horber, J. K. H., Voegel, J.-C., and Senger, B. (2002). Multibead-and-spring model to interpret protein detachment studied by AFM force spectroscopy. Biophys. J. 83, 706–722. Gergely, C., Senger, B., Voegel, J. C., Horber, J. K. H., Schaaf, P., and Hemmerle, J. (2001). Semiautomatized processing of AFM force-spectroscopy data. Ultramicroscopy 87, 67–78. Gibson, C. T., Weeks, B. L., Lee, J. R. I., Abell, C., and Rayment, T. (2001). A nondestructive technique for determining the spring constant of atomic force microscope cantilevers. Rev. Sci. Instrum. 72, 2340–2343. Goddenhenrich, T., Lemke, H., Hartmann, U., and Heiden, C. (1990). Force microscope with capacitive displacement detection. J. Vacuum Sci. Technol., A: Vacuum, Surf., Films 8, 383–387. Grabarek, Z., and Gergely, J. (1990). Zero-length crosslinking procedure with the use of active esters. Anal. Biochem. 185, 131– 135. Grandbois, M., Beyer, M., Rief, M., Clausen-Schaumann, H., and Gaub, H. E. (1999). How strong is a covalent bond? Science 283, 1727–1730. Grubmuller, H., Heymann, B., and Tavan, P. (1996). Ligand binding: molecular mechanics calculation of the streptavidin –biotin rupture force. Science 271, 997 –999. Hafner, J. H., Cheung, C. L., and Lieber, C. M. (1999). Growth of nanotubes for probe microscopy tips. Nature 398, 761– 762. Hafner, J. H., Cheung, C. L., Woolley, A. T., and Lieber, C. M. (2001). Structural and functional imaging with carbon nanotube AFM probes. Progress in Biophys. & Molec. Biol. 77, 73– 110. Han, T., Williams, J. M., and Beebe, T. P. (1995). Chemical-bonds studied with functionalized atomic-force microscopy tips. Analytical Chemica Acta 307, 365–376. Hansen, D. C., McCafferty, E., Lins, C. W., and Fitzpatrick, J. J. (1995). An FT-IR investigation of parabactin adsorbed onto aluminium. Appl. Surf. Sci. 84, 85–90. Hartmann, U. (1991). van der Waals interactions between sharp probes and ﬂat sample surfaces. Phys. Rev. B 43, 2404– 2407. Heinz, W. F., and Hoh, J. H. (1999). Spatially resolved force spectroscopy of biological surfaces using the atomic force microscope. Trends Biotechnol. 17, 143–150. Hersman, L. E. (2000). The role of siderophores in iron oxide dissolution. In “Environmental microbemetal interactions” (D. R. Lovley, Ed.), p. 395. ASM Press, Washington DC. Hersman, L., Lloyd, T., and Sposito, G. (1995). Siderophore-promoted dissolution of hematite. Geochim. Cosmochim. Acta 59, 3327– 3330. Hiebert, F. K., and Bennett, P. C. (1992). Microbial control of silicate weathering in organic-rich ground water. Science 258, 278–281. Hinterdorfer, P., Baumgartner, W., Gruber, H. J., Schilcher, K., and Schindler, H. (1996). Detection and localization of individual antibody-antigen recognition events by atomic force microscopy. Proc. Natl. Acad. Sci. USA 93, 3477–3481. Hinterdorfer, P., Gruber, H. J., Kienberger, F., Kada, G., Riener, C., Borken, C., and Schindler, H. (2002). Surface attachment of ligands and receptors for molecular recognition force microscopy. Colloid Surf., B: Biointerf. 23, 115–123. Holmen, B. A., and Casey, W. H. (1996). Hydroxamate ligands, surface chemistry, and the mechanism of ligand-promoted dissolution of goethite [a-FeOOH(s)]. Geochim. Cosmochim. Acta 22, 4403–4416. Holmen, B. A., Tejedor-Tejedor, M. I., and Casey, W. H. (1997). Hydroxamate complexes in solution and at the goethite-water interface: a cylindrical internal reﬂection fourier transform infrared spectroscopy study. Langmuir 13, 2197– 2206. Hutter, J. L., and Bechhoefer, J. (1993). Calibration of atomic-force microscope tips. Rev. Sci. Instrum. 64, 1868–1873.

50

T. A. KENDALL AND S. K. LOWER

Israelachvili, J. N. (1992). “Intermolecular and Surface Forces”, 2nd edn. Academic Press, San Diego. Israelachvili, J. N., and McGuiggan, P. M. (1988). Forces between surfaces in liquids. Science 241, 795–800. Ito, T., Citterio, D., Buehlmann, P., and Umezawa, Y. (1999). Observation of silver and hydrogen ion binding to self-assembled monolayers using chemically modiﬁed AFM tips. Langmuir 15, 2788–2793. Izrailev, S., Stepaniants, S., Balsera, M., Oono, Y., and Schulten, K. (1997). Molecular dynamics study of unbinding of the avidin–biotin complex. Biophys. J. 72, 1568– 1581. Jarvis, S. P., Duerig, U., Lantz, M. A., Yamada, H., and Tokumoto, H. (1998). Feedback-stabilized force sensors. A gateway to the direct measurement of interaction potentials. Appl. Phys. A: Mater. Sci. Process. A66, S211–S213. Jarvis, S. P., Yamada, H., Yamamoto, S. I., Tokumoto, H., and Pethica, J. B. (1996). Direct mechanical measurement of interatomic potentials. Nature (London) 384, 247–249. Jeppesen, C., Wong, J. Y., Kuhl, T. L., Israelachvili, J., Mullah, J. N., Zalipsky, S., and Marques, C. M. (2001). Impact of polymer tether length on multiple ligand–receptor bond formation. Science 293, 465 –468. Jericho, S. K., and Jericho, M. H. (2002). Device for the determination of spring constants of atomic force microscope cantilever and micromachined springs. Rev. Sci. Instrum. 73, 2483–2485. Johnson, K. L., Kendall, K., and Roberts, A. D. (1971). Surface energy and the contact of elastic solids. Proc. R. Soc. Lond. A 324, 301–313. Joyce, S. A., and Houston, J. E. (1991). A new force sensor incorporating force-feedback control for interfacial force microscopy. Rev. Sci. Instrum. 62, 710–715. Kalinowski, B. E., Liermann, L. J., Brantley, S. L., Barnes, A., and Pantano, C. G. (2000). X-ray photoelectron evidence for bacteria-enhanced dissolution of hornblende. Geochim. Cosmochim. Acta 64, 1331–1343. Karlsson, J.-J., Kadziola, A., Rasmussen, A., Rostrup, T. E., and Ulstrup, J. (1996). Electron transfer of the di-heme protein: Pseudomonas stutzeri cytochrome c4. In “Protein Folds: A DistanceBased Approach” (S. Brunak, Ed.), pp. 56–67. CRC Press, Boca Raton. Kasas, S., Riederer, B. M., Catsicas, S., Cappella, B., and Dietler, G. (2000). Fuzzy logic algorithm to extract speciﬁc interaction forces from atomic force microscopy data. Rev. Sci. Instrum. 71, 2082–2086. Kendall, K. (1994). Adhesion: Molecules and mechanics. Science 263, 1720–1725. Kendall, T. A., and Hochella, M. F. Jr. (2003). The measurement and interpretation of molecular level forces of interaction between the siderophore azotobactin and mineral surfaces. Geochim. Cosmochim. Acta. Kienberger, F., Kada, G., Gruber, H. J., Pastushenko, V. P., Riener, C., Trieb, M., Knaus, H.-G., Schindler, H., and Hinterdorfer, P. (2000). Recognition force spectroscopy studies of the NTAHis6 bond. Single Mol. 1, 59–65. Kosmulski, M. (2001). “Chemical properties of material surfaces”. Marcel Dekker, New York. Kraemer, S. M., Cheah, S. F., Zapf, R., Xu, J. D., Raymond, K. N., and Sposito, G. (1999). Effect of hydroxamate siderophores on Fe release and Pb(II) adsorption by goethite. Geochim. Cosmochim. Acta 63, 3003–3008. Kraemer, S. M., Xu, J. D., Raymond, K. N., and Sposito, G. (2002). Adsorption of Pb(II) and Eu(III) by oxide minerals in the presence of natural and synthetic hydroxamate siderophores. Environ. Sci. Technol. 36, 1287–1291. Kreller, D. I., Gibson, G., van Loon, G. W., and Horton, J. H. (2002). Chemical Force Microscopy Investigation of Phosphate Adsorption on the Surfaces of Iron(III) Oxyhydroxide Particles. J. Colloid Interf. Sci. 254, 205–213. Kummert, R., and Stumm, W. (1980). The surface complexation of organic acids on hydrous.gamma. alumina. J. Colloid Interf. Sci. 75, 373–385.

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

51

Lawrence, J. R., Korber, D. R., Hoyle, B. D., Costerton, J. W., and Caldwell, D. E. (1991). Optical sectioning of microbial bioﬁlms. J. Bacteriol. 173, 6558–6567. Leckband, D., and Israelachvili, J. (2001). Intermolecular forces in biology. Q. Rev. Biophy. 34, 105 –267. Lee, G. U., Chrisey, L. A., and Colton, R. J. (1994a). Direct Measurement of the Forces between Complementary Strands of DNA. Science 266, 771–773. Lee, G. U., Kidwell, D. A., and Colton, R. J. (1994b). Sensing discrete streptavidin–biotin interactions with atomic force microscopy. Langmuir 10, 354– 357. Liermann, L. J., Kalinowski, B. E., Brantley, S. L., and Ferry, J. G. (2000). Role of bacterial siderophores in dissolution of hornblende. Geochim. Cosmochim. Acta 64, 587–602. Lo, Y. S., Huefner, N. D., Chan, W. S., Stevens, F., Harris, J. M., and Beebe, T. P. (1999). Speciﬁc interactions between biotin and avidin studied by atomic force microscopy using the Poisson statistical analysis method. Langmuir 15, 1373–1382. Lo, Y. S., Simons, J., and Beebe, T. P. (2002). Temperature dependence of the biotin– avidin bondrupture force studied by atomic force microscopy. Journal of Physical Chemistry B 106, 9847–9852. Lower, S. K., Tadanier, C. J., and Hochella, M. F. (2000). Measuring interfacial and adhesion forces between bacteria and mineral surfaces with biological force microscopy. Geochim. Cosmochim. Acta 64, 3133–3139. Lower, S. K., Hochella, M. F., and Beveridge, T. J. (2001a). Bacterial recognition of mineral surfaces: Nanoscale interactions between Shewanella and a-FeOOH. Science 292, 1360–1363. Lower, S. K., Tadanier, C. J., and Hochella, M. F. (2001b). Dynamics of the mineral-microbe interface: Use of biological force microscopy in biogeochemistry and geomicrobiology. Geomicrobiol. J. 18, 63–76. Lower, S. K., Hochella, M. F. Jr., Banﬁeld, J. F., and Rosso, K. (2002). Nanogeoscience: From the movement of electrons to lithosphere plates. Eos: Trans. Am. Geophys. Union 83, 53 –56. Ludwig, C., Casey, W. H., and Rock, P. A. (1995). Prediction of ligand-promoted dissolution rates from the reactivities of aqueous complexes. Nature (London) 375, 44 –47. Maeda, N., and Senden, T. J. (2000). A Method for the Calibration of Force Microscopy Cantilevers via Hydrodynamic Drag. Langmuir 16, 9282–9286. Marszalek, P. E., Oberhauser, A. F., Pang, Y. -P., and Fernandez, J. M. (1998). Polysaccharide elasticity governed by chair-boat transitions of the glucopyranose ring. Nature (London) 396, 661 –664. Maurice, P. A., Lee, Y. J., and Hersman, L. E. (2000). Dissolution of Al-substituted goethites by an aerobic Pseudomonas mendocina var. bacteria. Geochim Cosmochim Acta 64, 1363–1374. Mazzola, L. T., Frank, C. W., Fodor, S. P. A., Mosher, C., Lartius, R., and Henderson, E. (1999). Discrimination of DNA hybridization using chemical force microscopy. Biophys. J. 76, 2922–2933. Merkel, R., Nassoy, P., Leung, A., Ritchie, K., and Evans, E. (1999). Energy landscapes of receptorligand bonds explored with dynamic force spectroscopy. Nature 397, 50– 53. Meyer, G., and Amer, N. M. (1990). Simultaneous measurement of lateral and normal forces with an optical-beam-deﬂection atomic force microscope. Appl. Phys. Lett. 57, 2089–2091. Moy, V. T., Florin, E. L., and Gaub, H. E. (1994a). Intermolecular forces and energies between ligands and receptors. Science 266, 257–259. Moy, V. T., Florin, E.-L., and Gaub, H. E. (1994b). Adhesive forces between ligand and receptor measured by AFM. Colloid Surf., A: Physicochem. Engng. Aspects 93, 343–348. Muller, D. J., and Engel, A. (1997). The height of biomolecules measured with the atomic force microscope depends on electrostatic interactions. Biophy. J. 73, 1633– 1644.

52

T. A. KENDALL AND S. K. LOWER

Muller, D. J., Baumeister, W., and Engel, A. (1999). Controlled unzipping of a bacterial surface layer with atomic force microscopy. Proc. Nat. Acad. Sci. USA 96, 13170– 13174. Myers, C. R., and Nealson, K. H. (1988). Bacterial manganese reduction and growth with manganese oxide as the sole electron acceptor. Science 240, 1319–1321. Myers, C. R., and Nealson, K. H. (1990). Respiration-linked proton translocation coupled to anaerobic reduction of manganese(IV) and iron(III) in Shewanella putrefaciens MR-1. J. Bacteriol. 172, 6232–6238. Myers, C. R., and Myers, J. M. (1992). Localization of cytochromes to the outer membrane of anaerobically grown. Shewanella putrefaciens MR-1. J. Bacteriol. 174, 3429–3438. Myers, C. R., and Myers, J. M. (1993). Ferric reductase is associated with the membranes of anaerobically grown Shewanella putrefaciens MR-1. Fems Microbiol. Lett. 108, 15–22. Myers, C. R., and Myers, J. M. (1997). Outer membrane cytochromes of Shewanella putrefaciens MR-1: Spectral analysis, and puriﬁcation of the 83-kDa c-type cytochrome. Biochim. Biophys. Acta 1326, 307– 318. Myers, J. M., and Myers, C. R. (1998). Isolation and sequence of omcA, a gene encoding a decaheme outer membrane cytochrome c of Shewanella putrefaciens MR-1, and detection of omcA homologs in other strains of S. putrefaciens. Biochim. Biophys. Acta 1373, 237–251. Myers, J. M., and Myers, C. R. (2000). Role of the tetraheme cytochrome CymA in anaerobic electron transport in cells of Shewanella putrefaciens MR-1 with normal levels of menaquinone. J. Bacteriol. 182, 67–75. Myers, J. M., and Myers, C. R. (2001). Role for outer membrane cytochromes OmcA and OmcB of Shewanella putrefaciens MR-1 in reduction of manganese dioxide. Appl. Environ. Microbiol. 67, 260–269. Myers, J. M., and Myers, C. R. (2002). Genetic complementation of an outer membrane cytochrome omcB mutant of Shewanella putrefaciens MR-1 requires omcB plus downstream DNA. Appl. Environ. Microbiol. 68, 2781–2793. Nealson, K. H., and Saffarini, D. (1994). Iron and manganese in anaerobic respiration: Environmental signiﬁcance, physiology, and regulation. Ann. Rev. Microbiol. 48, 311– 343. Neubauer, U., Nowack, B., Furrer, G., and Schulin, R. (2000). Heavy metal sorption on clay minerals affected by the siderophore desferrioxamine B. Environ. Sci. Technol. 34, 2749–2755. Noy, A., Vezenov, D. V., and Lieber, C. M. (1997). Chemical force microscopy. Ann. Rev. Mater. Sci. 27, 381 –421. Nubel, U., Engelen, B., Felske, A., Snaidr, J., Wieshuber, A., Amann, R. I., Ludwig, W., and Backhaus, H. (1996). Sequence heterogeneities of genes encoding 16S rRNAs in Paenibacillus polymyxa detected by temperature gradient gel electrophoresis. J. Bacteriol. 178, 5636–5643. Ohmura, N., Kitamura, K., and Saiki, H. (1993). Selective adhesion of Thiobacillus ferrooxidans to pyrite. Appl. Environ. Microbiol. 59, 4044–4050. Ong, Y.-L., Razatos, A., Georgiou, G., and Sharma, M. M. (1999). Adhesion forces between E. coli bacteria and biomaterial surfaces. Langmuir 15, 2719–2725. Oreskes, N., Shrader-Frechette, K., and Belitz, K. (1994). Veriﬁcation, validation, and conﬁrmation of numerical models in the Earth Sciences. Science 263, 641– 646. Pleuddemann, E. P. (1991). “Silane Coupling Agents”. Plenum Press, New York. Rabinovich, Y. I., and Yoon, R. H. (1994). Use of atomic force microscope for the measurements of hydrophobic forces between silanated silica plate and glass sphere. Langmuir 10, 1903–1909. Razatos, A., Ong, Y. L., Sharma, M. M., and Georgiou, G. (1998a). Evaluating the interaction of bacteria with biomaterials using atomic force microscopy. J. Biomater. Sci. Polym. Edu. 9, 1361–1373. Razatos, A., Ong, Y.-L., Sharma, M. M., and Georgiou, G. (1998b). Molecular determinants of bacterial adhesion monitored by atomic force microscopy. Proc. Nat. Acad. Sci. USA 95, 11059– 11064.

FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES

53

Rezania, A., Johnson, R., Lefkow, A., and Healy, K. E. (1999). Bioactivation of metal oxide surfaces. 1. Surface characterization and cell response. Langmuir 15, 6931–6939. Rief, M., Gautel, M., Oesterhelt, F., Fernandez, J. M., and Gaub, H. E. (1997a). Reversible unfolding of individual titin immunoglobulin domains by AFM. Science (Washington, D.C.) 276, 1109–1112. Rief, M., Oesterhelt, F., Heymann, B., and Gaub, H. E. (1997b). Single molecule force spectroscopy on polysaccharides by atomic force microscopy. Science (Washington, D.C.) 275, 1295–1297. Roden, E. E., and Zachara, J. M. (1996). Microbial reduction of crystalline iron(III) oxides: Inﬂuence of oxide surface area and potential for cell growth. Environ. Sci. Technol. 30, 1618–1628. Rugar, D., Mamin, H. J., and Guethner, P. (1989). Improved ﬁber-optic interferometer for atomic force microscopy. Appl. Phys. Lett. 55, 2588–2590. Ryan, J. N., and Gschwend, P. M. (1994). Effects of ionic-strength and ﬂow-rate on colloid release — relating kinetics to intersurface potential-energy. J. of Colloid and Interface Science 164, 21–34. Ryan, J. N., and Elimelech, M. (1996). Colloid mobilization and transport in groundwater. Colloid Surf. a-Physicochem. Eng. Aspects 107, 1–56. Sader, J. E., and White, L. (1993). Theoretical analysis of the static deﬂection of plates for atomic force microscope applications. J. Appl. Phys. 74, 1–9. Sader, J. E., Larson, I., Mulvaney, P., and White, L. R. (1995). Method for the calibration of atomic force microscope cantilevers. Rev. Sci. Instrum. 66, 3789– 3798. Sader, J. E. (1998). Frequency response of cantilever beams immersed in viscous ﬂuids with applications to the atomic force microscope. J. Appl. Phys. 84, 64–76. Sader, J. E., Chon, J. W. M., and Mulvaney, P. (1999). Calibration of rectangular atomic force microscope cantilevers. Rev. Sci. Instrum. 70, 3967–3969. Schmitt, L., Ludwig, M., Gaub, H. E., and Tampe, R. (2000). A metal-chelating microscopy tip as a new toolbox for single-molecule experiments by atomic force microscopy. Biophy. J. 78, 3275–3285. Schultze-Lam, S., Harauz, G., and Beveridge, T. J. (1992). Participation of a cyanobacterial S-layer in ﬁne-grain mineral formation. J. Bacteriol. 174, 7971–7981. Schwesinger, F., Ros, R., Strunz, T., Anselmetti, D., Guntherodt, H. J., Honegger, A., Jermutus, L., Tiefenauer, L., and Pluckthun, A. (2000). Unbinding forces of single antibody-antigen complexes correlate with their thermal dissociation rates. Proc. Nat. Acad. Sci. USA 97, 9972–9977. Seaman, J. C., Alexander, D. B., Loeppert, R. H., and Zuberer, D. A. (1992). The availability of iron from various solid-phase iron sources to a siderophore producing pseudomonas strain. J. Plant Nutr. 15, 2221– 2233. Sposito, G. (1989). “The Chemistry of Soils”. Oxford University Press, New York. Stevens, F., Lo, Y. S., Harris, J. M., and Beebe, T. P. (1999). Computer modeling of atomic force microscopy force measurements: Comparisons of Poisson, histogram, and continuum methods. Langmuir 15, 207–213. Stone, A. T. (1997). Reactions of extracellular organic ligands with dissolved metal ions and mineral surface. “Geomicrobiology: Interactions Between Microbes and Minerals” (K. H. Nealson, Ed.), Vol. 35, pp. 309–341. Mineralogical Society of America, Washington D.C. Strunz, T., Oroszlan, K., Schafer, R., and Guntherodt, H.-J. (1999). Dynamic force spectroscopy of single DNA molecules. Proc. Nat. Acad. Sci. USA 96, 11277– 11282. Stumm, W. (1992). “Chemistry of the Solid-Water Interface”. Wiley, Inc., New York. Swamy, M. J. (1995). Thermodynamic analysis of biotin binding to avidin. A high sensitivity titration calorimetric study. Biochem. Mol. Biol. Int. 36, 219 –225. Telford, J. R., and Raymond, K. N. (1996). Siderophores. Compr. Supramol. Chem. 1, 245– 266. Toikka, G., Hayes, R. A., and Ralston, J. (1996). Surface forces between spherical ZnS particles in aqueous electrolyte. Langmuir 12, 3783– 3788.

54

T. A. KENDALL AND S. K. LOWER

Tokunaga, M., Aoki, T., Hiroshima, M., Kitamura, K., and Yanagida, T. (1997). Subpiconewton intermolecular force microscopy. Biochem. Biophys. Res. Commun. 231, 566–569. Torii, A., Sasaki, M., Hane, K., and Okuma, S. (1996). A method for determining the spring constant of cantilevers for atomic force microscopy. Meas. Sci. Technol. 7, 179– 184. van Loosdrecht, M. C. M., Lyklema, J., Norde, W., and Zehnder, A. J. B. (1989). Bacterial adhesion: A physicochemical approach. Microbial. Ecol. 17, 1– 15. van Oss, C. J. (1993). Acid-base interfacial interactions in aqueous media. Colloid Surf. A: Physicochem. Engg. Aspects 78, 1 –49. Verwey, E. J., and Overbeek, J. T. G. (1948). “Theory of the Stability of Lyophobic Colloids”. Elsevier Publishing, Amsterdam. Vigeant, M. A.-S., Ford, R. M., Wagner, M., and Tamm, L. K. (2002). Reversible and irreversible adhesion of motile Escherichia coli cells analyzed by total internal reﬂection aqueous ﬂuorescence microscopy. Appl. Environ. Microbiol. 68, 2794–2801. Voet, D., and Voet, J. (1995). “Biochemistry”, 2nd edn. John Wiley & Sons, New York. Wagner, P. (1998). Immobilization strategies for biological scanning probe microscopy. FEBS Lett. 430, 112 –115. Watteau, F., and Berthelin, J. (1994). Microbial dissolution of iron and aluminum from soil minerals — efﬁciency and speciﬁcity of hydroxamate siderophores compared to aliphatic-acids. Eur. J. Soil Biol. 30, 1– 9. Weisenhorn, A. L., Maivald, P., Butt, H. J., and Hansma, P. K. (1992). Measuring Adhesion, Attraction, and Repulsion Between Surfaces in Liquids With an Atomic-Force Microscope. Phys. Rev. B 45, 11226–11232. Wenzler, L. A., Moyes, G. L., Olson, L. G., Harris, J. M., and Beebe, T. P. (1997). Single molecule bond-rupture force analysis of interactions between AFM tips and substrates modiﬁed with organosilanes. Analytical Chemistry 69, 2855–2861. Whitman, W. B., Coleman, D. C., and Wiebe, W. J. (1998). Prokaryotes: The unseen majority. Proc. Nat. Acad. Sci. USA. 95, 6578–6583. Williams, J. M., Han, T. J., and Beebe, T. P. (1996). Determination of single-bond forces from contact force variances in atomic force microscopy. Langmuir 12, 1291–1295. G. Winkelmann, (Ed.), (1991). In “CRC Handbook of Microbial Iron Chelates” CRC Press, Boca Raton, Florida. Wong, S. S., Joselevich, E., Wooley, A. T., Cheung, C. L., and Lieber, C. M. (1998a). Covalently functionalized nanotubes as nanometre-sized probes in chemistry and biology. Nature 394, 52–55. Wong, S. S., Woolley, A. T., Joselevich, E., Cheung, C. L., and Lieber, C. M. (1998b). Covalentlyfunctionalized single-walled carbon nanotube probe tips for chemical force microscopy. J. of American Chemical Society 120, 8557–8558. Wong, J., Chilkoti, A., and Moy, V. T. (1999). Direct force measurements of the streptavidin–biotin interaction. Biomol. Engg. 16, 45 –55. Yamamoto, S.-i., Yamada, H., and Tokumoto, H. (1997). Precise force curve detection system with a cantilever controlled by magnetic force feedback. Rev. Sci. Instrum. 68, 4132–4136. Yao, H.-L., and Yeh, H.-H. (1996). Fumarate, Maleate, and Succinate Adsorption on Hydrous.delta.A12O3. 1. Comparison of the Adsorption Maxima and Their Signiﬁcance. Langmuir 12, 2981–2988.

RICE FUNCTIONAL GENOMICS: LARGE-SCALE GENE DISCOVERY AND APPLICATIONS TO CROP IMPROVEMENT Hei Leung1 and Gynleung An2 1

International Rice Research Institute, DAPO Box 7777, Metro Manila, Philippines 2 Pohang University of Science and Technology, San 31 Hyoja-dong, Nam-gu, Pohang, Korea 790-784 I. Introduction II. Rice as a Model Genetic System A. Diversity of Oryza B. Grass as a Single Genetic System III. Rice Genome A. Genome Sequencing B. Salient Features of the Rice Genome C. Comparison Between Indica and Japonica Rice IV. Key Ingredients for Gene Discovery in Rice A. Genetic Resources B. High-throughput Technologies C. Biological Evaluation D. Bioinformatics V. Forward and Reverse Genetics A. Insertional Mutants B. Chemical- and Irradiation-induced Mutants C. Natural Genetic Variation VI. Functional Validation of Rice Genes A. Gene Expression Analysis B. Gene Silencing C. Heterologous Bioassays D. Gene Replacement E. Allelic Series VII. Applications to Crop Improvement A. Candidate Gene Approach and Allele Mining B. Pathways and Genetic Regulation C. Cross-species Inference of Gene Function VIII. International Collaboration and the Role of Developing Countries IX. Concluding Remarks Acknowledgments References

55 Advances in Agronomy, Volume 82 Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved. 0065-2113/03 $35.00

56

H. LEUNG AND G. AN

I. INTRODUCTION Rice has an old and rich cultivation history. Archeological ruins found in China and India suggest that domestication of rice might date back 7000 years (Watanabe, 1997). Several regions in Asia, including Assam in India, Myanmar, northern Thailand, and southwest and southern China, have been hypothesized as the origins of cultivated rice. Also, rice cultivation probably originated independently in several locations (Matsuo, 1997). Rice is more than just a food crop. As a key sustenance for people since ancient times, rice has become part of the social and religious fabric of many cultures (see White 1994 for an interesting popular article about rice culture in National Geographic). Today, rice remains the staple food of more than half the world population and supplies up to 60% of the calories in the diet of people living in Asia. In many developing countries, consumers spend more than 20% of disposable income just to buy sufﬁcient rice to sustain daily food needs. Rice cultivation is also a major source of employment in rural areas. The rice economy — production, distribution, and consumption — therefore plays an important role in the livelihood of many people. Rice is cultivated in diverse ways. Rice production is often classiﬁed according to the ecosystem in which it is grown. In terms of area, a majority (58%) of rice is grown in the irrigated environment, commonly called paddy rice. This is a highly intensive rice production environment, which accounts for about 75% of the rice produced worldwide. Most investment in production technologies has been applied to this environment. The remaining portion of rice is grown in the rainfed environment, that is, on land that relies on natural rainfall. About half of the rainfed rice is grown in areas with a reasonably reliable water supply (the favorable rainfed area), but the other half is grown in very difﬁcult conditions with too little or too much water and poor soil conditions. Figure 1 summarizes yield trends of rice over the course of the Green Revolution that started more than four decades ago in response to the food crisis in many parts of the developing world (Hossain et al., 2000). As a result of concentrated efforts devoted to improving rice germplasm and agronomy, a dramatic and steady increase in rice productivity has occurred across Asia (Hossain, 1998). Rice yield in irrigated areas nearly doubled from 1960 to 1980. This has resulted in lower rice prices that beneﬁt consumers. The yield increase, however, is low in unfavorable environments where the water supply is unreliable and the soil is often infertile or problematic for rice production. Recognizing this historical trend is relevant in the context of applying genomics to crop improvement. The trend enables us to identify the key areas where advances in science can help solve intractable problems. In the highly intensive system, technologies are needed to increase production efﬁciency with less cost. The system also needs improved resilience against pests and diseases to reduce

RICE FUNCTIONAL GENOMICS

57

Figure 1 Yield trend of rice from 1967 to 1997 in three rice production environments (modiﬁed from Hossain et al., 2000).

reliance on chemical inputs. In the unfavorable rainfed environment, there is plenty of room to improve productivity and the nutritional value of germplasm to meet the need of the rural poor. Although rice is such an important food crop, it has only a relatively short history as a model organism for biological investigation. Before the 1980s, rice research — from basic to applied — was mostly done by countries where rice is the staple crop and where sufﬁcient infrastructure existed to conduct research. This limited the majority of rice research to a few countries or international agricultural research centers where varietal improvement is a mandate or a national necessity. There was hardly any private interest in rice research. In many respects, rice was an “orphan” crop in terms of research investment. This picture changed in the mid-80s when advances made in plant molecular biology and gene transfer techniques were seen as promising avenues to convert biological discoveries to applications. The Rockefeller Foundation, a long-time supporter of international agricultural development, made a strategic decision to support the application of molecular biology to solve rice production problems. The beginning of the International Rice Biotechnology Program in 1985 launched the “molecular revolution” in rice science. For the next 16 years (1985 – 2001), we witnessed the blossoming of basic research in rice. The sustained support of rice biotechnology produced a large cadre of scientists in the developing world, well trained and ready to apply useful genes in production (O’Toole et al., 2001).

58

H. LEUNG AND G. AN

While gene identiﬁcation and cloning were advanced toward the end of the Rockefeller Biotechnology Program, there was a growing sense that many biological problems limiting rice production required a comprehensive understanding of multiple genes and their interactions with the environment. In the last Rice Biotechnology meeting held in Phuket, Thailand, in 1998, questions were raised about the next quantum leap in rice productivity. What would the investment in research capacity and infrastructure in the rice-growing world do to future rice production? Around this time, the genome of the ﬁrst eukaryote — Saccharomyces cerevisiae — was completed (Williams, 1996), followed by several initiatives to sequence other model organisms — Drosophila, nematode, Arabidopsis, and human. These events marked a genomics decade that has put rice at the center of plant science research and provided a logical extension of the Rockefeller Foundation Rice Biotechnology Program. Genomics can be broadly deﬁned as the understanding of all genes and their encoded biological function in an organism in totality. It has emerged as an integrative discipline that provides the means to understand biological complexity and to solve difﬁcult practical problems. Rice, with the smallest genome size among major cereals and crop plants, is seen as the obvious target for genomics research. The Rockefeller Foundation sponsored a meeting in 1997 at the International Congress of Plant Molecular Biology in Singapore to discuss the merits and feasibility of sequencing the entire rice genome. At that time, physical mapping using YAC and BAC libraries was progressing rapidly, producing the critical resources for sequencing. However, with the available sequencing technologies at that time, completing the rice genome would cost about $200 million over 10 years, a formidable undertaking for any individual institute or program. Thus, a sensible way to achieve the task was to launch an international collaboration. This led to the formation of an International Rice Genome Sequencing Project (IRGSP) in 1998 under the coordination of the Rice Genome Program of Japan. The IRGSP represents a consortium of multiple countries collaborating to sequence the rice genome. In the initial stage, over 10 countries pledged to take part to sequence individual chromosomes. The IRGSP was a unique undertaking that has galvanized government commitment and it subscribed to the principle of releasing all sequence information to the public. Since then, rice genome sequencing has been the topic of scientiﬁc drama and excitement involving the private and public sector (Normile and Pennisi, 2002). To a large extent, the public initiative has been a key factor to leverage the contribution of rice genome data from the private sector. Together, this has led to the availability of genome sequence of two rice subspecies ( japonica and indica), a favorable situation for public researchers working not only on rice but on other plant species. With a genetic blueprint in place, we ask, what are the new opportunities to understand gene functions and their interactions? What are the most efﬁcient approaches to ﬁnd genes of agronomic importance? Why is international

RICE FUNCTIONAL GENOMICS

59

collaboration necessary for gene discovery in rice? How can we turn gene discovery into practical products? This review is organized with these questions in mind. In particular, we hope to lay out the framework for large-scale gene discovery that will beneﬁt the developing world, where access to knowledge and tools is often limited. For more discussion of rice gene function analyses and comparative genomics, readers are referred to a recent review by Shimamoto and Kyozuka (2002). In this chapter, we will begin with an introduction of the rice plant as a model species. We will discuss the ingredients needed for large-scale gene discovery and the approaches that will beneﬁt the improvement of traits needed in rice production. We will discuss the importance of producing public resources and the need for international collaboration, and how gene discovery can be integrated into plant improvement programs.

II. RICE AS A MODEL GENETIC SYSTEM Rice has emerged as a model in plant science research because of its many positive attributes for genetic studies. Rice has a simple genetic system (diploid and disomic inheritance) and one of the most amenable transformation systems among grasses (Hiei et al., 1994; Christou, 1997; Cheng et al., 1998). The small genome size and relatively low amount of repetitive sequences make rice a tractable system for whole-genome sequencing. Other than these technical advantages, an important, but perhaps less appreciated attribute is the great store of genetic diversity in the genus Oryza. With a long history of cultivation and selection, and a large and heterogeneous geographical niche, the wealth of phenotypic and genetic variation in all aspects of plant life, particularly genes controlling agronomically important traits, is enormous in rice. Below, we highlight the genetic diversity available in rice that is relevant for functional genomics research.

A. DIVERSITY

OF

ORYZA

Taxonomically, rice is in the subfamily Ehrhartoideae (former Oryzoideae) of the grass family Poaceae (Gramineae). Figure 2 shows the overall relationship of rice to other cereals and the recognized genomes within the genus Oryza (Katayama, 1997; Ge et al., 1999; Lu, 1999; Ge et al., 2001). The most widely cultivated rice, O. sativa, has the AA genome. Another cultivated rice species, O. glaberrima Steud., with the AA genome is grown only in West Africa (for an interesting anthropological perspective of African rice, see Linares, 2002). Both O. sativa and O. glaberrima have 12 chromosomes as their basic complement

60

H. LEUNG AND G. AN

Figure 2 Diversity of rice as part of the broad gene pool of the grass families. Dendrograms depict the relationships among cereal crops in the grass family (Kellogg, 1998), domesticated and wild rice with different genomes (from Ge et al. 1999), and cultivated rice species. Dendrograms are not drawn to scale.

(2n ¼ 24). F1 hybrids between the two species can be produced, but advancing beyond F2 generation is difﬁcult because of sterility. Nonetheless, the small amount of fertility has allowed the successful transfer of useful genes from O. glaberrima into O. sativa (Jones et al., 1997). The cultivated O. sativa has two subspecies — indica and japonica — that provide the main gene pool for rice breeding. There is a third minor group called “tropical japonica” (also called javanica) that is common in Southeast Asia. Japonica rice is generally adapted to the temperate area, whereas indica rice is planted in the tropics. Indica rice is the most widely grown type, constituting about 80% of world rice production. Chang et al. (1991) summarized the main agronomic characteristics of the main rice types (Table I). There is much overlap in agronomic characteristics between the two subspecies. Whether indica and japonica are monophyletic or diphyletic remains a research question (Morishima, 2001). Although divergence of these subspecies has occurred over time, sufﬁcient fertility exists to allow routine cross hybridization and gene ﬂow between the two subspecies. Also, because of their divergence, a high level of hybrid heterosis can be harnessed from indica £ japonica crosses. Such extra hybrid vigor is being exploited in hybrid rice production (Virmani and Ilyas-Ahmed, 2001). Now that the genome sequence of these two subspecies is available, this will open up opportunities to understand the molecular basis of japonica £ indica sterility as well as hybrid heterosis. There are more than 30 named Oryza species (Watanabe, 1997) with a broad geographic distribution. Table II lists some of the species that have been used in rice improvement programs or genetic studies (Brar and Khush, 2002). This is by no means a comprehensive list of all known wild rice species but it illustrates the diversity of the domesticated and wild germplasm that forms the basis for gene

RICE FUNCTIONAL GENOMICS

61

Table I Diverse Morphological and Physiological Features of indica, japonica, and Tropical japonica ( javanica) Rice Reﬂect a Broad Genetic Base for Gene Discovery Rice type

Characteristics Leaf morphology Grain Tillering Plant stature Awn Lemma and palea Shattering Plant tissue Photosensitivity Amylose Gelatinization temperature

Indica

Japonica

Tropical japonica (javanica)

Broad to narrow, light green Long to short, slender, ﬂat High Tall to intermediate Mostly awnless Thin, short hair Easy Soft Varying 23–31% Varying

Narrow, dark green Short, roundish Medium Short to intermediate Awnless to long-awned Dense, long hair Low Hard Zero to low sensitivity 10 –24% Low

Broad, stiff, light green Long, broad, thick Low Tall Long-awned or awnless Long hair Low Hard Low sensitivity 20–25% Low

Adapted from Chang et al. (1991).

discovery. The wild species that have evolved in diverse, heterogeneous, and often harsh environments are potentially a rich pool of useful genes for breeding. Wild rice with a non-AA genome is not interferile with cultivated rice, thus making gene transfer between these wild species difﬁcult, though not impossible, through intervention with embryo rescue and tissue culture techniques (Brar and Khush, 1997). There are examples of useful genes that have been extracted from wild germplasm for breeding (Brar and Khush, 2002). One well-known example is the grassy stunt resistance gene obtained from O. nivara. This resistance gene has been used in nearly all rice cultivars developed for the tropical lowland and it remains effective in conditioning resistance to the dominant strain of the grassy stunt virus until now (Khush, 1977). The ﬁrst cloned disease resistance gene, Xa21, was also originally introgressed from the wild rice O. longistaminata from an accession in Mali, Africa. After many years of genetic and breeding work, Xa21 was put into an agronomically acceptable background that has facilitated its molecular cloning and use in rice production (Song et al., 1995). Yet, one may also argue that the number of genes extracted for use represents only a handful of the potential useful genes in the wild species. Tanksley and McCouch (1999) popularized the use of unadaptive wild germplasm through systematic introgression into elite backgrounds. However, the large-scale introgression of wild species alleles into domesticated rice is limited by the laborious process of embryo rescue. This picture may change if we have a way to predetermine

62

H. LEUNG AND G. AN Table II Reservoir of Useful Genes and Alleles in Domesticated and Wild Rice Germplasm

Genome

Speciesa

AA AA AA AA AA

O. sativa O. glaberrima O. barthii O. nivara O. ruﬁpogon

AA AA AA BB, BBCC BBCC CC

O. longistaminata O. glumaepatula O. meridionalis O. punctata O. minuta O. ofﬁcinalis

CC CC CCDD CCDD CCDD EE FF HHKK HHJJ

O. rhizomatis O. eichingeri O. alta O. latifolia O. grandiglumis O. australiensis O. brachyantha O. schlechteri O. longiglumis

HHJJ GG GG GG

O. ridleyi O. granulata O. meyeriana O. neocaledonica

Distributionb

Number of accessions at IRRI Gene Bankc

Worldwide West Africa Africa Tropical, Subtropical Asia Tropical, Subtropical Asia, Tropical Australia Africa South, Central America Tropical Australia Africa Philippines, Papua New Guinea Tropical, Subtropical Asia, Tropical Australia Sri Lanka South Asia, East Africa South, Central America South, Central America South, Central America Tropical Australia Africa Papua New Guinea Indonesia (Irian Jaya), Papua New Guinea South Asia South, Southeast Asia Southeast Asia New Caledonia

91,963 1543 218 1258 1048 206 54 56 71 64 279 19 30 6 58 10 36 19 1 6 15 24 11 1

a

Lu (1999). Brar and Khush (2002). c Information from IRGCIS database as of February 7, 2003, http://grcsvr4/irgc/main.htm. b

the potential value of speciﬁc alleles in useful donors. The application of reverse genetics to determine the allelic state of candidate genes can be an attractive strategy to select wild germplasm for introgression. How to tap this genetic potential is the core agenda of functional genomics.

B. GRASS

AS A

SINGLE GENETIC SYSTEM

The grass family diverged from the dicotyledon (dicot) lineage about 200 million years ago (Wolfe et al., 1989; Crane et al., 1995). Within the grass family,

RICE FUNCTIONAL GENOMICS

63

an estimated . 10,000 species diverged from each other about 70 million years ago (Kellogg, 2001). During this period of evolution, there has been an expansion and contraction of the ancestral genome estimated to be about 3.5 pg/2C nucleus. Kellogg (1998) examined the DNA contents of representative species in relation to the phylogenetic tree of four subfamilies of grasses that contain the cereals and showed that the genome size had expanded in some but contracted in others (see Fig. 4 in Kellogg 1998). The Pooideae, which contains wheat, barley, rye, and oats, has a large genome (10 pg DNA per 2C nucleus), whereas members of the Ehrhartoideae have a generally smaller genome (3.3 pg). The large difference in genome size is attributed, to a large extent, to the expansion of repetitive sequences. The large genome size has been an impediment to the systematic sequencing of certain cereals, although progress has been made in identifying gene-rich regions in large genomes (Gill et al., 1996; Farris et al., 2000). That the grasses belong to a monophyletic group and all share a set of common “grass alleles” argues favorably for using a few reference species to avoid the difﬁculties (and expenditure involved) in sequencing individual species (Freeling, 2001). The concept of grass as a single genetic system rests on two main premises. First, the gene order on a chromosome is conserved in different related genomes. Second, there is sufﬁcient conservation of gene content across species, and that function can be inferred by sequence structure. In dicots, Bonierbale et al. (1988) ﬁrst demonstrated that tomato cDNA probes could be used to establish an RFLP linkage map in potato that is colinear with that in tomato. In monocots, Ahn and Tanksley (1992) used a common set of probes to show conserved linkage in maize and rice. Since then, comparative mapping in both dicots and monocots has led to the general conclusion that the gene order of related plant species is sufﬁciently conserved even after millions of years of evolution, making it possible to infer the location of a particular gene based on linkage information from another species (Devos and Gale, 2000). In most cases, colinearity between plant species can be demonstrated at the macro chromosomal level. There are, however, exceptions to colinearity due to the accumulation of chromosomal changes over millions of years of evolution. Such exceptions are particularly common at the micro level because of inversion, translocation, insertion, gene duplications, and deletions. A distinction between macro- and micro-colinearity is important because success in ﬁnding a gene through a syntenic relationship depends on the maintenance of micro-colinearity between related species. Freeling (2001) offers a useful clariﬁcation of the terms “colinearity” and “synteny”. Macro-colinearity is inferred when the same marker orders are observed in a typical mapping experiment with 10 — 20 cM resolution. Occasional chromosomal rearrangements may break up this gene order, but the pair of chromosomes in question is considered to be syntenic to each other. Thus, even for related species that do not have perfect macro-colinearity, they can still be syntenous.

64

H. LEUNG AND G. AN

However, macro-colinearity may or may not enable one to predict gene position at a ﬁne scale. For a syntenic relationship to be useful for ﬁnding a gene, we need markers with a high probability of ﬁnding the same adjacent markers in another species. Currently, the number of successful cases of gene isolation strictly based on positional relationships is still too few. The predictive power of the syntenic relationship has not yet been vigorously tested (Gaut, 2002). The exceptions in micro-colinearity, however, by no means argue against the use of grass as a single genetic system (Bennetzen and Freeling, 1997; Freeling, 2001). With the whole-genome sequence available, the beneﬁt of using a model species to ﬁnd a gene will be based more on gene sequence conservation rather than on map locations. For example, a candidate gene for disease resistance in barley can be used to ﬁnd its orthologs in rice or vice versa without knowledge of map location. Thus, map locations are useful data, but they are not a prerequisite for ﬁnding orthologous genes to evaluate function. In the future, the syntenic relationship will be more relevant for mapping unique traits in a species with little sequence information. For example, drought tolerant traits are common in many millets which do not have detailed sequence information. In such a case, moderate investment in mapping in the “trait-bearing” crops can lead to identiﬁcation of candidate genes in a species with a detailed sequence map. In summary, the original discovery of a colinear relationship between markers in different species triggered a large number of studies on the extent of gene order conservation. While macrosynteny (10 – 20 cM range) has usually been supported, evidence of microsynteny (within 1 cM, or in the 100-kb range) is more difﬁcult to obtain. Rearrangements, insertions, and deletions appear to have disrupted the syntenic relationship, which potentially leads to futile pursuit (Feuillet and Keller, 2002; Li and Gill, 2002). Notwithstanding these exceptions, comparative analyses of genome organization and gene function within the grasses will elevate the studies of genome evolution and practical applications to a new level. With the completion of the rice genome sequence, a comparative genomic approach can now be pursued with even greater intensity.

III. RICE GENOME Not long ago, sequencing the genome of an organism was a daunting proposition. However, with advances in sequencing technologies, the cost of sequencing a genome has dropped dramatically. For the past decade, genome sequencing projects, either completed or begun, have produced such excitement in biological sciences that many young students have been attracted to the discipline of genomics. The complete genome sequence represents a public resource that opens up opportunities for new understanding of biological systems.

RICE FUNCTIONAL GENOMICS

65

With sequence available, identiﬁcation and isolation of a gene can be largely done electronically or by simple PCR ampliﬁcation, making gene cloning in the conventional sense something of the past. This may prove to be one of the biggest returns from the investment in sequencing projects.

A. GENOME SEQUENCING As of January 2003, four versions of rice genome sequences exist: three draft versions produced by Monsanto, Syngenta, and the Beijing Genomics Institute (BGI) and a highly accurate version produced by the IRGSP (for a summary, see Buell, 2002). The Monsanto and Syngenta sequences, originally produced by independent efforts, have now been integrated with the IRGSP sequence. The BGI sequence is from indica rice and has now been improved from the initial 4X draft to a higher-quality 6X draft (Yu Jun, personal communication). Different versions of the genome sequence serve different purposes and require different levels of investment. The Syngenta and BGI genome sequences were produced by a “shot gun” approach, whereas the IRGSP sequence used a map-based approach. In shot-gun sequencing, the whole genome is sheared into small pieces (1 – 2 kb), subcloned in plasmids, and sequenced. On average, each sequence is analyzed 4 to 6 times. Then the sequences are assembled by a computer program. In high eukaryotes, the high proportion of repetitive sequences can obscure the alignment of truly contiguous sequences. Computationally demanding programs are therefore needed to solve this problem to obtain a reliable assembly. The shotgun sequencing approach has been successfully applied to the complex genomes of human and Drosophila. Depending on the objectives, the shot-gun approach has the advantage of producing a draft sequence quickly and allowing researchers to identify similar sequences in different species. On the other hand, there are compelling reasons for developing a high-quality sequence using a map-based approach. Here, each stretch of sequence is physically anchored to a chromosome. At the outset, the IRGSP committed to sequencing the rice genome at an accuracy of , 1 error per 10,000 bases to provide the gold standard for accurate sequence annotation and for detecting natural variation. On December 18, 2002, the IRGSP announced the completion of the “phase 2” sequence data with all contigs anchored on the 12 rice chromosomes. As of February 2003, 94% of the genome is covered, spanning a genetic length of 1530 cM. By deﬁnition, ﬁnishing the rice genome means that: (1) all sequences are physically aligned to the chromosome map, (2) there are no gaps, and (3) all sequences are annotated. Currently, chromosomes 1, 3, 4, 10 are considered to be ﬁnished (phase 3). It is expected that a completely ﬁnished genome sequence will be accomplished by the end of 2004.

66

H. LEUNG AND G. AN

B. SALIENT FEATURES

OF THE

RICE GENOME

Although sequencing of some chromosomes remains to be ﬁnished, the existing sequence information together with the completed Arabidopsis genome (The Arabidopsis Genome Initiative, 2000) give us a good picture of the structure and organization of the rice genome (Table III). For details, readers are referred to the original papers by Goff et al. (2002), Feng et al. (2002), Sasaki et al. (2002), and Yu et al. (2002). The rice genome is predicted to contain a large number of genes with an upper estimate as high as 62,000. On average, a gene can be found every 5– 7 kb, a density similar to that observed in Arabidopsis. The functional classiﬁcation of rice genes is similar to the Arabidopsis sequence [see Fig. 9 in Yu et al. (2002) and Fig. 1 in Goff et al. (2002)]. About 30% of the rice genes are plant-speciﬁc. There are signs that rice experienced many duplication events at the level of genes or chromosomal segments. To determine the extent of gene duplication, Goff et al. (2002) examined homology among all predicted genes (longer than 300 bp) using the Syngenta draft sequence and found that 77% were homologous to at least another gene, yielding about 15,000 gene families, a picture similar to that observed in Arabidopsis and other eukaryotes. Local duplications (i.e., homologous genes mapped to a single BAC contig) range from 15.4 to 30.4% depending on the chromosomes. The authors concluded that a genome duplication event might have occurred in rice 40– 50 million years ago and that chromosomes 11 and 12 were probably a result of duplication about 25 million years ago. The extent of gene or chromosomal duplication has implications for the strategy for assigning function to each gene. If gene redundancy is widespread, simple knockouts at individual genes would not produce detectable phenotypes. However, it is too early to tell the extent to which structural duplication is translated into functional redundancy. Genes that are structurally similar may have diverged to provide developmental and tissue speciﬁc expression. We have yet to determine what proportion of the duplicated genes is indeed functionally equivalent. Also, the number of gene families in tight clusters will affect the way we conﬁrm the function of a candidate gene. When we encounter a cluster of genes with similar sequences, how can we determine whether the QTL effect is due to a single member of the cluster or a collective effect of multiple members? For these reasons, better gene annotation and understanding of the functional relationship of paralogous and orthologous genes and members within gene families are important. While Arabidopsis and rice share common characteristics of plant genomes, there are signiﬁcant differences. The average size of rice genes is larger than that of Arabidopsis. The larger gene size in rice is apparently due to larger introns. Thus, the transcript size of a rice gene is not much larger than that of Arabidopsis (Sasaki et al., 2002). The rice genome also has a high CG content.

Table III Features and Composition of the Rice Genome Revealed by Different Versions of the Rice Genome Sequencea Finished genome sequence Chromosome 1 of Nipponbare

Features

Chromosome 4 of Nipponbare

Draft Sequence Nipponbareb

Indica 93.11c

Arabidopsis thaliana

3.4 6.4 4.8 230 3.8 605 58.2

2.7 7.4 4.4 340 – 370 47.2

2.4 5.7 4.2 296 3.2 371 54.9

4.5 4.5 3.3 201 2.3 356 51.4

2.4 4.5 5.2 250 4.2 170 44.1

40.7

42.3

38.9

37.0

32.7

43.8 214 kb/cM for short arm; 288 kb/cM for long arm

44.0 253 kb/cMd

43.3 –

34.9 213 kb/cMd

Estimated genome size (Mb) Estimated gene number for whole genome

– 62,500

44.2 208 kb/cM for euchromatic region; 636 kb/cM for heterochromatic region – 57,000

420 32,000– 50,000

466 46,000– 55,000

125 25,498

RICE FUNCTIONAL GENOMICS

Gene size (kb) Gene density (kb/gene) No. exon/gene Exon size (bp) No. intron/gene Intron size (bp) CG content in coding region (%) CG content in non-coding region (%) CG content overall (%) Physical distance/genetic distance

a Compiled from published data in Yu et al., 2002, Goff et al., 2002, Feng et al., 2002, Sasaki et al., 2002, and The Arabidopsis Genome Initiative (2000). – ¼ no information provided in these papers. b TIGR IRGSP annotation statistics as of 23 Jan 2003 (www.tigr.org/tdb/e2k1/osa1/riceInfo/info.shtml#Genes) c Beijing Genomics Institute published draft (4X genome coverage), Wong et al. (2002). d For rice, assuming a genetic length of 1699 cM and a genome size of 430 Mb; for Arabidopsis, assuming a genetic length of 586 cM and genome size of 125 Mb.

67

68

H. LEUNG AND G. AN

Wong et al. (2002) observed a negative gradient of CG content starting from the 50 end to 30 end of the coding region. This CG gradient may reﬂect codon usage, and hence may inﬂuence the accuracy of gene prediction as most gene prediction programs do not take into account such a bias. It has been suggested that the large difference between the number of predicted genes in rice and Arabidopsis may in part be caused by artifacts of gene prediction. Nearly all Arabidopsis genes can be found in rice but not vice versa. Based on the draft sequences of indica and japonica and the completed chromosomes 1 and 4, nearly 50% of the predicted rice genes have no known function in Arabidopsis. Yu et al. (2002) suggested that the rice genome represents a “superset” of plant genes, only part of which is present in Arabidopsis. Although part of the difference could be due to imperfect sequence annotation, rice probably has many unique genes that are diverged from Arabidopsis genes. Since a large sequence dataset is not yet available in other cereals, a pairwise genome comparison with other cereals is not yet possible. Nonetheless, the utility of rice sequences for other cereals has been demonstrated by using rice sequences as substrates for investigating gene expression and genetic variation in other species. Goff et al. (2002) reported that 90% of the rice sequences arrayed as oligos on a chip can be used to detect expression in maize, barley, and wheat.

C. COMPARISON BETWEEN INDICA

AND

JAPONICA RICE

Because genetic variation within a species represents the raw material for adaptive changes, special attention is given to examining the occurrence of single nucleotide polymorphisms (SNP), deletions and insertions between indica and japonica rice on a genome-wide scale. At the whole-chromosome level, Sasaki et al. (2002) used sequences from ﬁnished chromosome 1 of Nipponbare to query the assembled contigs of the BGI draft indica genome data. They found about 78% of the japonica sequence in the indica database but failed to identify matches to the remaining 22% of the sequences in chromosome 1. How much of this discrepancy is due to genetic differences between the indica and japonica sequences or to artifacts in the assembly of the indica sequence remains to be determined. At a ﬁner scale, Feng et al. (2002) compared a 2.3-Mb region on chromosome 4 of japonica (Nippponbare) and indica (Guanglui 4) rice. While the overall colinearity is maintained, there were many insertions and deletions. Insertions occurred not only in the intergenic region but also in the coding regions. Interestingly, there were more transposable elements in the japonica chromosome, contrary to the common perception that the total genome size of indica rice is larger than that of japonica rice. Overall, there was one SNP per 268 bp (0.37%) between the two sequenced varieties in this region of the chromosome.

RICE FUNCTIONAL GENOMICS

69

Using genome sequence of indica variety 93-11 as a reference, Yu et al. (2002) examined the polymorphisms with three varieties — Nipponbare, PA64s, and Guanglui 4 — and found frequencies of SNP in unique sequences (relative to 93-11) at 0.50, 0.35, and 0.50%, respectively. Nasu et al. (2002) examined 417 chromosomal regions in three japonica, two indica varieties and a wild relative O. ruﬁpogon and found one SNP in 232 bp between two randomly chosen lines. In comparison, about 1 SNP in 1000 bp was observed between Arabidopsis ecotypes Columbia and Landsberg (Cho et al., 1999). More recently, Katagiri (2003) presented preliminary data on the sequence variation between chromosome 1 of indica variety Kasalath and Nipponbare. Of 189,912 bp BAC-end sequences compared, one SNP/122 bp was observed. About 0.15% of the SNP was found in expressed genes. Monna (2003) analyzed six japnoica varieties (Nipponbare, Koshihikari, Kitaake, Akihikari, Sensho, and Itadaki) and two indica varieties (Kasalath and Guang-lu-ai 4) and found more SNP between the two indica varieties than between japonica and indica varieties. It is too early to conclude that these statistics are indicative of the degree of variation between the subspecies as these differences may simply reﬂect differences between varieties (rather than between two subspecies). A generalization of level of SNP can only be made after more sequence data become available for multiple varieties within japonica and indica rice. Such a determination of the within- and between-subspecies variation will be important for applying nucleotide-level variation in genetic association tests.

IV. KEY INGREDIENTS FOR GENE DISCOVERY IN RICE The genetic blueprint of rice can be viewed as a dictionary of characters, but without explanation of their function. To ﬁnd the gene function encoded by each gene or set of genes requires special kinds of genetic resources and methodologies. Figure 3 depicts such a platform of resources and technologies for large-scale gene discovery. The platform has three basic components: genetic resources, phenotypic/biological evaluation, and genomic tools. These three components are integrated by bioinformatics, which provides integration of data produced from different activities. Furthermore, bioinformatics helps develop hypotheses for further experimentation. This reiteration of hypothesis testing and data production eventually leads to the discovery and validation of gene functions. The framework brings multidisciplinary teams and different expertise together. It also highlights the importance of delivering the products of gene discovery to end-users through collaboration with breeding programs that can incorporate genes into appropriate genetic backgrounds.

70

H. LEUNG AND G. AN

Figure 3 Public research platform for rice functional genomics: genetic resources, phenotyping, high-throughput genomic tools, and bioinformatics. Linkage to breeding networks in rice growing countries represents the practical dimension of rice gene discovery.

A. GENETIC RESOURCES Genetic stocks harboring the traits of interest and associated variation are the essential materials for revealing biological function of a sequence. A range of genetic stocks can be used to identify or validate gene function. These include mutants, segregating mapping populations, backcross lines, well-characterized breeding lines and traditional germplasm. Among these genetic stocks, mutants are the “work horse” of gene discovery because they reveal phenotypic changes that correspond to a precise change in the genome. Furthermore, mutations often lead to over- or under-production of biochemical intermediates that can lead to identiﬁcation of pathways. Induced mutations alone, however, are not sufﬁcient to understand phenotypic variation caused by allelic diversity. Therefore, it is important to couple functional analysis with the diverse germplasm developed for genetic mapping and breeding. Allelic variation present in conserved or selected germplasm is particularly valuable because they represent products of natural selection under various evolutionary forces.

B. HIGH-THROUGHPUT TECHNOLOGIES A distinct feature of functional genomics is the application of high-throughput technologies that can reveal the behavior of thousands of genes at one time (Hieter and Boguski, 1999). Large-scale and genome-wide methodologies

RICE FUNCTIONAL GENOMICS

71

become an essential part of experimental design, hypothesis testing, and data analysis. We consider only brieﬂy the current technologies that are sufﬁciently robust for broad applications. With available genome sequence information from the large-scale sequencing of many organisms, it is possible to produce arrays carrying thousands of genes, providing a high-throughput method for gene expression analysis (essentially reverse northern analysis) and for the detection of genetic variation (reverse Southern analysis) (DeRisi et al., 1997; Schena and Davis, 1999). Originally, there were two main technologies for producing gene arrays: (1) spotting DNA [cDNA clones, PCR products, oligonucleotides (oligos)] and (2) in situ synthesis of oligos on slides by photolithography (Fodor et al., 1991). New and “hybrid” methodologies have been developed: ink-jet deposition (Hughes et al., 2001) and maskless array synthesis (Nuwaysir et al., 2002), which improve the ﬂexibility and utility of the arrays. Furthermore, some arrays (e.g., Diversity Array) assay the genetic diversity of any organism without depending on available sequence information (Jaccoud et al., 2001). While spotted cDNA arrays are currently the most commonly used, increasing attention is given to oligo arrays (50 or 60 mers) because of their increased speciﬁcity, quality control, and portability (Kane et al., 2000). Depending on the design of the oligos (choice of gene sequence, oligo size and number), oligo arrays can be used effectively to detect nucleotide polymorphism or deletions in the target genomes. Currently, rice gene arrayscDNA or oligos- of varying sizes and coverage are produced by individual groups or national programs. Progress is being made to develop large-scale oligo arrays for rice that are publicly accessible (S. Kikuchi, personal communication; Beijing Genomics Institute, personal communication). Although gene arrays provide a picture of the transcription proﬁle, they do not reveal post-transcriptional regulation of the genes. As in transcript analysis, proteomics provides a high-throughput platform to survey many proteins simultaneously under different experimental conditions (Zivy and de Vienne, 2000). Proteins can be separated in two-dimensional polyacrylamine gel electrophoresis (2D-PAGE), ﬁrst by isoelectric focusing in the ﬁrst dimension followed by molecular weight in the second dimension. Protein spots can then be extracted for microsequencing. In general, 2000 –3000 proteins can be resolved and analyzed using the 2D gel technique coupled with mass spectrometry. New technologies for proteomic analyses are being developed (Koller et al., 2002). Readers are referred to an overview of proteomic technology and its applications for gene discovery by Salekdeh et al. (2002a).

C. BIOLOGICAL EVALUATION High-quality phenotypic data are central to a gene discovery program yet infrastructure and expertise for phenotyping are often taken for granted. The lack

72

H. LEUNG AND G. AN

of adequate phenotyping capacity can be a bottleneck for gene discovery. High-quality phenotyping means a precise, reproducible biological assay of the trait in question. Phenotyping may entail different levels of sophistication from simple visual evaluation to the measure of a detailed metabolic response under highly controlled environmental conditions. Overall, in a chemical- or irradiation-induced mutant population, 6 – 10% of the mutant lines has visible phenotypes in either the vegetative or reproductive stage. While the obvious phenotype (e.g., plant size and color) can be discerned by inexperienced observers, subtle morphological changes often require trained eyes. Plant breeders, for instance, are particularly keen to observe overall plant architecture, panicle shape, and grain ﬁlling as a means to improving plant type and yield. To maximize the use of any large collection of mutants, it would be beneﬁcial to have regular “grow-out” ﬁeld days for expert observers to select phenotypes of interest. For a lot of agronomic traits, the target phenotypes must be observed through conditional screens in which mutants, isogenic lines and breeding lines are subject to speciﬁc conditions to express the phenotypes. For examples, in screening for drought-tolerance, the developmental stages (vegetative or reproductive) at which the plants are subject to water stress may determine the kinds of genes that can be identiﬁed. The nature and parameters of conditional screens are therefore vital for successful detection of the desired traits. Mutants with a uniform genetic background have the advantage that any deviation in phenotype can be identiﬁed readily. As we search for speciﬁc variants, innovative screening techniques using reporter gene constructs will prove useful (Page and Grossniklaus, 2002). Furthermore, screening for secondary mutations will be increasingly important for deﬁning the effect of the ﬁrst mutations and revealing interaction between genes and pathways.

D. BIOINFORMATICS Bioinformatics brings together biological and computational tools that can integrate data from structural and functional analyses and present new models and hypotheses for further testing or validation. Bioinformatics helps us integrate information to formulate new questions to be examined, hence bringing new insights into the process of gene discovery. Because of the increasingly large amount of information produced, it is no longer adequate to examine data in an isolated manner. A key function of bioinformatics is to provide a system by which information can be retrieved, analyzed, and distributed to a broad community of researchers and users. In summary, the four components — genetic resources, biological evaluation, genomics tools, and bioinformatics — require diverse skills in biology and modern technologies. They require infrastructure in modern science and

RICE FUNCTIONAL GENOMICS

73

engineering (nanotechnology, computer science, etc.) and conventional biological studies (phenotyping and innovative screening procedures). While individual laboratories or institutions may be well suited to pursue some of these activities, a single laboratory will unlikely have all the resources and expertise to examine all questions. Thus, functional genomics is necessarily a collaborative process and a public resource platform provides a framework for the division of labor and formation of partnerships.

V. FORWARD AND REVERSE GENETICS Forward genetics refers to ﬁnding gene function by starting with a phenotype and then identifying the genes responsible through inheritance studies, followed by genetic mapping to locate the target gene on a chromosomal region. By “walking” down the chromosome using overlapping markers, the DNA sequence responsible for the trait can eventually be identiﬁed. The ﬁnal step of gene isolation can be hastened considerably with the whole-genome sequence available. Reverse genetics, in contrast, begins with a known DNA sequence and works backward to identify the biological function ascribed to that sequence. Typically, a sequence with a predictive function is used to search for knockout lines in a pool of DNA or electronically from databases of insertion mutants (see below). Once a mutant is found, it can be assayed for phenotypes based on the functional prediction. For organisms with whole-genome sequences available, this provides a new paradigm for gene discovery. With forward and reverse genetics, it is within our reach to systematically assign a function to every gene of interest in rice. For this reason, there have been many initiatives around the world to produce rice mutants. Below, we will use several representative mutant stocks to demonstrate the various strategies.

A. INSERTIONAL MUTANTS Insertional mutagenesis has been the favored approach for tagging genes of interest. The most advanced systems have been developed in maize and Arabidopsis using both transposons and integrative plasmids as the mutagenic agents. Although the initial investment in producing insertion mutants is high, the pay off is great because the mutated genes are potentially tagged. Using a variety of vectors originally designed and tested in Arabidopsis, rapid progress has been made in the production and characterization of rice insertion mutants over the past several years.

74

H. LEUNG AND G. AN

1.

T-DNA Tagging

The Arabidopsis knockout facility at the University of Wisconsin recently established a population of 60,480 T-DNA-tagged lines (Krysan et al., 1999). This is a signiﬁcant step toward the production of genome-wide mutations. T-DNA insertion in Arabidopsis is probably a random event, with the inserted sequences being stable through multiple generations (Azpiroz-Leehan and Feldmann, 1997). Recently, this T-DNA tagging strategy was employed in rice (Jeon et al., 2000). An’s laboratory at Pohang University of Science and Technology in Korea has produced approximately 100,000 fertile rice lines that have been tagged by T-DNA using binary vectors of Agrobacterium tumefaciens (Jeon et al., 2000). Figure 4 illustrates the basic principle of insertional mutagenesis using known DNA elements. Analysis of randomly selected transgenic plants has indicated an average of 1.4 loci of T-DNA inserts per line. Therefore, the T-DNA knockout stocks would provide a 63% probability of ﬁnding an insertion within a given gene if T-DNA inserts randomly. It appears that T-DNA prefers transcriptionally active regions in rice (Barakat et al., 2000). Therefore, the chance of ﬁnding a knockout gene in the rice T-DNA insertional stocks would be higher than the estimated value. One difﬁculty in establishing a population of T-DNA-tagged lines is obtaining enough seeds from the ﬁrst generation. In experiments with 10,743 primary transgenic plants, 39% bore fewer than 50 seeds and only 50% yielded more than 100 (Jeon et al., 2000). Yin and Wang (2000) observed similar results from their analysis of 2633 transgenic rice plants. Although the majority of the transgenic plants became fertile in the T2 generation, 22% of the lines showed

Figure 4 Production of insertion lines using introduced T-DNA and endogenous retrotransposon Tos17. T-DNA can be inserted into either a gene (line 1) or an intergenic region (line 2). During the transformation procedure, the endogenous retrotransposon Tos17 can be activated and inserted into a gene (line 3).

RICE FUNCTIONAL GENOMICS

75

less than 50% fertility. Reduced fertility is probably due, in part, to the lethal effects of T-DNA insertion. However, a considerable number of lethal mutations are likely induced during tissue culture. Those lines that show low fertility must be individually ampliﬁed before being used for further analysis.

2.

Transposon Tagging

Transposon tagging has been a powerful tool for isolating new genes since the controlling element was ﬁrst recognized by McClintock (Fedoroff et al., 1983). The ﬁrst successful cloning of a plant gene was achieved via the Ac/Ds (Activator – Dissociation) transposon system (Fedoroff et al., 1984). Other transposon systems, such as En/Spm (Enhancer/Suppressor – mutator) and Mu (Mutator), have been used for cloning several genes of maize (Walbot, 2000). These elements are active not only in their natural hosts but also in other plants, including Arabidopsis. The maize Ac/Ds system has been tested for gene tagging in rice. First, the autonomous Ac element was cloned between a promoter and the hygromycin phosphotransferase coding region. The construct was then introduced into rice chromosomes by direct transformation. Transposition of the Ac element was proven by recovering hygromycin-resistant plants (Izawa et al., 1991; Murai et al., 1991). Enoki et al. (1999) analyzed the behavior of 559 plants from four transgenic rice families through three successive generations; 18.9% of the plants contained newly transposed Ac insertions. DNA pools from 6000 Ac-inserted plants were organized in a 3D matrix (row — column –plate pools) and subjected to PCR screening. Of the 14 randomly selected genes, two knockouts were identiﬁed, one of which encodes the rice cytochrome P450 (CYP86) gene. Molecular analysis of Ac transmission to the progeny revealed that its germinal transmission was detected in 9 of the 12 insertions, indicating that some Ac transmissions occurred in somatic tissues that would not transmit to the progeny. Sequence analysis of 99 ﬂanking regions over the 50 region of Ac indicated that it preferentially transposed into protein-coding regions. The non-autonomous Ds element has been transposed in the presence of Ac transposase via the direct gene transfer method (Shimamoto et al., 1993; Sugimoto et al., 1994). Germinal transposition of Ds was observed at a high frequency in the R2 progeny when a transgenic plant containing the Ds element was crossed with a transgenic plant carrying Ac transposase under the control of the cauliﬂower mosaic virus (CaMV) 35S promoter. A wide spectrum of mutations, affecting growth, morphogenesis, ﬂowering time, and disease resistance, was observed in the Ds population (Izawa et al., 1997). Whether these mutations are due to Ds must still be determined. The frequency of Ds transposition declined signiﬁcantly in subsequent generations, even in

76

H. LEUNG AND G. AN

the presence of Ac transposase. However, transposition could be recovered in certain lines through protoplast regeneration. Similar Ac/Ds gene tagging systems have been developed for rice by employing the Agrobacterium-mediated gene delivery method (Chin et al., 1999; Nakagawa et al., 2000). This method mediates transfer of one to a few copies of T-DNA into plant cell chromosomes, thereby making the transgene more stable compared with the direct gene transfer method (Dai et al., 2001). As many as 80% of the Ds elements have been excised from the original integration sites in the presence of Ac transposase activity provided in trans. The germinal excision frequency was no more than 40% (Nakagawa et al., 2000). Repetitive ratoon culturing caused new transposition, at about 30% frequency, thereby demonstrating that the procedure is capable of producing a large population of mutants (Chin et al., 1999). Linkage analysis of the empty donor site and the transposed Ds insertion site revealed that four of ﬁve insertion sites were linked to the donor site (Nakagawa et al., 2000). This indicates that most of the transposition events occurred at a tightly linked site, as had been reported for Arabidopsis (Machida et al., 1997; Sundaresan, 1996). In that species, the physical map positions of 356 Ds insertions showed signiﬁcant preference for transposition to the regions adjacent to the nucleolus organizer regions (Parinov et al., 1999). For any given gene, insertions preferentially occurred at the 50 end of the gene. Therefore, the shortrange and highly preferential transposition system can be effectively used for the targeted mutagenesis of closely linked genes. In contrast, transpositions of another mobile element, En/spm, were well distributed over different chromosomes and were not clustered at a few genomic locations in Arabidopsis (Wisman et al., 1998). However, the tendency of the En/Spm element to amplify can potentially complicate the interpretation of phenotypes and molecular analyses. Ac/Ds in conjunction with Spm/dSpm has been proposed for use in developing a four-element system for directed tagging of crop-speciﬁc alleles (Phogat et al., 2000). In the ﬁrst step, stocks carrying Ds within dSpm would be crossed to a line with Spm to facilitate random transposition of the transposon. In the second step, individual lines would be crossed to a line with Ac to induce local transposition of the Ds element. This four-element system would therefore exploit the natural tendency of unlinked jumps of dSpm and linked jumps of Ds. Unfortunately, the Ac element in this four-element system would undergo frequent excision from the target gene, causing variegation. To overcome this, a self-stabilizing Ac derivative has been developed that undergoes autonomous transposition but is stable after integration (Schmitz and Theres, 1994). Charng et al. (2000) have designed an inducible transposon system using an element that contains a PR-1a promoter::transposase fusion (PR-1a::Tpase). Treatment with salicylic acid (SA) induces the expression of transposase, resulting in the transposition of Ds in somatic tissues. Germinal transposition events also have

RICE FUNCTIONAL GENOMICS

77

been observed after SA treatment of ﬂowers. The development of inducible transposable element systems will also be valuable for rice. Recently, Qu et al. (2003) explored the use of a GVG chemically inducible system to control transposition of Ds. Preliminary data suggested that transposition of Ds could be induced by dexamethasone (DEX).

3. Retrotransposon Tagging Another effort toward gene tagging in rice involves introducing the tobacco retrotransposon Tto1 and demonstrating its autonomous transposition through reverse transcription (Hirochika et al., 1996). A transposable element, Tag1 from Arabidopsis, has also been analyzed in rice (Liu et al., 1999b). The transcription and excision behaviors of Tag1 were similar in both rice and Arabidopsis. However, the excision was tissue-speciﬁc only in rice. A survey of 73,000 sequence-tagged connectors (STC), corresponding to nearly 50 Mb of the rice genome, showed that retroelements are randomly distributed with respect to potential genes (Mao et al., 2002). Retrotransposons are both functionally and structurally different from well-characterized transposable elements such as Ac and Spm. These retrotransposons are believed to be involved in gene duplication as well as in the regulation of gene expression. Transposition of some retroelements can be induced by stresses caused by pathogen infection, cell culture, and wounding (Takeda et al., 1998; Parinov et al., 1999). Such endogenous elements in the rice genome have provided excellent tools for gene tagging. One such element is Tos17, which is activated during tissue culture and ampliﬁed up to 30 copies (Parinov et al., 1999). A gene knockout system using Tos17 has been developed for identifying insertional mutations in several genes. After screening 550 plants that were mutagenized by Tos17, Sato et al, (1999) identiﬁed a mutation in the homeobox gene OSH15. Rice phytochrome A ( phyA) mutant lines have also been isolated from a Tos17induced mutant population using a 3D DNA-pooling system (Takano et al., 2001). By screening for viviparous mutants, Agrawal et al. (2001) identiﬁed Tos17 insertions in the rice zeaxanthin epoxidase gene (OsABA1) and in a novel OsTATC gene, which shows a weak homology with bacterial Sec-independent translocase TATC. More than 32,000 Tos17 insertion lines have been produced and more than 8600 independent sequence-ﬂanking insertion sites have been determined (Miyao et al., 2001). The mutant stocks are available upon request (see http://tos.nias.affrc.go.jp/~miyao/pub/tos17). Analysis of insertion points in the rice genome sequence has revealed that insertions in the genic regions (exon and intron) were 2-fold higher than in the intergenic region, suggesting that Tos17 prefers the genic regions for target sites. In addition, hot spots of Tos17 insertions

78

H. LEUNG AND G. AN

were observed, with some of them being clustered. This may limit the efﬁciency of genome-wide gene tagging using Tos17 alone.

4.

Activation Tagging

Conventional strategies, such as T-DNA mutagenesis or transposon tagging, are not efﬁcient for analyzing the function of redundant genes. Likewise, neither T-DNA tagging nor transposon tagging will identify the genes that are required during multiple stages of a life cycle and whose loss-of-function results in early embryonic or gametophytic mortality (Weigel et al., 2000). Fewer than 10% of the genes tagged in the Arabidopsis genome are likely to produce a visible phenotypic change. Therefore, complementing technologies are needed to assess the function of the remaining genes. One such technique is the activation-tagging system developed in Arabidopsis and successfully used for cloning several genes (Kardailsky et al., 1999; Ito and Meyerowitz, 2000; Lee et al., 2000; Weigel et al., 2000). This system uses T-DNA vectors that contain multimerized CaMV 35S transcriptional enhancers positioned near the right T-DNA border (Hayashi et al., 1992). Gain-of-function mutagenesis in rice by activation tagging requires a strong enhancer element. Since the CaMV 35S enhancer element has enhancer activity in rice cells, they can be used to produce activation tagging lines in rice (Fig. 5). A transposon-mediated activation-tagging system has also been developed using the Ds element that carries the tetramerized CaMV 35S enhancer (Mori et al., 2000). From the T-DNA activation-tagging pools of Arabidopsis, Weigel et al. (2000) characterized over 30 dominant mutants with various phenotypes.

Figure 5 Activation tagging using an insertion construct carrying a 4X 35S promoter to drive the expression of endogenous genes. The multimerized 35S enhancer located at one end of the T-DNA can enhance the expression of a nearby gene from upstream (top) or downstream (bottom) of the gene.

RICE FUNCTIONAL GENOMICS

79

Analysis of a subset of the mutants has shown that the tagging vector causes overexpression of the gene immediately adjacent to the inserted enhancer. In rice, T-DNA activation-tagging vector pGA2715 was developed for both trapping and activation tagging of rice genes (Jeong et al., 2002). The binary vector pGA2715 contains the promoterless glucuronidase (GUS) reporter gene next to the right border and the multimerized transcriptional enhancers from the CaMV 35S promoter next to the left border. A total of 13,450 T-DNA insertional lines were produced using the binary vector pGA2715. Histochemical GUS assay revealed that the GUS-staining frequency from the pGA2715 lines was about two times higher than that from the lines transformed with binary vector pGA2707, which lacks the enhancer element. This result suggests that the enhancer sequence present in the T-DNA enhanced GUS tagging efﬁciency. RT-PCR analysis of a subset of randomly selected pGA2715 lines has shown that expression of the genes immediately adjacent to the inserted enhancer was increased signiﬁcantly. These results indicated that the large population of T-DNA-tagged lines transformed with pGA2715 could be used for trapping a gene using the GUS reporter as well as for isolation of gain-offunction mutants in rice.

5.

Entrapment tagging

An entrapment-tagging system allows for monitoring gene activity by creating fusions between tagged genes and a reporter gene, such as GUS and green ﬂuorescent protein (GFP). Insertion of the promoterless reporter will not only destroy normal gene function but also activate expression of the reporter gene. Three entrapment systems are available: enhancer trap, promoter trap, and gene trap (Springer, 2000). The devices for entrapment can be transferred into plant cells as a part of T-DNA or transposons (Sundaresan, 1996; Martienssen, 1998; Springer, 2000). This approach has been successfully applied to genes that are difﬁcult to identify by traditional methods. Some examples are regulatory sequences that drive reporter gene expression in nematode-feeding structures (Barthels et al., 1997), molecular markers for embryogenesis (Topping and Lindsey, 1997) or regulatory sequences that mediate guard-cell-speciﬁc expression (Plesch et al., 2000). Activation of a reporter gene in a promoter trap vector can be as high as 30% (Sundaresan, 1996). In rice, at least 5% of the T-DNAs and 10% of the transposed Ds elements become activated in various tissues, for examples, roots, leaves, ﬂowers, and seeds (Chin et al., 1999; Jeon et al., 2000). If reporter gene activation by certain environmental conditions or by chemicals such as growth substances is included, total tagging efﬁciency can be higher. Some of the tags have displayed tissue- or organ-speciﬁc reporter expression, while others have exhibited

80

H. LEUNG AND G. AN

ubiquitous expression patterns. One must study whether these lines will manifest any mutant phenotypes in organs where the reporter has been activated, and whether the phenotypes co-segregate with the inserted reporter gene. The GUS gene has frequently been used for gene trapping in plants because of the accurate detection of its gene product and tolerance for the N-terminal translational fusions in its enzyme activity. However, one problem with the GUS assay is the destructive nature of its staining and destaining procedures (Jefferson et al., 1987). Non-invasive and non-destructive reporter genes, such as GFP or luciferase, have not yet been widely used for gene trapping in plants. Recently, a visualization system with a charge-coupled device camera, band-pass ﬁlters, and a light source was used to demonstrate that green ﬂuorescence emitted from GFP could be visualized in the calli, dry seeds, roots, and seedlings of transgenic rice plants (Chung et al., 2000). Such an efﬁcient visualization system should facilitate the use of a non-destructive visual selection marker for entrapment tagging in rice. We are currently investigating the efﬁciency of GFP expression in rice plants transformed with the vector pGA2717 that carries the promoterless GFP next to the left T-DNA border (Gyn An, unpublished data). A new genomic DNA-based signal sequence trap method, the signal-exon trap (SET), may help identify genes that encode secreted and membrane-bound proteins (Peterfy et al., 2000). SET is based on the coupling of an exon trap to the translation of captured exons, thus allowing the screening of exon-encoded polypeptides for signal peptide function. This system may be helpful in the discovery of novel members of known secretory gene clusters as well as for other positional cloning approaches.

6.

Forward Screening

Insertional mutant pools are useful resources for studying gene function. First, they can be used to identify mutants whose function is altered by the insertion. Because most mutations are recessive, their phenotypes are easily detected in a segregating population. However, because rice plants are much larger than Arabidopsis and need to be grown in speciﬁc conditions, large-scale phenotypic screening is often limited by space and time. Systematic efforts by collaboration with several research groups are therefore needed to screen a large number of mutants. Because mutations could be induced by unknown endogenous transposons during tissue culture, one must determine whether the phenotypes induced by the rice transformation are due to the insertion. Once the mutant phenotype has been shown to co-segregate with the insertional element, the sequence ﬂanking the insert can then be isolated by PCR-based methods, such as thermal asymmetric interlaced (TAIL) PCR (Liu and Whittier, 1995),

RICE FUNCTIONAL GENOMICS

81

adapter-ligated PCR (Balzergue et al., 2001), inverse PCR (Triglia et al., 1988), a universal biotinylated adapter ampliﬁcation procedure (Hanley et al., 2000), a panhandle PCR (Walbot, 2000), or a plasmid rescue system (Weigel et al., 2000).

7.

DNA Pool Screening

Insertional mutants can be identiﬁed in a given gene via PCR-based screening (Krysan et al., 1999; Sato et al., 1999). Using a gene-speciﬁc primer and a primer located near the end of the insert, the fragment ﬂanking the insert can be ampliﬁed and its sequence then determined. Because screening of individual lines requires great effort, DNA pools of a large number of lines are commonly used for identifying a knockout. In Arabidopsis, it was reported that the maximum useful pool size is about 2350 lines per pool (Krysan et al., 1999). Because the rice genome is about four times larger than the Arabidopsis genome, a pool of 500 lines (less complexity) would be needed for PCR-based screening of the knockout mutations in rice. DNA pool screening has been performed from the Tos17 insertional pools. An’s laboratory has prepared DNA pools from 40,000 T-DNA lines and provided the screening service for the isolation of mutants from the T-DNA insertional lines. Figure 6 shows the overall scheme of detecting insertional rice mutants from the screening of DNA pool from T-DNA tagged lines. 8.

Insertion Sequence Database

Random sequence analysis of the fragments ﬂanking the insertional element can help identify gene function. Flanking sequences from Ds insertions of 931 independent transgenic lines in Arabidopsis have been characterized (Parinov et al., 1999). In a similar study, Ito et al. (2002) determined the genomic sequences ﬂanking the Ds elements and mapped the elements’ insertion sites in 1173 transposed lines by comparison with the published genomic sequences. One-half of the lines contained Ds on the same chromosome, in particular, the hot spots near the three starting loci. In the other lines, the Ds elements were transposed across chromosomes. The Ds elements tended to transpose near the chromosome ends and rarely transposed near the centromeres. Likewise, directsequence characterization of 761 Mutator-tagged fragments in maize has shown that random sequencing of transposon-tagged fragments can produce signiﬁcant numbers of interesting transposon-tagged genes and mutant lines (Hanley et al., 2000). In rice, the ﬂanking sequences of Tos17 were determined by TAIL- and suppression-PCR (Hirochika, 2001). As of October 2002, 23,709 ﬂanking sequences from 2424 insertion lines had been determined (A Miyao et al.,

82

H. LEUNG AND G. AN

Figure 6 PCR screening of DNA pool obtained from T-DNA insertion lines. a) DNA from individual plants pooled in different levels of complexity. Each pool contains DNA from 50 to 100 lines. A super pool consists of 500 lines. b) A rice gene OsMADS6 is used to illustrate PCR screening to detect the tagged mutant. Four screens were conducted using two primers from the gene and two

RICE FUNCTIONAL GENOMICS

83

unpublished data). These are classiﬁed into 8653 independent ﬂanking sequences, about 39% of which showed homology to known genes. The ﬂanking sequences of T-DNA insertional lines and Ds-tagged lines of rice are being determined. This strategy should be useful for both the identiﬁcation of mutants and the discovery of new genes. Since the rice genome sequence is almost completed, disrupted genes can be readily identiﬁed and it will be possible to map the insertion sites on the chromosomes by comparing the ﬂanking sequences with the genomic sequence. An insertion mutant line of any gene can be identiﬁed in silico and, thus, ﬂanking-sequence databases will become the basis for reverse genetics studies. Construction of such databases through collaboration with laboratories producing rice insertional mutant lines will be valuable for determining gene function on a large scale.

B. CHEMICAL-

AND IRRADIATION-INDUCED

MUTANTS

Although insertional mutagenesis represents the most direct approach for tagging a particular gene, there are technical and logistical problems in achieving complete genome coverage based on insertional mutagenesis alone. First, the investment of producing a large collection of insertion mutants is high and not necessarily feasible for many laboratories. Second, insertional mutagenesis based on the introduction of exogenous transponsons and T-DNA is often genotypedependent. Currently, only certain japonica rice genotypes are amenable to efﬁcient large-scale transformation. Activation of retrotransposons such as Tos17 depends on the ability to perform tissue culture and the level of transposition also varies among genotypes. Third, the transgenic nature of the mutants often prevents large-scale ﬁeld cultivation in certain parts of the world. Finally, parts of the genome may not be accessible to insertional inactivation, thus preventing complete genome coverage. In contrast, chemical and irradiation mutagenesis offer several advantages. Production of the mutants is relatively inexpensive. Any genotype can be mutagenized and the distribution of mutations is probably random in the genome. The main shortcoming of the chemical- and irradiationinduced mutants is that they are not tagged, making subsequent gene isolation difﬁcult. However, with the new high-throughput genotyping techniques, it is now possible to detect mutations rapidly, hence increasing the utility of these mutants. primers located near the right or left border of the T-DNA. In the ﬁrst step, super pool 7 was detected to carry a T-DNA insertion in the OsMADS6 gene. To increase sensitivity and ﬁdelity, a DNA-gel blot analysis was performed with the radioactively labeled OsMADS6 probe. In step 2, pool 65 was identiﬁed to carry the insertion. An individual line that carried the insertion was identiﬁed in step 3. The exact insertion position can be determined by sequencing the PCR product from step 3.

84

H. LEUNG AND G. AN

1.

Choice of Mutagens

To produce a comprehensive mutant library useful for forward and reverse genetics, we need to consider the efﬁciency and speciﬁcity of the mutagens. Efﬁciency means how much potency the mutagen has for producing mutations. It also means that the mutagenesis procedure should be straightforward such that a sufﬁcient number of mutations in the genome can be produced in a few largescale experiments. In this respect, chemical mutagenesis is more convenient than irradiation experiments as special radiation facilities are not needed. Speciﬁcity refers to the predominant kinds of mutations (small or large deletions, point mutations) produced by a particular mutagen (Koornneef et al., 1982). Depending on the approaches to be used to identify mutations, speciﬁc mutagens can be selected to produce the desired type and size of genetic lesions. Ethyl methanesulfonate (EMS) is perhaps the most commonly used mutagen in plants because of its potency and ease of use. The mutagen causes mostly transition point mutations. In rice, EMS mutagenesis involves soaking the seeds in an aqueous solution at a chosen concentration (from 0.2 – 2.0%) for a given time. It should be emphasized that rice genotypes have a different sensitivity to mutagens. It is therefore important to produce a kill curve to determine the desirable level of survival. An advantage of EMS mutagenesis is that a relatively small mutant population (, 10,000) is needed to provide near-genome coverage because the genome can tolerate a large number of point mutations without showing deleterious effects. In Arabidopsis, point mutation density can be as high as four mutations per Mb (L. Comai, personal communication). Deletion can be induced by a variety of chemical and irradiation treatments. Ionizing radiation has been widely used to induce mutations for plant breeding and classical genetic analysis, but the consequences of ionizing radiation have been examined at the molecular level only in a few organisms. Shirley et al. (1992) examined the effects of fast neutron (FN) and X-ray irradiation at two genetic loci encoding chalcone ﬂavonone isomerase and dihydroﬂavonol 4-reductase in Arabidopsis. They found not only deletions at the loci but also distant chromosomal rearrangements. From a population of FN-induced M2 seedlings, Bruggemann et al. (1996) isolated 20 mutations at the HY4 locus in Arabidopsis; 15 of these were deletions detectable by RFLP analysis. Okubara et al. (1994) used FN to produce lettuce mutants with increased susceptibility to downy mildew. Of 14 mutations at four Dm genes isolated from 2211 FN-treated M2 lines, 10 were shown to be deletions by analysis of RAPD markers linked to the Dm genes. Besides ionizing irradiation, several chemicals have been shown to induce predominantly small to kilobase-size deletions. Using diepoxybutane (DEB) to induce mutations at the rosy locus in Drosophila, Reardon et al. (1987) found that 43% of the mutations were deletions ranging from 50 to 8 kb. Yandell et al. (1994) found that trimethylpsoralen was effective in inducing deletions, ranging

RICE FUNCTIONAL GENOMICS

85

from 0.1 to 15 kb in Caenorhabditis elegans. Aguirrezabalaga et al. (1995) showed that hexamethylphosphoramide (HMPA), a cross-linking agent, caused deletions in 96% of all mutants recovered in Drosophila. How many mutants do we need to have sufﬁcient coverage of the genome? In Arabidopsis, Krysan et al. (1996) estimated that the probability of any gene being tagged by T-DNA insertion depends on the size of the gene and the total genome size. Dividing the total genome by an average gene size will give the number of targets in the genome to be tagged or mutated. In Arabidopsis, to have a 95% chance of ﬁnding an insertion in a 3.2-kb gene, one would need 95,000 T-DNA lines (assuming 1.5 insertions per line). Extending the same reasoning to rice (with approximately 4 £ the genome size of Arabidopsis), the population size needed for genome coverage would be . 400,000 lines, assuming one deletion per line. Since multiple mutations are induced by chemical or irradiation mutagenesis, fewer independent lines are needed. For example, assuming an average of 10 deletions per line, one would need approximately 40,000 lines to have a 95% chance of mutating any gene in the genome. These are obviously rough estimates as we do not have an accurate estimate of the number of induced mutations per genome. Nonetheless, it appears that a population of 50,000 mutant lines should provide reasonable genome coverage.

2. Deletion and Point Mutation Stocks To our knowledge, there are currently two large collections of chemical- and irradiation-induced rice mutants. The ﬁrst one is an FN-induced population originally developed by Pamela Ronald at the University of California-Davis using japonica variety M202 and subsequently acquired by a private company. The collection consists of 24,660 M2 lines that have been used for highthroughput PCR screening (Li et al., 2001). The second collection is produced and maintained at IRRI. Four mutagenic agents — fast neutron, gamma ray, DEB, and EMS — were used to produce mutants with different sizes of genetic lesions. Variety IR64 was chosen as the standard genotype for producing a comprehensive mutant population in indica rice. IR64 is the most widely grown rice variety in the tropics and it carries many valuable agronomic traits related to yield, plant architecture, grain quality, and tolerance for biotic and abiotic stresses. Creating mutations in such an elite genetic background can facilitate the detection of phenotypic changes in important agronomic traits (Leung et al., 2001). Currently, the population consists of approximately 30,000 lines at the M4 stage. The goal is to produce . 40,000 independent mutant lines toward the end of 2003 to give a high probability of detecting a mutation in most genes. By producing a population of mutants with an identical genotype using different mutagens, we aim to produce

86

H. LEUNG AND G. AN

a large allelic series for any locus as a tool for gene validation. Furthermore, the population is advanced to individual M4 lines such that sufﬁcient seeds are available for distribution for the screening and evaluation of quantitative traits. Deletions. The concept of using deletion mutations for forward and reverse genetics is straightforward though it could be technically challenging. If sizable DNA regions (e.g., 100 –1000 base pairs) are deleted in a genome, it should be possible to detect such lesions by various genotyping approaches. As a proof of concept, Chang et al. (2003) evaluated the utility of a genome wide chip to detect deletions in known mutations. DNA from a gamma ray induced d1 mutation with a known deletion in the locus encoding a G-protein (Fujisawa et al., 1999) was used to hybridize to 21,000 predicted genes (each represented by a 16 25-mer probe set) on a Rice GeneChip. A hybridization signal ﬁve times below the wildtype was used as a criterion to indicate deletion. A stretch of deletions encompassing four genes was detected on chromosome 5 precisely at the map location containing the d1 gene. Results from experiments with a deletion at the Xa21 locus, however, were less deﬁnitive due to cross hybridization between similar sequences of the disease resistance gene family. Nonetheless, these experiments open up a powerful approach to apply genome-wide oligo chip to identify deletions in mutants that have been phenotypically well characterized. A successful reverse genetic screen using deletion mutants was ﬁrst demonstrated in C. elegans (Jansen et al,. 1997; Westlund et al., 1999; Liu et al., 1999a). Unlike screening for insertion lines where the sequence of one end of the T-DNA or inserted transposon is known, screening for deletion requires sequence information ﬂanking the target region on both sides. In C. elegans, a high-throughput approach was developed by growing the worms and extracting genomic DNA in a 96-well microtiter plate format. DNA is pooled at different levels of complexity (pooling of mutant lines, plates, and multiple plates) and PCR primers are then designed to bracket a region of the gene for deletion detection. The PCR amplicon of a deletion mutant, despite its low frequency, would be preferentially ampliﬁed in a complex DNA pool as a small amplicon (due to deletion). Differential ampliﬁcation can be further enhanced by adjusting the extension time in the PCR procedure. Liu et al. (1999a) described a systematic PCR screen using C. elegans mutant libraries produced by four mutagens: EMS, ethylnitrosourea, diepoxyoctane, and ultraviolet-activated trimethylpsoralen. These mutagens produce deletion sizes from 700 to 2900 bp (average 1400 bp). From screening a 3-kb window for more than 100 genes, this method detected a deleted target gene per 600,000 mutagenized genomes. The approach was successfully extended to plants by Li et al. (2001). Using a population of 51,840 FN-induced Arabidopsis lines, they succeeded in detecting deletion mutants in a DNA pool containing 2592 lines. The scanning window chosen was from 3 to 17 kb, somewhat wider than that used in C. elegans. Twenty-one out of 25 (84%) genes attempted were successful. The detected deletions ranged from 0.8 to 12 kb. The authors emphasized the need to use PCR

RICE FUNCTIONAL GENOMICS

87

conditions that favored the ampliﬁcation and detection of the smaller amplicons. The screening technique was applied to a rice mutant library of 24,660 FN-induced lines and a 2.5-kb deletion in a target gene, RG1, was found. We are developing a modiﬁed PCR screening and detection protocol for the IR64 mutants (Fig. 7). More than 8000 M3 lines derived from FN and DEB were pooled (10 lines per pool, with 10 plants per line) and amplicons were observed on polyacrylamide gel to increase the sensitivity of detection. In a pilot experiment, a deletion at the PAL1 gene (phenylammonia lyase) was identiﬁed (P. Manosalva, unpublished data). By using a DNA pool of lesser complexity (1/100 dilution), the need for preferential ampliﬁcation of the small amplicons is less, but the pooling efﬁciency is lower than that reported by Li et al. (2001). Point mutations. In genetic analysis, it is often useful to have many point mutations that give a range of allelic variants. The value of a point mutation stock becomes even greater with the development of the high-throughput detection system called Targeting Induced Local Lesions IN Genome (TILLING) (McCallum et al., 2000a, 2000b). TILLING is a reverse genetics strategy based on high-throughput detection of point mutations in targeted genetic loci. This approach makes use of DNA strand mismatches formed between mutant and wild-type DNA. EMS is the preferred mutagen because it is known to produce predominantly A/T to G/C transitions, allowing the prediction of nucleotide positions that would yield missense,

Figure 7 Reverse genetics in rice using populations of plants mutagenized with fast neutron and diepoxybutane that create sizable deletions. An approach similar to that in screening insertion lines but this protocol requires primer sequences on both side of the target gene. DNA is extracted from mutagenized plants and grouped into pools. PCR reactions are performed on DNA pools using oligonucleotide primer combinations from the gene of interest. PCR products are then loaded on an agarose gel, transferred to a membrane, and hybridized to probes from the gene of interest. Detection of a smaller amplicon suggests the presence of a deletion in the mutant DNA pool. Once the deletion has been conﬁrmed by PCR, the pools can be deconvoluted into individual lines.

88

H. LEUNG AND G. AN

nonsense, or splice site mutations. DNA from individual M2 plants is isolated, pooled, and arrayed in 96-well plates. Primers are designed using CODDLE (www.proweb.org/input/) to bracket a 1-kb region that most likely contains a deleterious mutation in a target gene. The primers are then used to amplify the gene of interest and, by denaturing and re-annealing DNA, homoduplexes and heteroduplexes will be formed in the DNA pool. Originally, denatured HPLC was used to detect the presence of a DNA mismatch; however, a high-throughput system making use of enzymatic detection of heteroduplexes has proved to be more efﬁcient (Colbert et al., 2001). For TILLING to be cost-effective, mutation density (number of mutations per Mb) has to be high. It is therefore necessary to evaluate the relationship between the mutagen dose and mutation density. We ﬁrst produced a mutant population using 0.8% EMS. DNA was isolated individually from approximately 2000 M2 plants, pooled (8 genotypes per pool), and subject to screening. Of 10 genes screened, independent mutations were detected in two genes — pp2A4 encoding the serine/threonine protein phosphatase catalytic subunit and a gene encoding callose synthase. Sequencing of the mutated loci conﬁrmed that they were A/T to G/C transition mutations (B. Till, personal communication). Although the frequency of the mutations is low, the preliminary data suggest that TILLING can be a robust strategy for identifying an allelic series for any gene of interest. High-dose mutagenesis is being evaluated to produce a high mutation density in the mutant population.

C. NATURAL GENETIC VARIATION Natural variants have been selected purposefully to produce favorable agronomic traits in plant breeding. The widely used semidwarf gene sd1 in rice is a mutation in a gene encoding gibberellic acid (GA)-20 oxidase, which controls a step in gibberellin biosynthesis (Spielmeyer et al., 2002; Sasaki et al., 2002a). Mutations in Os20ox2 (280-bp deletion and a point mutation in the conserved motif of the dioxygenase) result in a semidwarf phenotype. Another example is the use of allelic variation to adjust ﬂowering time and maturity for local adaptation. The gene Hd1 controlling heading date in rice encodes a zinc ﬁnger transcription factor (homolog of CONSTANS (CO) of Arabidopsis). The “wildtype” Hd1 allele in variety Nipponbare inhibits heading under long days (natural ﬁeld condition) but promotes heading under short days. Sequence analysis showed that the Nipponbare allele has a 36-bp deletion in the zinc ﬁnger domain, which probably affects the regulatory role of the transcription factor (Yano et al., 2000). These examples suggest that many useful functional variants used in plant breeding are in fact mutant forms of ancestral “wild-type” alleles. Therefore, the natural variation present in traditional rice germplasm represents a reservoir of useful alleles awaiting discovery.

RICE FUNCTIONAL GENOMICS

89

The conventional way of investigating genes controlling quantitative variation is to ﬁrst map the QTL to a chromosomal region. For genes with large phenotypic effects, one can produce a large mapping population with sufﬁcient resolution to ﬁne-map the locus. For traits that require quantitative measurement, inbred families (such as recombinant inbred lines or advanced backcross lines) are needed for replicated trait measurement. With sequence information in hand, it is now possible to begin with a gene sequence and look for sequence variation that is likely to affect function. To some extent, genetic variation observed in traditional or modern improved varieties can be viewed as a large collection of mutations under the inﬂuence of natural or artiﬁcial selection. Many highthroughput techniques developed for forward and reverse genetics are therefore well suited for characterizing germplasm. For example, the same principle of TILLING can be applied to detecting spontaneous mutations present in the natural population (L. Comai and S. Henikoff, personal communication). In this case, instead of pooling DNA from mutants, DNA of individual germplasm accessions is annealed with a reference template DNA to detect mismatches. Once a variant is detected, the PCR fragment can be sequenced to determine the precise location of the variant.

VI. FUNCTIONAL VALIDATION OF RICE GENES So far, we have discussed the kinds of genetic materials available for discovering gene function. Without these genetic stocks, it is not possible to assign biological function. However, the utility of these genetic materials can be maximized when used with genome-wide analytical tools. It is important to note that genome-wide approaches can help identify candidate genes and pathways in a broad sense, but the data produced are often correlative in nature. Thus, standard genetic and biochemical techniques are necessary to pinpoint the functions and to dissect the pathways. Ultimately, the convergence of multiple lines of experimental evidence holds promise in understanding gene function and gene interactions. We highlight below several approaches that will probably be common tools for gene function validation in the next few years.

A. GENE EXPRESSION ANALYSIS 1.

Transcription Proﬁling

Transcription proﬁling refers to the display of a large set of genes that are expressed under speciﬁed conditions. Genetic messages from genetic stocks with different characteristics (e.g., salinity-tolerant versus sensitive) can be used to

90

H. LEUNG AND G. AN

assay which genes are expressed and the quantity of messages produced. Gene expression proﬁling also reveals what genes are regulated in coordination. If genome sequence information is available, one can look for a common sequence upstream of co-regulated genes to ﬁnd important motifs involved in gene regulation. A parallel analysis of many genes at one time can reveal which genes are collectively responsible for complex traits. While the literature on the use of microarrays and gene chips is growing, few published papers are available on genome-wide expression analysis in rice. The ﬁrst analysis was conducted by Kawasaki et al. (2000) who applied a set of 1728 cDNA libraries of salt-stressed roots obtained from salt-tolerant variety Pokkali. The expression patterns of Pokkali and sensitive variety IR29 under salt-induced conditions were compared in this set of genes. The data suggested that the most signiﬁcant difference in gene expression occurred within the initial phase of salt stress. Within 15 min after salt stress (NaCl, 150 mM), about 10% of the genes were either up-regulated or down-regulated in Pokkali. This expression pattern was delayed in sensitive variety IR29. The results suggest that the salttolerant property observed in Pokkali may in part be attributed to the rapid expression of a small set of genes, transmitted as signals leading to a downstream response. Failure to mount this rapid response, as in the case of IR29, could be responsible for the down-regulation of transcription and eventual death in 24 h. A second example of gene expression analysis focuses on nutrient partitioning during rice grain ﬁlling (Zhu et al., 2003). The objective of this study was to identify sets of co-regulated genes involved in grain ﬁlling. In this case, the array platform is an oligo chip designed to contain 16 probes (25-mer per probe) covering the 30 end of each of 21,000 predicted genes. Thirty-three samples of rice RNA representing different stages of grain ﬁlling were used as targets. Based on sequence annotation, the authors narrowed down from 21,000 genes to 491 candidate genes that are potentially involved in the synthesis and transport of carbohydrates, proteins, and fatty acids. Expression analysis led to the identiﬁcation of 269 “grain-ﬁlling genes” potentially involved in grain development. By examining the upstream regions of these genes, Zhu et al. (2003) identiﬁed a cis-element AACA that is over-represented in the promoters of 103 genes presumably involved in grain ﬁlling. The AACA element is hypothesized to interact with certain transcription factors to regulate nutrient partitioning in the grain. Both the stress tolerance and nutrient partitioning studies illustrate the power of using gene arrays to identify a core set of genes that are potentially important in regulating or being regulated in the traits of interest. The power of microarray analysis will be much enhanced when expression data are compiled and made available for data mining by the public. Efforts are underway to make available the expression data of a variety of experiments from more than 60 laboratories in Japan (http://red.dna.affrc.go.jp/RED; S. Kikuchi, personal communication).

RICE FUNCTIONAL GENOMICS

2.

91

Protein Proﬁles

Proteomic analysis of rice is still in its infancy but the few studies published so far have suggested its potential to gain a deeper understanding of gene expression in rice under various environmental and developmental conditions. Koller et al. (2002) reported a comprehensive proteomic analysis of rice using tissues from leaf, root, and seed. Complementary protein separation and analytical techniques were applied (2D PAGE followed by HPLC-tandem mass spectrometry and multidimensional protein identiﬁcation technology). A total of 2528 unique proteins was identiﬁed. About 32% of all the proteins identiﬁed in leaves, roots, and seeds could not be functionally classiﬁed based on known databases. Salekdeh et al. (2002a,b) investigated the protein proﬁles of rice plants in response to drought and salt stresses. In their studies, the 2D PAGE system can reproducibly display about 2000 proteins. Salt-tolerant variety Pokkali and sensitive IR29 were subjected to 50 and 100 mM NaCl at 14 and 21 days, respectively, and root proteins were compared. Two proteins — ascorbate peroxidase and caffeoyl-CoA O-methyltransferase — showed particularly interesting patterns in Pokkali that are distinct from those of the salt-sensitive IR29. Ascorbate peroxidase was 4.4 times more abundant in Pokkali than in IR29 in the absence of salt stress. The constitutively high level of enzyme may confer a greater antioxidant capacity to tolerate salt stress. Caffeoyl-CoA O-methyltransferase, an enzyme involved in lignin biosynthesis, was induced upon salt stress, and it may be responsible for enhancing ligniﬁcation as a mechanism to counter salt stress. In their drought-stress studies, Salekdeh et al. (2002b) compared protein proﬁles of seedlings of two genotypes, CT9993 and IR62266, under different water regimes. CT9993 is an upland variety known for its droughttolerance attributes, whereas IR62266 is a drought-sensitive lowland variety. Of more than 1000 protein spots analyzed, 42 showed a signiﬁcant change in abundance under drought stress. Four proteins — S-like RNase homologue, actin depolymerizing factor, Rubisco activase, and isoﬂavone reductase-like protein — showed distinct behavior under drought stress, suggesting that these proteins may play a role in different drought-response mechanisms not considered previously. As of today, transcript and protein proﬁling represent accessible techniques for examining gene expression on a genome scale. Undoubtedly, new and improved high-throughput techniques will become available in the coming years, but the principles of designing informative experiments remain the same. By carefully choosing the appropriate experimental conditions as well as genetic stocks such as mutants and genetically well-deﬁned near-isogenic lines, gene expression experiments can help us narrow down genes from thousands to a manageable number, whose function can be conﬁrmed by genetic and physiological experiments.

92

H. LEUNG AND G. AN

B. GENE SILENCING Post-transcriptional gene silencing (PTGS) was discovered in plants and other organisms as a means to “defend” the host cell against foreign or unwanted transcripts. This mechanism involves degradation of mRNA mediated by dsRNA sharing homology with the target sequence, resulting in the suppression of the expression of the gene (Chuang and Meyerowitz, 2000; Baulcombe, 2002). Since the process is highly sequence-speciﬁc, it can be used to selectively degrade the transcript of a chosen sequence. This targeted mRNA destruction technique, called RNA inference (RNAi), has been used to suppress gene expression in a wide range of organisms, including fungi, animals, and plants. Details of the mechanisms and genetics of RNAi can be found in several recent reviews (Waterhouse et al., 2001; Dernburg and Karpen, 2002; Hannon 2002). RNAi has been demonstrated as a powerful tool to systematically suppress gene expression in C. elegans (Fire et al., 1998). RNAi can be induced by injecting nematodes or feeding bacteria expressing dsRNA. The technology has been used to suppress genes in chromosomes I and III of C. elegans (Fraser et al., 2000; Gonczy et al., 2000). Because RNAi is sequence homology-dependent, it can be used to silence the expression of individual genes or members of a gene family by designing a gene-silencing sequence carrying a 26-bp region identical to that of the target (Parrish et al., 2000). Compared to gene knockout, gene silencing via RNAi is not always complete. This may be an advantage because partial silencing may allow the analysis of genes whose knockout mutations would cause lethality. In plants, Wesley et al. (2001) produced a series of vector constructs that encode self-complementary “hairpin” RNA (hpRNA). This modiﬁcation enhances the efﬁciency of silencing over a dsRNA construct. Furthermore, a new vector (pHELLSGATE) has been made to produce a gene library for high-throughput silencing. In rice, there is increasing interest in applying gene silencing as a complementary tool to gene knockout for investigating gene function. Shimamoto et al. (2002) applied RNAi to investigate the seven members of the OsRac gene family. By speciﬁcally targeting the 30 UTR region that is less conserved among the family members, it is possible to selectively suppress each gene member of the family. We expect that RNAi will be an important tool kit for gene validation as we encounter more and more gene families with multiple members.

C. HETEROLOGOUS BIOASSAYS A standard way to validate gene function is to transform a plant deﬁcient in a phenotype with a candidate gene to look for restored function in the transformant. Technologically, genetic complementation is no longer a problem

RICE FUNCTIONAL GENOMICS

93

in rice as various useful transformation vectors are available for performing complementation tests. The problem is more on the robustness of the system to assay a large collection of candidate genes in a high-throughput manner. Also, technical problems exist in cases where the candidate allele is recessive and where the recipient genotypes (e.g., a recessive mutant) could be compromised in vigor which makes them unsuitable for transformation. Alternative and efﬁcient ways to obtain data equivalent to genetic complementation are therefore needed. One possibility is to use a heterologous expression system in which rice genes are introduced into an alternative host and assayed for function. This is similar to the use of shuttle vectors between yeast and bacteria for testing the gene function of S. cerevisiae, a method widely used in fungal genetics. For plants, Arabidopsis probably represents the most versatile system for transformation. The fact that Arabidopsis has a comprehensive mutant collection offers a good chance that an Arabidopsis mutant can be used for complementation with a rice candidate gene. The same approach can be used to test rice genes in yeast mutants for phenotypes that can be assayed in yeast cells. The use of a virus vector to express foreign genes has been practiced for many years in dicots, but much less has been done in monocots (Scholthof et al., 1996; Karrer et al., 1998). Recently, a gene expression system has been developed for monocots using the wheat streak mosaic virus (WSMV) (Choi et al., 2000). WSMV has a broad host range, including important cereals such as wheat, maize, barley and oats. In this system, the WSMV is engineered to accept foreign genes as cDNA cassettes (e.g., candidate genes from rice). Production of in-frame transcripts is done in vitro and the infectious transcripts are used to infect wheat plants mechanically. The foreign gene carried by the virus will be expressed in the cytoplasm and transmitted systemically throughout the plant, enabling observation of the phenotypes in nearly all tissues and stages of plant growth. Phenotypes conferred by the foreign genes can be assayed under screening conditions that may last for more than a month (Fig. 8). The system has been successfully tested using the GUS gene and antibiotic resistance. Application of this approach to evaluate rice gene function is being explored (R. Choi, personal communication). Using wheat as a heterologous host for assaying rice genes can be attractive because of the similar phenology between rice and wheat. This approach may be useful for testing a large number of candidate genes with predicted functions based on sequence information or gene expression data. The use of heterologous systems for gene functional assay obviously has its limitation as not all rice genes will be properly expressed in the alternative host. Also, some genes may not give the correct phenotypes. Nonetheless, a heterologous expression system with high throughput potential can offer an efﬁcient way to produce supporting functional evidence for rice genes.

94

H. LEUNG AND G. AN

Figure 8 Wheat streak mosaic virus (WSMV) as a vehicle to assay rice gene function in wheat plants. Individual candidate genes as cDNA clones are inserted into the virus vector. Viral transcripts carrying the foreign genes are obtained by in vitro transcription and used to infect wheat plants by mechanical inoculation. The foreign genes are expressed and spread systemically along with the WSMV. Wheat plants are assayed for phenotypes within one to two months.

D. GENE REPLACEMENT When integrative vectors are introduced into plants, the vector-borne sequences nearly always integrate into the host chromosome by non-homologous recombination, a process called ectopic integration. In lower eukaryotes (e.g., fungi), however, there is a high frequency of homologous recombination when the introduced vector has a stretch of sequence identical to a sequence in the host genome. The ability to conduct site-speciﬁc homologous recombination in higher plants has been a long-sought goal because it offers perhaps the most precise form of genetic engineering by replacing one allele with another carrying the desirable effects. The problem of positional effect caused by integration into different chromosomes can also be eliminated. The recent breakthrough by Terada et al. (2002) holds great promise for adding gene replacement to the tool kit of rice functional genomics. To test the gene replacement technique, Terada et al. (2002) used the Waxy locus as the target gene because the phenotype can be easily assayed. The targeting vector pRW1 was engineered to carry the Waxy locus of rice plus selective markers (hygromycin B resistance and diphtheria toxin) for positive and negative selection of homologous recombination events. Out of about

RICE FUNCTIONAL GENOMICS

95

600 hygromycin-resistant transformants obtained, six showed precise homologous recombination between the vector and the Waxy locus, yielding a success rate of 1%. Since the vector construct is not designed for speciﬁc genes, the authors believe that the method is of general applicability to other rice genes. If the system is robust enough, any gene sequence, native or modiﬁed, can be used to replace the resident gene. Thus, the technology is not only suitable for validating the function of a gene but also suitable for creating the desirable allelic state at a speciﬁed locus.

E. ALLELIC SERIES Direct functional tests by genetic complementation can be substituted if there is a large series of allelic mutations at the candidate gene locus. A perfect correlation between phenotypic changes and multiple alleles at the locus can provide convincing evidence that the candidate gene is responsible for a certain phenotype. This approach was used in proving the identity and function of the Mlo locus in barley (Buschges et al., 1997). With the large-scale production of mutant lines worldwide, allelic series will be increasingly useful for conﬁrming the function of gene sequences.

VII. APPLICATIONS TO CROP IMPROVEMENT Given the large number of traits amenable for improvement using genomic tools, it will not be possible to describe all of them in meaningful detail. Instead, it is more relevant to discuss what tools and approaches will become common in plant improvement programs when the functions of most rice genes are known. We can think of three important ways in which crop improvement programs will beneﬁt from having a functional dictionary of rice genes. First, with the identiﬁcation of actual genes controlling certain traits, we will shift from using linked markers to actual genes as units of selection. Second, we will make use of the knowledge of genetic regulation and biochemical pathways to create genotypes that confer large phenotypic effects. We will increasingly be able to recognize genetic switches or regulatory circuitry that could lead to a signiﬁcant change in phenotype. This will drastically improve selection theories that are primarily based on quantitative genetics models with little integration of genetic knowledge in biochemical and metabolic terms. Third, sequence-guided comparative biology will maximize the use of knowledge of gene function across related or distant plant species. These approaches can be used in combination to solve problems of a varying degree of complexity.

96

H. LEUNG AND G. AN

A. CANDIDATE GENE APPROACH

AND

ALLELE MINING

Although marker-aided selection (MAS) has been used successfully to accumulate genes with large effects, its use for selecting quantitative traits has not been widespread. The main reason is that most of the markers available are loosely linked to the target gene and are easily uncoupled by recombination. As functional annotation of the genome sequence continues to improve, we expect that candidate gene approach to be widely adopted by breeding programs. Immediate gain will come from the selection of traits such as tolerance for biotic stresses where there is a good knowledge of host defense genes (Ramalingam et al., 2003). Conventionally, determining the phenotypic contribution of a particular allele is based on setting up a cross between lines with high and low trait values and examining the segregating progeny. Because genetic mapping experiments often take 1 – 2 years to complete, functional association through genetic mapping is not satisfactory for examining multiple traits at the same time. The ability to establish a function– structure relationship at the level of nucleotide variation for target alleles across a large number of genetic loci would be of great value in breeding (Drenkard et al., 2000; Rafalski, 2002). Several studies in maize have demonstrated the potential of association genetics by examining sequence variation to establish structure –function relationships (Remington et al., 2001; Thornsberry et al., 2001; Vigouroux et al., 2002; Whitt et al., 2002). If candidate gene loci are known, it will be possible to identify sequence variants at the loci and observe patterns of structure –function association — an emerging area called allele mining. Implicit in allele mining is that the gene pool is so large that it is not feasible to conduct controlled genetic crosses to determine the function of each allele. The gene pool could be represented by uncharacterized raw germplasm or advanced breeding lines in a breeding program. One may consider both forward and reverse genetics approaches to search for useful alleles. Forward genetics is applicable to advanced germplasm that is known to be associated with an improved trait. The advanced germplasm represents an enriched pool of useful alleles and, by examining the shift in allele frequency of certain genes under selection, one can detect a pattern of association that gives meaningful prediction of allele function. If breeding records and pedigree information are well kept, it is possible to trace the transmission of favorable alleles from the donor germplasm that gives rise to improved varieties. The second set of materials represents the “raw” germplasm that has not been widely used and whose phenotypes are largely unknown. In this case, the reverse genetics approach can be used to examine a stratiﬁed collection of germplasm grouped by geographic origin or historical performance data. Using a known candidate gene sequence, the appropriate region of the gene sequence can be ampliﬁed from germplasm and examined for association with phenotypes. Several association mapping methods have been developed for such analyses

RICE FUNCTIONAL GENOMICS

97

(Rigoutsos and Floratos, 1998; Pritchard and Rosenberg, 1999; Pritchard et al., 2000). As more knowledge is accumulated on the function of various sequence motifs and gene structures, functional inference based on association genetics will continue to improve.

B. PATHWAYS

AND

GENETIC REGULATION

We envision that a signiﬁcant outcome of functional genomics research will be the discovery of genes that can contribute a large change in the value of breeding targets. These include new pathways or biochemical mechanisms not obvious in current rice genotypes. In this context, genes that determine the hierarchical regulation of pathways are potential targets. A major class of such genes are those encoding transcription factors that control genetic switch points affecting multiple downstream genes. In Arabidopsis, about 1700 genes belong to members of known classes of transcription factors and there are 1300 such genes in rice (The Arabidopsis Genomics Initiative, 2000; Goff et al., 2002). Manipulation of regulatory elements can bring about dramatic changes in phenotypes that are often viewed as being controlled by many genes with minor effects. The dramatic effect of transcriptional factors is well illustrated by the CBFs (C-repeat binding factors) and the related DREB (drought responsive element binding) factors. These factors control multiple downstream genes that confer tolerance for cold, salt, and drought stress (Stockinger et al., 1997; Jaglo-Ottosen et al.,1998; Kasuga et al., 1999; Thomashaw 1999). Sequences with speciﬁc domains/motifs can be extracted from public plant genomics databases, and families of transcription factors can be categorized phylogenetically and putative gene functions inferred based on sequence similarity and experimental data. Chen et al. (2002) examined the expression patterns of 402 transcription factor genes in Arabidopsis using a variety of defense-response mutants and plants under different stress conditions. The results revealed coordinated regulation of speciﬁc “regulons”. Transcription factors with putative roles on target traits can be further assayed by overexpression and knockout mutations. Expansion of these studies will open up opportunities for the manipulation of transcription control in plant breeding (Zhu et al., 2003). We can expect rapid progress in this area as all ingredients for conducting these experiments are already in place in rice.

C. CROSS-SPECIES INFERENCE

OF

GENE FUNCTION

We expect comparative genetics to become an integral part of future plant improvement programs. Comparative analysis between genomes will broaden

98

H. LEUNG AND G. AN

the “gene pool” for trait manipulation in rice. Here, “gene pool” refers to the pool of genetic knowledge. Kaplinsky et al. (2002) showed that comparison of conserved and diverged sequences in the non-coding regions of genes in different grass genomes can lead to the identiﬁcation of critical regions of genes that control function. Similarity in function will enable us to extend knowledge about one species to another. A main objective of comparative genetic analyses is to test whether orthologous genes (of common descent) function similarly in different species. With more knowledge accumulated in allied species, we will increasingly look beyond rice for a solution to genetic and breeding problems in rice. For example, the introgression of wild species alleles from wild to domesticated rice is limited by a lack of chromosome pairing between AA and alien genomes. In wheat, the Ph1 locus controls the pairing of homoeologous chromosomes of different wheat genomes. The wild-type Ph1 gene, located on chromosome 5B, prevents the pairing of homoeologous chromosomes so that only homologous chromosomes of each genome can pair in meiosis. In the absence of the Ph1 gene, homoeologous chromosomes can pair, allowing genetic recombination between different genomes. By comparative mapping, Foote et al. (1997) found a region on rice chromosome 9 syntenic to the wheat 5B region carrying Ph1. Using deletion mapping, Robert et al. (1999) narrowed down the region to 4 Mb that potentially carry an ortholog of the Ph1 gene in rice. We can now ask whether an orthologous Ph1 gene indeed exists in rice, and, if so, can we manipulate the gene to facilitate gene transfer between wild and domesticated rice (D. Brar, personal communication)? This sort of cross-species inquiry will likely be routinely used in future plantimprovement programs. Even if the hypothesis of similar function is not supported, the ﬁnding of distinct functions in genes of high sequence similarity points to exciting new avenues for investigation. Shimamoto and Kyozuka (2002) recently reviewed the genes controlling disease-response signaling and reproductive development in rice and their counterparts in related species. They concluded that, between rice and Arabidopsis, sequence similarities do not always imply functional similarity. Orthologous genes in different species can play a divergent role. Changes in function and regulation of orthologous genes could be the basis for divergence of plant species.

VIII. INTERNATIONAL COLLABORATION AND THE ROLE OF DEVELOPING COUNTRIES Given the resources already available or being developed by different national programs, the scenario of ﬁnding all agronomically important rice genes by 2010

RICE FUNCTIONAL GENOMICS

99

seems realistic. But the agenda is large and it requires diverse inputs. Unlike sequencing projects that can be carried out by a few laboratories with high capacity in sequencing and bioinformatics, functional genomics requires a wide range of experience from the cell to the whole plant to crop biology. To assign functions to gene sequences, we must have the means and skills to examine a wide range of phenotypes. Collaboration in rice functional genomics is therefore a necessity if we are to achieve the goal of ﬁnding the function of all important rice genes in a timely manner. Besides technical considerations, there is an impetus to ensure that modern science is put to use in solving difﬁcult production problems, particularly in the developing world, where food security and access are a major concern. Despite the impressive gains in rice productivity over the past four decades, yield in marginal agricultural areas has remained low. The development of adaptive, high-yielding, and nutritious varieties must therefore be part of a strategy for improving the livelihood of people living in these unfavorable regions (Hossain et al., 2000). To solve these problems, there must be open access to genetic knowledge and empowering tools. Collaboration in gene discovery ensures such access. A useful collaborative model in plant science has been provided by the Multinational Arabidopsis Consortium (www.Arabidopsis.org/info/ 2010_projects/MASC_info.html). The Arabidopsis consortium has participation by 21 members from 13 countries where signiﬁcant programs in Arabidopsis research are taking place. It serves to promote communication and the exchange of data and to coordinate research activities to maximize efﬁciency. The goal of the Arabidopsis community is to ﬁnd out the function of all genes in the plant by 2010. A similar goal is envisioned by the rice research community (Fig. 9). But rice is different from Arabidopsis; many traits have to be measured under realistic agronomic conditions. A rice consortium therefore needs to incorporate the practical aspects of plant improvement and at the same time make use of the expertise and experience from rice-breeding institutions around the world. In this context, the roles of developing countries become important although they often lack the infrastructure for genomics research. Rice-breeding institutions have the genetic resources, phenotyping capacity, and breeding and evaluation networks for testing gene function and genotype £ environment interactions. Breeding institutions also have the capacity to assemble genes in varieties that will be accepted by farmers and consumers. Finally, genomics, because of the excitement and publicity involved, provides a fertile ground to attract young students in the developing world to enter into plant science research, hence building the human resources for future agricultural research and development. Efforts are already under way to forge international collaboration in rice functional genomics. Since the beginning of the international sequencing project, a consortium approach to accelerate rice gene discovery is being promoted by

100

H. LEUNG AND G. AN

Figure 9 Functional characterization of rice genes will be the challenge after completion of the rice genome sequence. Delivery of research results for the development of adaptive and nutritious varieties can be accelerated by international collaboration within this decade.

the International Rice Functional Genomics Working Group organized by IRRI (Fischer et al., 2000). With the completion of the rice genome sequence, there is consensus that a more structured consortium is needed to facilitate collaboration and sharing of information and resources. In January 2003, the International Rice Functional Genomics Consortium was formed with scientists from 17 institutions of 12 countries exploring ways to consolidate resources and build common strategies (http://www.IRIS.IRRI.org/IRFGC).

IX. CONCLUDING REMARKS Recognizing the rapid progress in plant genomics, this chapter does not attempt to provide a comprehensive review of the past and on-going work related to rice genomics. Rather, we provide a snapshot of the current status of the ﬁeld and focus on the framework needed to make gene discovery relevant to plant improvement. Rice is special because of its botanical history and biological attributes. The completion of the rice genome sequence is a major milestone in plant science; however, for the sequence information to be useful, it must be converted into a format that allows investigation with a variety of genetic materials. Functional genomics concerns the development of a resource platform

RICE FUNCTIONAL GENOMICS

101

by which the function of every gene and its interaction with other genes can be characterized and understood in an efﬁcient manner. The excitement in functional genomics lies in the creative application of high-throughput tools to probe into the biological complexity of genetic networks and biological processes. In the post-genome-sequencing era, we may save efforts in gene cloning in the conventional sense, but there is no substitute for detailed studies involving genetics, physiology, and biochemistry if we are to understand the biological functions of genes. Because rice is a major food crop in many societies and cultures, discovering rice genes and having access to the knowledge inevitably has a social impact. Starting in the 1960s, the Green Revolution has dramatically increased the food supply in the developing world (Evenson et al., 1996; Hossain 1996). The yield increase in the major cereals has largely come from genetic improvement of the plants supported by improved agronomic practices. The higher efﬁciency in production has reduced the rice price to about 50% in real terms over the past three decades (Khush, 2001; M. Hossain, IRRI, personal communication). Despite this impact, the problem of access to technology and resources-improved seeds, fertilizer-has fueled debates about the achievements of the Green Revolution. As we move forwards with the large-scale discovery of rice genes — the ingredients for plant improvement — we must be mindful of public access to the empowering knowledge and tools. In this chapter, we have highlighted the resources available and being developed to accelerate gene discovery and we also speculated on the future adoption of new knowledge and tools in plant improvement programs. Developing countries have a unique role to play in rice functional genomics because they are rich in genetic resources and agronomic knowledge of the rice crop. The beneﬁts of engaging developing countries in gene discovery are obvious but this requires the will and efforts to support the partnerships. The power of bringing together a full range of expertise in basic and applied research is enormous. The move toward more international collaboration in rice functional genomics is very encouraging. Success in this effort will rest upon the continuing goodwill and determination of the research community to sustain the spirit of collaboration to channel gene discovery into practical crop improvement.

ACKNOWLEDGMENTS We thank Alice Bordeos and Ramil Mauleon for assistance in the preparation of this manuscript, Il Ryong Choi for providing Fig. 8 and valuable discussion, and Bill Hardy for technical editing. We are grateful to Rebecca Nelson and Jan Leach for their helpful comments and to many colleagues who share their knowledge on the topic.

102

H. LEUNG AND G. AN

REFERENCES Agrawal, G. K., Yamazaki, M., Kobayashi, M., Hirochika, R., Miyao, A., and Hirochika, H. (2001). Screening of the rice viviparous mutants generated by endogenous retrotransposon Tos17 insertion. Tagging of a zeaxanthin epoxidase gene and a novel OsTATC gene. Plant Physiol. 125, 1248–1257. Aguirrezabalaga, I., Nivard, M. J. M., Comendador, M. A., and Vogel, E. W. (1995). The crosslinking agent hexamethylphosphoramide predominantly induces intra-locus and multi-locus deletions in postmeiotic germ cells of Drosophila. Genetics 139, 649–658. Ahn, S., and Tanksley, S. D. (1992). Comparative linkage maps of the rice and maize genomes. Proc. Natl Acad. Sci. USA 90, 7980–7984. Azpiroz-Leehan, R., and Feldmann, K. A. (1997). T-DNA insertion mutagenesis in Arabidopsis: going back and forth. Trends Genet. 13, 152 –156. Balzergue, S., Dubreucq, B., Chauvin, S., Le-Clainche, I., Le Boulaire, F., de Rose, R., Samson, F., Biaudet, V., Lecharny, A., Cruaud, C., Weissenbach, J., Caboche, M., and Lepiniec, L. (2001). Improved PCR-walking for large-scale isolation of plant T-DNA borders. BioTechniques 30, 496–504. Barakat, A., Gallois, P., Raynal, M., Mestre-Ortega, D., Sallaud, C., Guiderdoni, E., Delseny, M., and Bernardi, G. (2000). The distribution of T-DNA in the genomes of transgenic Arabidopsis and rice. FEBS Lett. 471, 161 –164. Barthels, N., van der Lee, F. M., Klap, J., Goddijn, O. J., Karimi, M., Puzio, P., Grundler, F. M., Ohl, S. A., Lindsey, K., Robertson, L., Robertson, W. M., van Montagu, M., Gheysen, G., and Sijmons, P. C. (1997). Regulatory sequences of Arabidopsis drive reporter gene expression in nematode feeding structures. Plant Cell 9, 2119– 2134. Baulcombe, D. (2002). RNA silencing. Curr. Biol. 12, R82–R84. Bennetzen, J., and Freeling, M. (1997). Review: the uniﬁed grass genome: synergy in synteny. Genome Res. 7, 301–306. Bonierbale, M. W., Plaisted, R. L., and Tanksley, S. D. (1988). RFLP maps based on a common set of clones reveal modes of chromosomal evolution in potato and tomato. Genetics 120, 1095– 1103. Brar, D. S., and Khush, G. S. (1997). Alien introgression in rice. Plant Mol. Biol. 35, 35–47. Brar, D. S., and Khush, G. S. (2002). Transferring genes from wild species into rice. In “Quantitative Genetics, Genomics and Plant Breeding” (M. S. Kang, Ed.), pp. 197–217. CAB International, London. Brown, P. O., and Botstein, D. (1999). Exploring the new world of the genome with DNA microarrays. Nat. Genet. 21, 33–37. Bruggemann, E., Handwerger, K., Essex, C., and Storz, G. (1996). Analysis of fast neutron-generated mutants at the Arabidopsis thaliana HY4 locus. Plant J. 10, 755– 760. Buell, R. (2002). Current status of the sequence of the rice genome and prospects for ﬁnishing the ﬁrst monocot genome. Plant Physiol. 130, 1585–1586. Buschges, R., Holricher, K., Panstruga, R., Simons, G., Wolter, M., Frijters, A., van Daelen, R., van der Lee, T., Diergaarde, P., Groenendijk, J., Topsch, S., Vos, P., Salamini, F., and Schulze-Lefert, P. (1997). The barley Mlo gene: a novel control element of plant pathogen resistance. Cell 88, 695–705. Chang, T. T., Pan, Y., Chu, Q., Peiris, R., and Loresto, G. C. (1991). Cytogenetic, electrophoretic, and root studies of javanica rices. In “Rice Genetics II” (G. S. Khush, D. S. Brar, and B. Hardy, Eds.), pp. 21– 32. International Rice Research Institute, Los Ban˜os, Philippines. Chang, H. S., Wu, C., Zeng, L., Dunn, M., Wang, G. L., Leung, H., Goff, S., Wang, X., Zhu, T., and Leach, J. E. (2003). Detection of deleted genes in rice mutants using the Rice GeneChip genome array, Page 100 in Abstracts of Plant and Animal Genomes XI. January 11 –15, 2003, San Diego, California.

RICE FUNCTIONAL GENOMICS

103

Charng, Y.-C., Pﬁtzner, A. J. P., Pﬁtzer, U. M., Charng-Chang, K.-F., Chen, C., Tu, J., and Kuo, T.-T. (2000). Construction of an inducible transposon, INAc, to develop a gene tagging system in higher plants. Mol. Breed. 6, 353 –367. Chen, W., Provart, N., Glazebrook, J., Katagiri, F., Chang, H. S., Eulgem, T., Mauch, F., Luan, S., Zou, G., Whitnam., S., Budworth, P. R., Tao, Y., Xie, Y., Chen, X., Lam, S., Kreps, J., Harper, J., Si-Ammour, A., Mauch-Mani, B., Heinlein, M., Kobayashi, K., Hohn, T., Dangl, J., Wang, X., and Zhu, T. (2002). Expression proﬁle matrix of Arabidopsis transcription factor genes suggests their putative functions in response to environmental stresses. Plant Cell 14, 559– 574. Cheng, X., Sardana, R., Kaplan, H., and Altosaar, I. (1998). Agrobacterium-transformed rice plants expressing synthetic cry1A(b) and cry1A(c) genes are highly toxic to striped stemborer and yellow stemborer. Proc. Natl Acad. Sci. USA 95, 2767–2772. Chin, H. G., Choe, M. S., Lee, S. H., Park, S. H., Koo, J. C., Kim, N. Y., Lee, J. J., Oh, B. G., Yi, G. H., Kim, S. C., Choi, H. C., Cho, M. J., and Han, C. (1999). Molecular analysis of rice plants harboring an Ac/Ds transposable element-mediated gene trapping system. Plant J. 19, 615–623. Cho, R. J., Mindrinos, M., Richards, D. R., Sapolsky, R. J., Anderson, M., Drenkard, E., Dewdney, J., Beuber, T. L., Stemmers, M., Federspiel, N., Theologis, A., Yang, W.-H., Hubbell, E., Au, M., Chung, E. Y., Lashkari, D., Lemieux, B., Dean, C., Lipshutz, R. J., Ausubel, F. M., Davis, R. W., and Oefner, P. J. (1999). Genome-wide mapping with biallelic markers in Arabidopsis thaliana. Nature Gen. 23, 203–207. Choi, I.-R., Stenger, D. C., Morris, J., and French, R. (2000). A plant virus vector for systemic expression of foreign genes in cereals. Plant J. 23, 547 –555. Christou, P. (1997). Rice transformation: bombardment. Plant Mol. Biol. 35, 197 –203. Chuang, C. F., and Meyerowitz, E. M. (2000). Speciﬁc and heritable genetic interference by doublestranded RNA in Arabidopsis thaliana. Proc. Natl Acad. Sci. USA 97, 4985–4990. Chung, B. C., Kim, J. K., Nahm, B. H., and Lee, C. H. (2000). In planta visual monitoring of green ﬂuorescent protein in transgenic rice plants. Mol. Cells 10, 411–414. Colbert, T., Till, B. J., Tompa, R., Reynolds, S., Steine, M. N., Yeung, A. T., McCallum, C. M., Comai, L., and Henikoff, S. (2001). High-throughput screening for induced point mutations. Plant Physiol. 126, 480 –484. Crane, P. R., Friis, E. M., and Raunsgaard-Pedersen, K. (1995). The origin and early diversiﬁcation of angiosperms. Nature 374, 27–33. Dai, S., Zheng, P., Marmey, P., Zhang, S., Tian, W., Chen, S., Beachy, R. N., and Fauquet, C. (2001). Comparative analysis of transgenic rice plants obtained by Agrobacterium-mediated transformation and particle bombardment. Mol. Breed. 7, 25–33. DeRisi, J. L., Iyer, V. R., and Brown, P. O. (1997). Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 278, 680 –686. Dernburg, A. F., and Karpen, G. H. (2002). A chromosome RNAissance. Cell 111, 159 –162. Devos, K. M., and Gale, M. D. (2000). Genome relationships: the grass model in current research. Plant Cell 12, 637 –646. Drenkard, E., Richter, B. G., Rozen, S., Stutius, L. M., Angell, N. A., Mindrinos, M., Cho, R. J., Oefner, P. J., Davis, R. W., and Ausubel, F. M. (2000). A simple procedure for the analysis of single nucleotide polymorphisms facilitates map-based cloning in Arabidopsis. Plant Physiol. 124, 1483–1492. Enoki, H., Izawa, T., Kawahara, M., Komatsu, M., Koh, S., Kyozuka, J., and Shimamoto, K. (1999). Ac as a tool for the functional genomics for rice. Plant J. 19, 605 –613. Evenson, R. E., Herdt, R. W., and Hossain, M. (1996). “Rice research in Asia: progress and priorities”, p. 418, CAB International, Wallingford, UK. Farris, J. D., Haen, K. M., and Gill, B. S. (2000). Saturation mapping of a gene-rich recombinant hot spot region in wheat. Genetics 154, 823– 835. Fedoroff, N. V., Wessler, S., and Shure, M. (1983). Isolation of the transposable maize controlling elements Ac and Ds. Cell 35, 235 –242.

104

H. LEUNG AND G. AN

Fedoroff, N. V., Furtek, D. B., and Nelson, O. E. (1984). Cloning of the bronze locus in maize by a simple and generalizable procedure using the transposable element Activator (Ac). Proc. Natl Acad. Sci. USA 81, 3825–3829. Feng, Q., Zhang, Y., Hao, P., Wang, S., Fu, G., Huang, Y., Li, Y., Zhu, J., Liu, Y., Hu, X., Jia, P., Zhang, Y., Zhao, Q., Ying, K., Yu, S., Tang, Y., Weng, Q., Zhang, L., Lu, Y., Mu, J., Lu, Y., Zhang, L. S., Yu, Z., Fan, D., Liu, X., Lu, T., Li, C., Wu, Y., Sun, T., Lei, H., Li, T., Hu, H., Guan, J., Wu, M., Zhang, R., Zhou, B., Chen, Z., Chen, L., Jin, Z., Wang, R., Yin, H., Cai, Z., Ren, S., Lv, G., Gu, W., Zhu, G., Tu, Y., Jia, J., Zhang, Y., Chen, J., Kang, H., Chen, X., Shao, C., Sun, Y., Hu, Q., Zhang, X., Zhang, W., Wang, L., Ding, C., Sheng, H., Gu, J., Chen, S., Ni, L., Zhu, F., Chen, W., Lan, L., Lai, Y., Cheng, Z., Gu, M., Jiang, J., Li, J., Hong, G., Xue, Y., and Han, B. (2002). Sequence and analysis of rice chromosome 4. Nature 420, 316–320. Feuillet, C., and Keller, B. (2002). Comparative genomics in the grass family: molecular characterization of grass genome structure and evolution. Ann. Bot. 89, 3– 10. Fire, A., Xu, S., Montgomery, M. K., Kostas, S. A., Driver, S. E., and Mello, C. C. (1998). Potent and speciﬁc genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391, 806–811. Fischer, K. S., Barton, J., Khush, G. S., Leung, H., and Cantrell, R. (2000). Collaboration in rice. Science 290, 279 –280. Fodor, S. P. A., Read, J. L., Pirrung, M. C., Styryer, L., Lu, A. T., and Solas, D. (1991). Light-directed, spatially addressable parallel chemical synthesis. Science 251, 767 –773. Foote, T., Roberts, M., Kurata, N., Sasaki, T., and Moore, G. (1997). Detailed comparative mapping of cereal chromosome regions corresponding to the Ph 1 locus in wheat. Genetics 147, 801– 807. Fraser, A. G., Kamath, R. S., Zipperlen, Martinez-Campos, M., Sohrmann, M., and Ahringer, J. (2000). Functional genomic analysis of C. elegans chromosome 1 by systematic RNA interference. Nature 408, 325–330. Freeling, M. (2001). Grasses as a single genetic system: reassessment. Plant Physiol. 15, 1191– 1197. Fujisawa, Y., Kato, T., Ohki, S., Ishikawa, A., Kitano, H., Sasaki, T., Asahi, T., and Iwasaki, Y. (1999). Suppression of the heterotrimetric G protein causes abnormal morphology, including dwarﬁsm in rice. Proc. Natl Acad. Sci. USA 96, 7575–7580. Gaut, B. (2002). Evolutionary dynamics of grass genomes. New Physiol. 154, 15–28. Ge, S., Sang, T., Lu, B. R., and Hong, D. U. (1999). Phylogeny of rice genomes with emphasis on origins of allotetraploid species. Proc. Natl Acad. Sci. USA 96, 14400–14405. Ge, S., San, T., Lu, B-R., and Hong, D.-Y. (2001). Phylogeny of the genus Oryza as revealed by molecular approaches. In “Rice Genetics IV” (G. S. Khush, D. S. Brar, and B. Hardy, Eds.), pp. 89–105. New Delhi Science Publishers Inc. and International Rice Research Institute, Los Ban˜os, Philippines. Gill, K. S., Gill, B. S., Endo, T. R., and Boiko, E. V. (1996). Identiﬁcation and high density mapping of gene-rich regions in chromosome group 5 of wheat. Genetics 143, 1001–1012. Goff, S. A., Ricke, D., Lan, T. H., Presting, G., Wang, R., Dunn, M., Glazebrook, J., Sessions, A., Oeller, P., Varma, H., Hadley, D., Hutchison, D., Martin, C., Katagiri, F., Lange, B. M., Moughamer, T., Xia, Y., Budworth, P., Zhong, J., Miguel, T., Paszkowski, U., Zhang, S., Colbert, M., Sun, W., Chen, L., and Cooper, B. (2002). A draft sequence of the rice genome (Oryza sativa L. ssp. japonica). Science 296, 92–100. Gonczy, P., Echeverri, C., Oegema, K., Coulson, A., Jones, S. J. M., Copley, R. R., Duperon, J., Oegema, J., Brehm, M., Etienne, C., Hannak, E., Kirkham, M., Pichler, S., Flohrs, K., Goessen, A., Leidel, S., Alleaume, A.-M., Martin, C., Ozlu, N., Bork, P., and Hyman, A. A. (2000). Nature 408, 331 –336. Hanley, S., Edwards, D., Stevenson, D., Haines, S., Hegarty, M., Schuch, W., and Edwards, K. J. (2000). Identiﬁcation of transposon-tagged genes by the random sequencing of mutator-tagged DNA fragments from Zea mays. Plant J. 23, 557–566. Hannon, G. J. (2002). RNA interference. Nature 418, 244–251.

RICE FUNCTIONAL GENOMICS

105

Hayashi, H., Czaja, I., Lubenow, H., Schell, J., and Walden, R. (1992). Activation of a plant gene by T-DNA tagging: auxin-independent growth in vitro. Science 258, 1350–1353. Hieter, P., and Boguski, M. (1997). Functional genomics: it’s all how you read it. Science 278, 601 –602. Hiei, Y., Ohta, S., Komari, T., and Kumashiro, T. (1994). Efﬁcient transformation of rice (Oryza sativa L.) mediated by Agrobacterium and sequence analysis of the boundaries of the T-DNA. Plant J. 6, 271– 282. Hirochika, H., Otsuki, H., Yoshikawa, M., Otsuki, Y., Sugimoto, K., and Takeda, S. (1996). Autonomous transposition of the tobacco retrotransposon Tto1 in rice. Plant Cell 8, 725–734. Hirochika, H. (2001). Contribution of the Tos17 retrotransposon to rice functional genomics. Curr. Opin. Plant Biol. 4, 118–122. Hossain, M. (1996). Recent developments in the Asian rice economy: challenges for rice research. In “Rice Research in Asia: Progress and Priorities” (R. E. Evenson, R. W. Herdt, and M. Hossain, Eds.), pp. 17 –33. CAB International, London and International Rice Research Institute, Los Ban˜os, Philippines. Hossain, M. (1998). Sustaining food security in Asia: economic, social, and political aspects. In “Sustainability of Rice in the Global Food System” (N. G. Dowling, S. M. Greenﬁeld, and K. S. Fischer, Eds.), pp. 19 –43. Paciﬁc Basin Center, USA, and International Rice Research Institute, Los Ban˜os, Philippines. Hossain, M., Bennett, J., Datta, S., Leung, H., and Khush, G. S. (2000). Biotechnology research in rice for Asia: priorities, focus and directions. In “Agricultural Biotechnology in Developing Countries: Towards Optimizing the Beneﬁts of the Poor” (M. Qaim, et al., Eds.), pp. 99–120. Kluwer Academic Publishers, Netherlands. Hughes, T. R., Mao, M., Jones, A. R, Burchard, J., Marton, M. J., Shannon, K. W., Lefkowitz, S. M., Ziman, M., Schelter, J. M., Meyer, M. R., Kobayashi, S., Davis, C., Dai, H., He, Y. D., Stephaniants, S. B., Cavet, G., Walker, W. L., West, A., Coffey, E., Shoemaker, D. D., Stoughton, R., Blanchard, A. P., Friend, S. H., and Linsley, P. S. (2001). Expression proﬁling using microarrays fabricated by an ink-jet oligonucleotide synthesizer. Nat. Biotechnol. 19, 342 –347. Ito, T., and Meyerowitz, E. M. (2000). Overexpression of a gene encoding a cytochrome P450, CYP78A9, induces large and seedless fruit in Arabidopsis. Plant Cell 12, 1541–1550. Ito, T., Motohashi, R., Kuromori, T., Mizukado, S., Sakurai, T., Kanahara, H., Seki, M., and Shinozaki, K. (2002). A new resource of locally transposed dissociation elements for screening gene-knockout lines in silico on the Arabidopsis genome. Plant Physiol. 129, 1695– 1699. Izawa, T., Miyazaki, C., Yamamoto, M., Terada, R., Iida, S., and Shimamoto, K. (1991). Introduction and transposition of the maize transposable element Ac in rice (Oryza sativa L.). Mol. Gen. Genet. 227, 391–396. Izawa, T., Ohnishi, T., Nakano, T., Ishida, N., Enoki, H., Hashimoto, H., Itoh, K., Terada, R., Wu, C., Miyazaki, C., Endo, T., Iida, S., and Shimamoto, K. (1997). Transposon tagging in rice. Plant Mol. Biol. 35, 219–229. Jaccoud, D., Peng, K., Feinstein, D., and Kilian, A. (2001). Diversity arrays: a solid state technology for sequence information independent genotyping. Nucleic Acids Res. 29(4), e25. Jaglo-Ottosen, K. R., Gilmour, S. J., Zarka, D. G., Schabenberger, O., and Thomashaw, M. F. (1998). Arabidopsis CBF1 overexpression induces COR genes and enhances freezing tolerance. Science 280, 104–106. Jansen, G., Hazendonk, E., Thijssen, K. L., and Plasterk, R. H. (1997). Reverse genetics by chemical mutagenesis in Caenorhabditis elegans. Nature Gen. 17, 119–121. Jefferson, R. A., Kavanagh, T. A., and Bevan, M. W. (1987). GUS fusions: b-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J. 6, 3901–3907. Jeon, J.-S., Lee, S., Jung, K.-H., Jun, S.-H., Jeong, D.-H., Lee, J., Kim, C., Jang, S., Yang, K., Nam, J., An, K., Han, M.-J., Sung, R.-J., Choi, H.-S., Yu, J.-H., Choi, J.-H., Cho, S.-Y., Cha, S.-S.,

106

H. LEUNG AND G. AN

Kim, S.-I., and An, G. (2000). T-DNA insertional mutagenesis for functional genomics in rice. Plant J. 22, 561–570. Jeong, D.-H., An, S., Kang, H.-G., Moon, S., Han, J., Park, S., Lee, H. S., An, K., and An, G. (2002). T-DNA insertional mutagenesis for activation tagging in rice. Plant Physiol. 130, 1636– 1644. Jones, M. P., Dingkuhn, M., Aluko, G. K., and Semon, M. (1997). Interspeciﬁc Oryza sativa L. £ O. glaberrima Steud. progenies in upland rice improvement. Euphytica 92, 237–246. Kane, M. D., Jatkoe, T. A., Stumpf, C. R., Lu, J., Thomas, J. D., and Madore, S. J. (2000). Assessment of the sensitivity and speciﬁcity of oligonucleotide (50 mer) microarrays. Nucleic Acids Res. 28, 4552–4557. Kaplinsky, N. J., Braun, D. M., Penterman, J., Goff, S. A., and Freeling, M. (2002). Utility and distribution of conserved non-coding sequences in the grasses. Proc. Natl Acad. Sci. USA 99, 6147–6151. Kardailsky, I., Shukla, V. K., Ahn, J. H., Dagenais, N., Christensen, S. K., Nguyen, J. T., Chory, J., Harrison, M. J., and Weigel, D. (1999). Activation tagging of the ﬂoral inducer FT. Science 286, 1962–1965. Karrer, E. E., Beachy, R. N., and Holt, C. A. (1998). Cloning of tobacco genes that elicit the hypersensitive response. Plant Mol. Biol. 36, 681 –690. Kasuga, M., Liu, Q., Miura, S., Yamaguchi-Shinozaki, K., and Shinozaki, K. (1999). Improving plant drought, salt, and freezing tolerance by gene transfer of a single stress-inducible transcription factor. Nat. Biotechnol. 17, 287–291. Katagiri, S. (2003). Comparative genomics of rice chromosome 1: map construction and preliminary genome sequencing of an indica variety, Kasalath, In “Abstracts of Rice Genome Workshop. February 5 – 7, 2003”, p. 13. Rice Genome Research, Program, National Institute of Agrobiological Resources/STAFF Institute, Tsukuba, Japan. Katayama, T. (1997). Relationships between chromosome numbers and genomic constitutions in genus Oryza. In “Science of the Rice Plant” (T. Matsuo, Y. Futsuhara, F. Kikuchi, and H. Yamaguchi, Eds.), pp. 39–48. Food and Agriculture Policy Research Center, Tokyo. Kellogg, E. A. (1998). Relationships of cereal crops and other grasses. Proc. Natl Acad. Sci. USA 95, 2005–2010. Kellogg, E. A. (2001). Evolutionary history of the grasses. Plant Physiol. 125, 1198–1205. Kawasaki, S., Borchert, C., Deyholos, M., Wang, H., Brazille, S., Kawai, K., Galbraith, D., and Bohnert, H. (2000). Gene expression proﬁles during the initial phase of salt stress in rice. Plant Cell 13, 889–905. Khush, G. S. (1977). Disease and insect resistance in rice. Adv. Agron. 29, 265–341. Khush, G. S. (2001). Green revolution: the way forward. Nat. Rev. Genet. 2, 815 –822. Koller, A., Washburn, M. P., Lange, B. M., Andon, N. L., Deciu, C., Haynes, P. A., Hays, L., Schieltz, D., Ulaszek, R., Wei, J., Wolters, D., and Yates, J. R. (2002). Proteomic survey of metabolic pathways in rice. Proc. Natl Acad. Sci. USA 99, 11968–11974. Koornneef, M., Dellaert, L. W., and van der Veen, J. H. (1982). EMS- and radiation-induced mutation frequencies at individual loci in Arabidopsis thaliana (L.) Heynh. Mutat. Res. 93, 109–123. Krysan, P. J., Young, J. C., Tax, F., and Sussman, M. R. (1996). Identiﬁcation of transferred DNA insertions within Arabidopsis genes involved in signal transduction and ion transport. Proc. Natl Acad. Sci. USA 93, 8145–8150. Krysan, P. J., Young, J. C., and Sussman, M. R. (1999). T-DNA as an insertional mutagen in Arabidopsis. Plant Cell 11, 2283–2290. Lee, H., Suh, S. S., Park, E., Cho, E., Ahn, J. H., Kim, S. G., Lee, J. S., Kwon, Y. M., and Lee, I. (2000). The AGAMOUS-LIKE 20 MADS domain protein integrates ﬂoral inductive pathways in Arabidopsis. Genes Dev. 14, 2366–2376. Leung, H., Wu, C., Baraoidan, M. R., Bordeos, A., Ramos, M., Madamba, S., Cabauatan, P., Vera Cruz, C. M., Portugal., A., Reves, G., Bruskiewich, R., McLaren, G., Gregorio, G., Bennett, J., Brar, D., Khush, G., Schnable, P., Wang, G. L., and Leach, J. E. (2001). Deletion mutants for

RICE FUNCTIONAL GENOMICS

107

functional genomics: progress in phenotyping, sequence assignment, and database development. In “Rice Genetics IV” (G. S. Khush, D. S. Brar, and B. Hardy, Eds.), pp. 239– 251. New Delhi Science Publishers, Inc. Li, W., and Gill, B. S. (2002). The colinearity of the Sh2/A1 orthologous region in rice, sorghum and maize is interrupted and accompanied by genome expansion in the Triticeae. Genetics 160, 1153–1162. Li, X., Song, Y., Century, K., Straight, S., Ronald, P., Dong, X., Lassner, M., and Zhang, Y. (2001). A fast neutron deletion mutagenesis-based reverse genetics system for plants. Plant J. 27, 235–242. Linares, O. F. (2002). African rice (Oryza glaberrima): history and future potential. Proc. Natl Acad. Sci. USA 99, 16360–16365. Liu, Y.-G., and Whittier, R. (1995). Thermal asymmetric interlaced PCR: automatable ampliﬁcation and sequencing of insert end fragments from P1 and YAC clones for chromosome walking. Genomics 25, 674–681. Liu, L. X., Spoerke, J. M, Mulligan, E. L., Chen, J., Reardon, B., Westlund, B., Sun, L., Abel, K., Armstrong, B., Hardiman, G., King, J., McCague, L., Basson, M., Clover, R., and Johnson, C. D. (1999a). High-throughput isolation of Caenorhabditis elegans deletion mutants. Genome Res. 9, 859 –867. Liu, D., Zhang, S., Fauquet, C., and Crawford, N. M. (1999b). The Arabidopsis transposon Tag1 is active in rice, undergoing germinal transposition and restricted, late somatic excision. Mol. Gen. Genet. 262, 413–420. Lu, B. R. (1999). Taxonomy of the genus Oryza (Poaceae): historical perspective and current status. Intl. Rice Res. Notes 24, 4–8. Machida, C., Onouchi, H., Koizumi, J., Hamada, S., Semiarti, E., Torikai, S., and Machida, Y. (1997). Characterization of the transposition pattern of the Ac element in Arabidopsis thaliana using endonuclease I-SceI. Proc. Natl Acad. Sci. USA 94, 8675–8680. Mao, L., Wood, T. C., Yu, Y., Budiman, M. A., Tomkins, J., Woo, S., Sasinowski, M., Presting, G., Frisch, D., Goff, S., Dean, R. A., and Wing, R. A. (2002). Rice transposable elements: a survey of 73,000 sequence-tagged-connectors. Genome Res. 10, 982 –990. Martienssen, R. A. (1998). Functional genomics: probing plant gene function and expression with transposons. Proc. Natl Acad. Sci. USA 95, 2021–2026. Matsuo, T. (1997). Origin and distribution of cultivated rice. In “Science of the Rice Plant” (T. Matsuo, Y. Futsuhara, F. Kikuchi, and H. Yamaguchi, Eds.), pp. 69–88. Food and Agriculture Policy Research Center, Tokyo. McCallum, C. M., Comai, L., Greene, E. A., and Henikoff, S. (2000a). Targeted screening for induced mutations. Nat. Biotechnol. 18, 455–457. McCallum, C. M., Comai, L., Greene, E. A., and Henikoff, S. (2000b). Targeting induced local lesions in genomes (TILLING) for plant functional genomics. Plant Physiol. 123, 439–442. Miyao, A., Murata, K., Tanaka, K., Onosato, K., Yamashita, Y., and Hirochika, H. (2001). Proﬁle of retrotransposon Tos17 insertions in rice genome sequence, OsCesA7 disrupted by insertion of the retrotransposon Tos17, In “9th Conference of Plant and Animal Genome P5b-1”. Monna, L. (2003). Single nucleotide polymorphisms (SNPs) in rice, In “Abstracts of Rice Genome Workshop. February 5 –7, 2003.” Rice Genome Research Program, National Institute of Agrobiological Resources/STAFF Institute, Tsukuba, Japan, p. 15. Mori, M., Sugimoto, K., Satoh, K., Okabe, K., Toki, S., Ugaki, M., Hirochika, H., and Kikuchi, S. (2000). Toward activation tagging in rice “6th International Congress of Plant Molecular Biology”, p. S1-57. Morishima, H. (2001). Evolution and domestication of rice. In “Rice Genetics IV” (G. S. Khush, D. S. Brar, and B. Hardy, Eds.), pp. 63– 77. New Delhi Science Publishers Inc., and International Rice Research Institute, Los Ban˜os, Philippines. Murai, N., Li, Z., Kawagoe, Y., and Hayashimamoto, A. (1991). Transposition of the maize activator element in transgenic rice plants. Nucleic Acids Res. 19, 617–622.

108

H. LEUNG AND G. AN

Nasu, S., Suzuki, J., Ohta, R., Hasegawa, K., Yui, R., Kitazawa, N., Monna, L., and Minobe, Y. (2002). Search for and analysis of single nucleotide polymorphisms (SNPs) in rice (Oryza sativa, Oryza ruﬁpogon) and establishment of SNP markers. DNA Research 9, 163–171. Nakagawa, Y., Machida, C., Machida, Y., and Toriyama, K. (2000). Frequency and pattern of transposition of the maize transposable element Ds in transgenic rice plants. Plant Cell Physiol. 41, 733 –742. Normile, D., and Pennisi, E. (2002). Rice: boiled down to bare essentials. Science 296, 32 –36. Nuwaysir, E. F., Huang, W., Albert, T. J., Singh, J., Nuwaysir, K., Pitas, A., Richmond, T., Gorski, T., Berg, J., Ballin, J., McCormick, M., Norton, J., Pollock, T., Sumwalt, T., Butcher, L., Porter, D. A., Molla, M., Hall, C., Blattner, F., Sussman, M., Wallace, R., Cerrina, F., and Green, R. (2002). Gene expression analysis using oligonucleotide arrays produced by maskless photolithography. Genome Res. 12, 1749–1755. Okubara, P. A., Anderson, P. A., Ochoa, O. E., and Michelmore, R. W. (1994). Mutants of downy mildew resistance in Lactuca sativa (lettuce). Genetics 137, 867–874. O’Toole, J. C., Toenniessen, G. H., Murashige, T., Harris, R. R., and Herdt, W. W. (2001). The Rockefeller Foundation’s International Program on Rice Biotechnology. In “Rice Genetics IV” (G. S. Khush, D. S. Brar, and B. Hardy, Eds.), pp. 39–59. New Delhi Science Publishers Inc., and International Rice Research Institute, Los Ban˜os, Philippines. Page, D. R., and Grossniklaus, U. (2002). The art and design of genetic screen: Arabidopsis thaliana. Nat. Rev. Genet. 3, 124 –135. Parinov, S., Sevugan, M., De, Y., Yang, W. C., Kumaran, M., and Sundaresan, V. (1999). Analysis of ﬂanking sequences from dissociation insertion lines: A database for reverse genetics in Arabidopsis. Plant Cell 11, 2263–2270. Parrish, S., Fleenor, J., Xu, S., Mello, C., and Fire, A. (2000). Functional anatomy of a dsRNA trigger: differential requirement for the two trigger strands in RNA interference. Mol. Cell 6, 1077– 1087. Peterfy, M., Gyuris, T., and Takacs, L. (2000). Signal-exon trap: a novel method for the identiﬁcation of signal sequences from genomic DNA. Nucleic Acids Res. 28, E26. Phogat, S., Burma, P. K., and Pental, D. (2000). A four-element based transposon system for allele speciﬁc tagging in plants: theoretical considerations. J. Biosci. 25, 57– 63. Plesch, G., Kamann, E., and Mueller-Roeber, B. (2000). Cloning of regulatory sequences mediating guard-cell-speciﬁc gene expression. Gene 249, 83 –89. Pritchard, J. K., and Rosenberg, NA. (1999). Use of unlinked genetic markers to detect population stratiﬁcation in association studies. Am. J. Hum. Genet. 65, 220 –228. Pritchard, J. K., Stephens, M., Rosenberg, N. A., and Donnelly, P. (2000). Association mapping in structured populations. Am. J. Hum. Genet. 67, 170 –181. Qu, S., Jeon, J.-S., Richter, T., Leach, J. E., Ronald, P., and Wang, G.-L. (2003). Development of chemically inducible Ac/Ds elements for gene mutagenesis in rice. In “Abstracts of Plant and Animal Genome XI. January 11–15, 2003, San Diego, California, USA”. p. 151, p. 515, Abstract P306. Rafalski, A. (2002). Applications of single nucleotide polymorphisms in crop genetics. Curr. Opin. Plant Biol. 5, 94–100. Ramalingam, J., Vera Cruz, C. M., Kukreja, K., Chittoor, J. M., Wu, J. L., Lee, S. W., Baraoidan, M. R., George, M. L., Cohen, M., Hulbert, S, Leach, J. E., and Leung, H. (2003). Candidate resistance genes from rice, barley, and maize and their association with qualitative and quantitative resistance in rice. Mol. Plant-Microbe Interact. 16, 14– 24. Reardon, J. T., Liljestrand-Golden, C. A., Dusenbery, R. L., and Smith, P. D. (1987). Molecular analysis of diepoxybutane-induced mutations at the rosy locus of Drosophila melanogaster. Genetics 115, 323–331. Remington, D. L., Thornsberry, J. M., Matsuoka, Y., Wilson, L. M., Whitt, S. R., Doebley, J., Kresovich, S, Goodman, M. M., and Buckler, E. S. (2001). Structure of linkage disequilibrium and phenotypic associations in the maize genome. Proc. Natl Acad. Sci. USA 98, 11479–11484.

RICE FUNCTIONAL GENOMICS

109

Rigoutsos, I., and Floratos, A. (1998). Combinatorial pattern discovery in biological sequences: the TEIRESIAS algorithym. Bioinformatics 14, 55– 67. Robert, M. A., Reader, S. M., Dalgliesh, C., Miller, T. E., Foote, T. N., Fish, L. J., Snape, J. W., and Moore, G. (1999). Induction and characterization of Ph1 wheat mutants. Genetics 153, 1909–1918. Salekdeh, Gh. H., Siopongco, J., Wade, L. J., Ghareyazie, B., and Bennett, J. (2002a). A proteomic approach to analyzing drought- and salt-responsiveness in rice. Field Crop Res. 76, 199–219. Salekdeh, Gh. H., Siopongco, J., Wade, L. J., Ghareyazie, B., and Bennett, J. (2002b). Proteomics analysis of rice leaves during drought stress and recovery. Proteomics 2, 1131–1145. Sasaki, A., Ashikari, M., Ueguchi-tanaka, M., Itoh, H., Nishimura, A., Swapan, D., Ishiyama, K., Saito, T., Kobayashi, M., Khush, G. S., Kitano, H., and Matsuoka, M. (2002). Green revolution: a mutant gibberellin-synthesis gene in rice. Nature 416, 701 –702. Sasaki, T., Matsumoto, T., Yamamoto, K., Sakata, K., Baba, T., Katayose, Y., Wu, J., Niimura, Y., Cheng, Z., Nagamura, Y., Antonio, B. A., Kanamori, H., Hosokawa, S., Masukawa, M., Arikawa, K., Chiden, Y., Hayashi, M., Okamoto, M., Ando, T., Aoki, H., Arita, K., Hamada, M., Harada, C., Hijishita, S., Honda, M., Ichikawa, Y., Idonuma, A., Iijima, M., Ikeda, M., Ikeno, M., Ito, S., Ito, T., Ito, Y., Ito, Y., Iwabuchi, A., Kamiya, K., Karasawa, W., Katagiri, S., Kikuta, A., Kobayashi, N., Kono, I., Machita, K., Maehara, T., Mizuno, H., Mizubayashi, T., Mukai, Y., Nagasaki, H., Nakashima, M., Nakama, Y., Nakamichi, Y., Nakamura, M., Namiki, N., Negishi, M., Ohta, I., Ono, N., Saji, S., Sakai, K., Shibata, M., Shimokawa, T., Shomura, A., Song, J., Takazaki, Y., Terasawa, K., Tsuji, K., Waki, K., Yamagata, H., Yamane, H., Yoshiki, S., Yoshihara, R., Yukawa, K., Zhong, H., Iwama, H., Endo, T., Ito, H., Hahn, J. H., Kim, H-I., Eun, M-Y., Yano, M., Jiang, J., and Gojobori, T. (2002). The genome sequence and structure of rice chromosome 1. Nature 420, 312– 316. Sato, Y., Sentoku, N., Miura, Y., Hirochika, H., Kitano, H., and Matsuoka, M. (1999). Loss-offunction mutations in the rice homeobox gene OSH15 affect the architecture of internodes resulting in dwarf plants. EMBO J. 18, 992– 1002. Schena, M., and Davis, R. W. (1999). Parallel analysis with biological chips. In “PCR Applications: Protocols for Functional Genomics” (M. A. Innis, D. H. Gelfand, and J. J. Sninsky, Eds.), Academic Press. Schmitz, G., and Theres, K. (1994). A self-stabilizing Ac derivative and its potential for transposon tagging. Plant J. 6, 781 –786. Scholthof, H. B., Scholthof, K.-B. G., and Jackson, A. O. (1996). Plant virus gene vectors for transient expression of foreign proteins in plants. Annu. Rev. Phytopathol. 34, 299– 323. Shimamoto, K., Miyazaki, C., Hashimoto, H., Izawa, T., Itoh, K., Terada, R., Inagaki, Y., and Iida, S. (1993). Trans-activation and stable integration of the maize transposable element Ds cotransfected with the Ac transposase gene in transgenic rice plants. Mol. Gen. Genet. 239, 354 –360. Shimamoto, K., and Kyozuka, J. (2002). Rice as a model for comparative genomics in plants. Annu. Rev. Plant Biol. 53, 399 –419. Shimamoto, K., Moritoh, S., and Miki Daisuke (2002). RNAi as a complementary method for functional genomics of rice. In “Abstracts of International workshop on towards building a global rice gene machine, November 11–12, 2002”, p. 27, CSIRO Plant Industry, Canberra, Australia. Shirley, B. S., Hanley, S., and Goodman, H. M. (1992). Effects of ionizing radiation on a plant genome: analysis of two Arabidopsis transparent testa mutations. Plant Cell 4, 333 –347. Song, W. H., Wang, G. L., Chen., L. L., Kim, H. S., Pi, L. Y., Holsten, T., Gardner, J., Wang, B., Zhai, W. X., Zhu, L. H., Fauquet, C., and Ronald, P. (1995). A receptor kinase-like protein encoded by the rice disease resistance gene, Xa21. Science 270, 1804–1806. Spielmeyer, W., Ellis, M. H., and Chandler, P. M. (2002). Semidwarf (sd-1), “green revolution” rice, contains a defective gibberellin 20-oxidase gene. Proc. Natl Acad. Sci., USA 99, 9043–9048.

110

H. LEUNG AND G. AN

Springer, P. S. (2000). Gene traps: tools for plant development and genomics. Plant Cell 12, 1007–1020. Stockinger, E. J., Gilmour, S. J., and Thomashaw, M. F. (1997). Arabidopsis thaliana CBF1 encodes an AP2 domain-containing transcriptional activator that binds the C-repeat/DRE, a cis-acting DNA regulatory element that stimulates transcription in response to low temperature and water deﬁcit. Proc. Natl Acad. Sci. USA 94, 1035–1040. Sugimoto, K., Otsuki, Y., Saji, S., and Hirochika, H. (1994). Transposition of the maize Ds element from a viral vector to the rice genome. Plant J. 5, 863– 871. Sundaresan, V. (1996). Horizontal spread of transposon mutagenesis: new uses of old elements. Trends Plant Sci. 1, 184–191. Takano, M., Kanegae, H., Shinomura, T., Miyao, A., Hirochika, H., and Furuya, M. (2001). Isolation and characterization of rice phytochrome A mutants. Plant Cell 13, 521–534. Takeda, S., Sugimoto, K., Otsuki, H., and Hirochika, H. (1998). Transcriptional activation of the tobacco retrotransposon Tto1 by wounding and methyl jasmonate. Plant Mol. Biol. 36, 365–376. Tanksley, S. D., and McCouch, S. R. (1997). Seed banks and molecular maps: unlocking genetic potential from the wild. Science 277, 1063–1066. Terada, R., Urawa, H., Inagaki, Y., Tsugane, K., and Iida, S. (2002). Efﬁcient gene targeting by homologous recombination in rice. Nat. Biotechnol. 20, 1030– 1034. Thornsberry, J. M., Goodman, M. M., Doebley, J., Kresovich, S., Nielsen, D., and Buckler, E. S. IV (2001). Dwarf8 polymorphisms associate with variation in ﬂowering time. Nat. Genet. 28, 286–289. The Arabidopsis Genome Initiative, (2000). Analysis of the genome sequence of the ﬂowering plant Arabidopsis thaliana. Nature 408, 796– 815. Thomashaw, M. F. (1999). Plant cold acclimation: freezing tolerance genes and regulatory mechanisms. Annu. Rev. Plant Physiol. Plant Mol. Biol. 50, 571–599. Topping, J. F., and Lindsey, K. (1997). Promoter trap markers differentiate structural and positional components of polar development in Arabidopsis. Plant Cell 9, 1713– 1725. Triglia, T., Peterson, M. G., and Kemp, D. J. (1988). A procedure for in vitro ampliﬁcation of DNA segments that lie outside the boundaries of known sequences. Nucleic Acids Res. 16, 81–86. Vigouroux, Y., McMullen, M., Hittinger, C. T., Houchins, K., Schulz, L., Kresovich, S., Matsuoka, Y., and Doebley, J. (2002). Identifying genes of agronomic importance in maize by screening microsatellites for evidence of selection during domestication. Proc. Natl Acad. Sci. USA 99, 9650–9655. Virmani, S. S., and Ilyas-Ahmed, M. (2001). Environment-sensitive genic male sterility (EGMS) in crops. Adv. Agron. 72, 139– 195. Walbot, V. (2000). Saturation mutagenesis using maize transposons. Curr. Opin. Plant Biol. 3, 103–107. Watanabe, Y. (1997). Phylogeny and geographical distribution of genus Oryza. In “Science of the Rice Plant” (T. Matsuo, Y. Futsuhara, F. Kikuchi, and H. Yamaguchi, Eds.), pp. 29–39. Food and Agriculture Policy Research Center, Tokyo. Waterhouse, P. M., Wang, M. B., and Lough, T. (2001). Gene silencing as an adaptive defence against viruses. Nature 411, 834–842. Weigel, D., Ahn, J. H., Blazquez, M. A., Borevitz, J. O., Christensen, S. K., Fankhauser, C., Ferrandiz, C., Kardailsky, I., Malancharuvil, E. J., Neff, M. M., Nguyen, J. T., Sato, S., Wang, Z. Y., Xia, Y., Dixon, R. A., Harrison, M. J., Lamb, C. J., Yanofsky, M. F., and Chory, J. (2000). Activation tagging in Arabidopsis. Plant Physiol. 122, 1003–1013. Wesley, S. V., Helliwell, C. A., Smith, N. A., Wang, M. B., Rouse, D. T., Liu, Q., Gooding, P. S., Singh, S. P., Abbott, D., Stoutjesdijk, P. A., Robinson, S. P., Gleave, A. P., Green, A. G., and Waterhouse, P. M. (2001). Construct design for efﬁcient, effective and high-throughput gene silencing in plants. Plant J. 27, 581–590.

RICE FUNCTIONAL GENOMICS

111

Westlund, B., Parry, D., Clover, R., Basson, M., and Johnson, C. D. (1999). Reverse genetic analysis of Caenorhabditis elegans presenilins reveals redundant but unequal roles for se-12 and hop-1 in Notch-pathway signaling. Proc. Natl Acad. Sci. USA 86, 2497–2502. White, P. T. (1994). Rice: the essential harvest. National Geographic, 48–79 (May issue). Whitt, S. R., Wilson, L. M., Tenaillon, M. I., Gaut, B. S., and Buckler, E. S. (2002). Genetic diversity and selection in maize starch pathway. Proc. Natl Acad. Sci. USA 99, 12959–12962. Williams, N. (1996). Genome projects: yeast genome sequence ferments new research. Science 272, 481. Wisman, E., Cardon, G. H., Fransz, P., and Saedler, H. (1998). The behaviour of the autonomous maize transposable element En/Spm in Arabidopsis thaliana allows efﬁcient mutagenesis. Plant Mol. Biol. 37, 989–999. Wolfe, K. H., Gouy, M., Yang, Y. W., Sharp, P. M., and Li, W. H. (1989). Date of the monocot-dicot divergence estimated from chloroplast DNA sequence data. Proc. Natl Acad. Sci. USA 86, 6201–6205. Wong, G. K.-S., Wang, J., Tao, L., Tan, J., Zhang, J., Passey, D. A., and Yu, J. (2002). Compositional gradients in Gramineae genes. Genome Res. 12, 851–856. Yandell, M. D., Edgar, L. G., and Wood, W. B. (1994). Trimethylpsoralen induces small deletion mutations in Caenorhabditis elegans. Proc. Natl Acad. Sci. USA 91, 1381–1385. Yano, M., Katayose, Y., Ashikari, M., Yamanouchi, U., Monna, L., Fuse, T., Baba, T., Yamamoto, K., Umehara, Y., Nagamura, Y., and Sasaki, T. (2000). Hd1, a major photoperiod sensitivity quantitative trait locus in rice, is closely related to the Arabidopsis ﬂowering time gene CONSTANS. Plant Cell 12, 2473– 2483. Yin, Z., and Wang, G.-L. (2000). Evidence of multiple complex patterns of T-DNA integration into rice genome. Theor. Appl. Genet. 100, 461–470. Yu, J., Hu, S., Wang, J., Wong, G. K., Li, S., Liu, B., Deng, Y., Dai, L., Zhou, Y., Zhang, X., Cao, M., Liu, J., Sun, J., Tang, J., Chen, Y., Huang, X., Lin, W., Ye, C., Tong, W., Cong, L., Geng, J., Han, Y., Li, L., Li, W., Hu, G., Huang, X., Li, W., Li, J., Liu, Z., Li, L., Liu, J., Qi, Q., Liu, J., Li, L., Li, T., Wang, X., Lu, H., Wu, T., Zhu, M., Ni, P., Han, H., Dong, W., Ren, X., Feng, X., Cui, P., Li, X., Wang, H., Xu, X., Zhai, W., Xu, Z., Zhang, J., He, S., Zhang, J., Xu, J., Zhang, K., Zheng, X., Dong, J., Zeng, W., Tao, L., Ye, J., Tan, J., Ren, X., Chen, X., He, J., Liu, D., Tian, W., Tian, C., Xia, H., Bao, Q., Li, G., Gao, H., Cao, T., Wang, J., Zhao, W., Li, P., Chen, W., Wang, X., Zhang, Y., Hu, J., Wang, J., Liu, S., Yang, J., Zhang, G., Xiong, Y., Li, Z., Mao, L., Zhou, C., Zhu, Z., Chen, R., Hao, B., Zheng, W., Chen, S., Guo, W., Li, G., Liu, S., Tao, M., Wang, J., Zhu, L., Yuan, L., and Yang, H. (2002). A draft sequence of the rice genome (Oryza sativa L. ssp. indica). Science 296, 79–92. Zhu, T., Budworth, P., Chen, W., Provart, N., Chang, H.-S., Guimil, S., Su, W., Estes, B., Zou, G., and Wang, X. (2003). Transcriptional control of nutrient partitioning during rice grain ﬁlling. Plant Biotechnol. 1, 59 –70. Zivy, M., and de Vienne, D. (2000). Proteomics: a link between genomics, genetics and physiology. Plant Mol. Biol. 44, 575–580.

This page is intentionally left blank

THE NATURE, PROPERTIES AND MANAGEMENT OF VOLCANIC SOILS R.A. Dahlgren,1 M. Saigusa2 and F.C. Ugolini3 1

Land, Air and Water Resources, University of California, Davis, California 95616, USA 2 Experimental Farm of Tohoku University, Kawatabi, Naruko, Tamatsukuri, Miyagi, 989-6711 Japan 3 Dipartimento di Scienza del Suolo e Nutrizione della Pianta, Universita` Degli Studi, Piazzale delle Cascine 15, 50144, Firenze, Italy

I. Introduction II. Volcanic Soil Distribution III. Soil Genesis in Volcanic Materials A. Allophanic and Nonallophanic Andisols B. General Trends in Soil Development on Volcanic Materials IV. Soil Classiﬁcation A. World Reference Base for Soil Resources—Andosols B. Soil Taxonomy—Andisols V. Chemical Weathering VI. Mineralogical Characteristics A. Common Colloidal Constituents of Volcanic Soils B. Formation and Transformation of Colloids in Volcanic Soils VII. Selected Chemical Characteristics of Volcanic Soils A. Organic Matter Accumulation B. Aluminum Dynamics VIII. Productivity and Management of Volcanic Soils A. Charge Characteristics and Chemical Fertility B. Physical Properties and Fertility C. Soil Management and Conservation of Volcanic Soils References

Soils formed in volcanic ejecta have many distinctive physical, chemical, and mineralogical properties that are rarely found in soils derived from other parent materials. These distinctive properties are largely attributable to the formation of noncrystalline materials (e.g., allophane, imogolite, ferrihydrite) containing variable charge surfaces, and the accumulation of organic matter. Formation of noncrystalline materials is directly related to the properties of volcanic ejecta as a parent material, namely the rapid weathering of glassy particles. The composition of the colloidal fraction forms a continuum between pure Al – humus complexes and pure allophane/imogolite, depending on the pH and organic matter characteristics 113 Advances in Agronomy, Volume 82 Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved. 0065-2113/03 $35.00

114

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI of the weathering environment. For soil management purposes, volcanic soils are often divided into two groups based on the colloidal composition of the surface horizons: allophanic soils dominated by allophane and imogolite, and nonallophanic soils dominated by Al –humus complexes and 2:1 layer silicates. Volcanic soils exhibit a wide range of agricultural productivity, depending on the degree or intensity of pedogenic development and the colloidal composition of the rooting zone. The different charge characteristics of allophanic and nonallophanic soils is the most important factor regulating chemical fertility attributes. Phosphorus ﬁxation, strong acidity, and aluminum toxicity are the primary chemical limitations to agricultural productivity. Volcanic soils generally have high physical fertility (tilth) and mature soils are relatively resilient to erosion and compaction. To maximize the productivity of volcanic soils, proper management based on an understanding of the unique physical, chemical, and mineralogical properties of these soils must be practiced. q 2004 Academic Press.

I. INTRODUCTION Volcanoes are revered and feared for their awesome and devastating eruptions that obliterate terrestrial ecosystems and often cause tremendous casualties to humans and wildlife (Fig. 1). There have been about 1500 active volcanoes during the last 10,000 years with approximately 60 eruptions each year (Simkin and Siebert, 1994; Global Volcanism Program, 2003). In terms of human casualties, the most catastrophic destruction is caused by pyroclastic ﬂows (e.g., 1902 Mont Pele´e, Martinique 29,025 casualties) and mudﬂows (e.g., 1985 Nevado del Ruiz, Colombia 25,000 casualties), but tsunamis (e.g., 1883 Krakatau, Indonesia 36,417 casualties), starvation (e.g., 1815 Tambora, Indonesia 92,000 casualties), and carbon dioxide degassing of volcanic lakes (e.g., 1986 Lake Nyos, Cameroon 1700 casualties) have also resulted in the loss of many human lives (Blong, 1984). Yet from these ashes of devastation arise some of the most productive soils in the world with the capacity to sustain high human population densities. Volcanism plays an important beneﬁcial role in sustaining the productivity of terrestrial ecosystems through soil rejuvenation. Volcanism is probably the most obvious mechanism of recycling large amounts of geological material and gases (e.g., CO2, SO2) on the planet Earth. Further, deposition of volcanic ejecta is not limited to the immediate vicinity of volcanoes; volcanic ash may be transported to great distances once injected into the atmospheric circulation. Because volcanism is responsible for bringing new material to the earth’s surface, volcanism is counteracting the effects of physical and chemical erosion. By providing a source of easily weatherable materials, this natural phenomenon represents an effective sink for carbon dioxide through

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 115

Figure 1 The 1980 eruptions of Mt. St. Helens, Washington, created a disturbance gradient that ranged from complete obliteration of the landscape near the mountain to a mere dusting of volcanic ash hundreds of kilometers from the vent. These eruptions rejuvenate soil-forming processes and renew the long-term fertility status of terrestrial ecosystems.

carbonic acid weathering ðCO2 ðgÞ þ H2 O ¼ Hþ þ HCO2 3 Þ. Soils formed in volcanic deposits also contain the largest accumulations of organic carbon among the mineral soil orders (Eswaran et al., 1993). Thus, volcanism plays an important role in the global carbon cycle, representing a primary source and sink for carbon. Soils formed in volcanic deposits have many distinctive properties that are rarely found in soils derived from other parent materials. Some unique properties of volcanic soils include variable charge, high water retention, high phosphate retention, low bulk density, high friability, highly stable soil aggregates, and excellent tilth (Shoji et al., 1993a). These distinctive properties are largely due to the formation of noncrystalline materials (e.g., active Al and Fe – allophane, imogolite, ferrihydrite, Al/Fe –humus complexes) and the accumulation of organic carbon, the dominant pedogenic processes occurring in volcanic soils. Formation of noncrystalline materials is directly related to the properties of volcanic ejecta as a parent material, namely the rapid weathering of the glassy particles. Volcanic soils exhibit a wide range of agricultural productivity, depending primarily on the degree or intensity of pedogenic development. The high potential of volcanic soils for agricultural production is illustrated by the fact that many of the most productive agricultural regions of the world are located near active or

116

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

dormant volcanoes (Shoji et al., 1993a). In volcanic regions such as Indonesia, the most densely populated areas are found near volcanoes. The intermittent additions of volcanic ash renew the long-term fertility status of terrestrial ecosystems by providing a source of nutrients from the rapid weathering of volcanic deposits. However, as soils become progressively more weathered, negative attributes such as P ﬁxation, low exchangeable base concentrations, and Al toxicity may severely limit agricultural production. To maximize the productivity of volcanic soils, proper management based on an understanding of the unique physical, chemical, and mineralogical properties of these soils must be practiced. The purpose of this chapter is to present a brief overview of soils derived from volcanic ejecta in terms of their distribution, genesis, classiﬁcation, and their mineralogical, chemical, physical, and agronomic characteristics. In this context, volcanic ejecta include ash, tephra, cinders, lahars, lapilli, tuff, pumice, volcanic bombs, and lava. These materials may also be reworked and mixed with materials from other geologic sources (e.g., loess, alluvium, and glacial drift). Research on volcanic soils would ﬁll volumes and several compilations on volcanic soils have been previously published (e.g.,Wada and Harward, 1974; Ugolini and Zasoski, 1979; Theng, 1980; Wada, 1980, 1985, 1986; Yoshinaga, 1983; Mizota and van Reeuwijk, 1989; Shoji et al. 1993a; Kimble et al., 2000; Ugolini and Dahlgren, 2002). We focus this review on recent developments and concepts affecting their properties and management as a soil resource.

II. VOLCANIC SOIL DISTRIBUTION The distribution of soils derived from volcanic materials closely parallels the global distribution of active and recently active volcanoes (Fig. 2). Volcanism is associated with interactions along tectonic plate boundaries (compressional and extensional margins) and with “hot spots” that occur within tectonic plates. The most extensive chain of volcanoes is associated with subduction of the Paciﬁc oceanic plate with continental masses along its boundary. This circum-Paciﬁc belt, known as the “Ring of Fire”, contains about 75% of the world’s approximately 520 currently active volcanoes (Harris, 1988). Other prominent chains of volcanoes associated with plate boundaries include subduction zones in the Mediterranean Sea, Indonesia (Java Trench), and the Caribbean, rift zones in northeast Africa (East Africa rift system), and spreading centers in Iceland and the Azores (Mid-Atlantic Ridge). Hot spot volcanoes occur where fractures in the continental plate or hot plumes of magma rise close to the earth’s surface resulting in magma ejection at the earth’s surface. About 5% of the world’s active volcanoes are associated with hot spots, the most famous including the Hawaiian Islands –Emperor Seamounts chain and Galapagos Islands on the Paciﬁc Plate and Yellowstone Park on the North American Plate.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 117

Figure 2 Global distribution of Andisols (Courtesy of USDA-NRCS, Soil Survey Division, World Soil Resources, 1998). Andisols cover about 120 million hectares or nearly 1% of the world’s land surface.

118

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

While it is difﬁcult to establish the area inﬂuenced by volcanic ejecta, soils with andic soil properties that are characterized as Andosols or Andisols cover from 110 to 124 million hectares, about 0.84% of the world’s land surface (Leamy, 1984; FAO, ISRIC and ISSS, 1998; Kimble et al., 2000). In addition, large land areas receive periodic additions of volcanic materials that inﬂuence soil properties, but are not sufﬁcient to meet the minimum requirements for Andosols or Andisols. For example, the cataclysmic eruption of Long Valley volcano in eastern California about 730,000 years ago ejected 600 km3 of pyroclastic material. These materials inundated 1500 km2 of central California and southwestern Nevada with several meters of pyroclastic materials, and deposited recognizable ash layers as far away as central Nebraska, a distance of over 1700 km (Harris, 1988). Volcanic soils occur across a wide range of temperature regimes: tropical (49%) . boreal (28%) . temperate (23%); and soil moisture regimes: udic (63%) . ustic (30%) . aridic (4%) . xeric (3%) (udic moisture regime includes perudic and aquic conditions — Wilding, 2000).

III. SOIL GENESIS IN VOLCANIC MATERIALS Formation of noncrystalline materials (active Al and Fe compounds) and accumulation of organic matter are the dominant pedogenic processes occurring in most soils formed in volcanic materials (Shoji et al., 1993a). This combination of processes, occurring preferentially in soils formed in volcanic materials, is termed “andosolization” (Duchaufour, 1977; Ugolini et al., 1988). Succinctly, andosolization is the darkening of the soil due to an accumulation of stable humic substances for considerable depth occurring in subacid conditions. Andosolization is a case of melanization in which Al3þ is the dominant cation; the other case of melanization occurs in the presence of Ca2þ and produces mollic epipedons. This relatively thick A horizon, that may or may not contain poorly crystalline secondary aluminosilicate material, is underlain by a Bw horizon. The colloidal composition of B horizons is dominated by allophane/imogolite and other noncrystalline materials including ferrihydrite (Childs et al., 1991; Shoji et al., 1993a). These soils lack distinct eluvial horizons, and illuvial horizons containing enrichment of clay or humus. Soil solution studies in volcanic deposits documented that andosolization is characterized by accumulation of Fe, Al, and dissolved organic carbon in A horizons with little translocation of these components to B horizons (Ugolini et al., 1988; Dahlgren et al., 1991). Thus, noncrystalline materials (e.g., allophane, imogolite, and ferrihydrite) in B horizons have formed in situ, rather than by translocation. Formation of noncrystalline materials is promoted by the carbonic acid weathering regime and hindered by the organic acid weathering regime. In areas near intermittently active

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 119

Figure 3 A polygenetic soil proﬁle in the vicinity of Mt. St. Helens consisting of four distinct ash deposits. Burial of A horizons contributes to organic matter preservation in volcanic soils experiencing intermittent ash deposition.

volcanoes, it is common to ﬁnd polygenetic proﬁles resulting from soil burial by successive depositional layers (Fig. 3).

A. ALLOPHANIC

AND

NONALLOPHANIC ANDISOLS 1

Under humid weathering conditions, the composition of the colloidal fraction forms a continuum between pure Al –humus complexes and pure allophane/ imogolite, depending on the pH and organic matter characteristics of 1 The classical expression of soils developed in volcanic materials is termed Andisols in Soil Taxonomy and Andosols in the WRB classiﬁcation system. For convenience, we will use the term Andisols to generically describe volcanic soils meeting Andisol and/or Andosol criteria in the text. Andosols will only be used when speciﬁcally referring to volcanic soils meeting WRB criteria.

120

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

the weathering environment (Mizota and van Reeuwijk, 1989). For soil management purposes, Andisols are often divided into two groups based on the mineralogical composition of the surﬁcial horizons (A horizons): allophanic Andisols dominated by allophane and imogolite, and nonallophanic Andisols dominated by Al –humus complexes and 2:1 layer silicates (Shoji and Ono, 1978; Shoji, 1985; Shoji et al., 1985). In Japan, allophanic Andisols are mainly distributed in areas having thick deposition of Holocene and/or late-Pleistocene tephras, while nonallophanic Andisols are found in areas with older tephra deposits. Older tephra deposits are generally more acidiﬁed and have been exposed to deposition of exogenous materials, such as loess from China dominated by 2:1 clay minerals (Inoue and Naruse, 1987; Mizota and Inoue, 1988; Mizota et al., 1990; Bautista-Tulin and Inoue, 1997; Saigusa and Matsuyama, 1998). Nonallophanic Andisols represent about 30% of all Andisols in Japan and are distributed throughout the world (e.g., Japan, USA, Indonesia, Spain, Italy, Portugal, Chile, New Zealand, West Samoa, Taiwan) (Shoji et al., 1985, 1987; Leamy et al., 1988; Madeira et al., 1994; Johnson-Maynard et al., 1997; Chen et al., 1999). The availability of Al3þ appears to be the critical factor regulating the formation of nonallophanic versus allophanic Andisols. Nonallophanic Andisols form preferentially in pedogenic environments that are rich in organic matter and have pH values of 5 or less (Fig. 4; Shoji and Fujiwara, 1984). Base-poor volcanic deposits (e.g., rhyolitic, dacitic, or andesitic) having noncolored volcanic glass and precipitation greater than about 1000 mm are two factors that contribute to soil acidiﬁcation. At pH values less than 5, organic acids are the dominant proton donor lowering pH and aqueous Al3þ activities through formation of Al– humus complexes. Aqueous Al3þ is also incorporated into the interlayer of 2:1 layer silicates when they are present (Dahlgren and Ugolini, 1989b). Under these conditions, humus and 2:1 layer silicates effectively compete for dissolved Al, leaving little Al available for co-precipitation with silica to form aluminosilicate materials, such as allophane/imogolite. Preferential incorporation of Al into Al-humus complexes and hydroxy-Al interlayers of 2:1 layer silicates has been termed the antiallophanic effect (Shoji et al., 1993a). Allophanic Andisols dominated by allophane/imogolite form preferentially in weathering environments with pH values in the range of 5 –7 and a low content of complexing organic compounds (Fig. 4; Ugolini and Dahlgren, 1991). As revealed by Shoji et al. (1982) and Shoji and Fujiwara (1984) allophanic Andisols are favored in basic parent materials (e.g., andesitic basalt, basalt) having colored volcanic glass, and climates having less than 1000 mm of precipitation. These conditions favor higher pH values (pH . 5), which promotes formation of Al-polymers relative to Al –humus complexes (Jackson, 1963a, b). The Al-polymers are able to react with silica and form allophane/imogolite. While Al– humus complexes also form in A horizons of allophanic Andisols,

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 121

Figure 4 Examples of Al–humus complex and allophane/imogolite distribution in allophanic and nonallophanic Andisols. Allophane/imogolite is most prominent at soil pH values greater than 5. Al –humus complexes and allophane/imogolite show an inverse relationship in nonallophanic Andisols. [Nonallophanic Andisol ¼ Mukaiyama (Shoji et al., 1993a); Allophanic Andisol ¼ Sirrah (Takahashi et al., 1993)].

122

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

there is an excess of Al relative to the complexing capacity of the humic substances. Parﬁtt and Saigusa (1985) found that when (Alp þ Fep)/Cp (p ¼ pyrophosphate extractable) was greater than about 0.1, allophane/imogolite was present. This implies that once the complexation capacity of humic substances is saturated, the excess Al can react with Si to form allophanic materials.

B. GENERAL TRENDS

SOIL DEVELOPMENT MATERIALS

IN

ON

VOLCANIC

Soils displaying a wide range of characteristics form in volcanic deposits (Table I). Climate, in conjunction with its inﬂuence on vegetation, and time of exposure to weathering are the two primary factors regulating soil developmental pathways in volcanic materials. Andisols generally form rapidly in humid climates and alter to other soil orders as soil age and degree of weathering increase.

Table I Examples of Soil Genesis Pathways in Contrasting Climates Extrapolated from the Literature Climatic regime Warm/Arid and Semi-arid Arid (400 mm): Entisols Semi-arid (700 mm): Ustivitrands Cold/Dry Entisols ! Vitricyrands Cold/Humid Entisols ! Spodosols Entisols ! Andisols ! Spodosols Humid/Temperate Warm/dry: Entisols ! Mollisols Warm/moist: Entisols ! Andisols Warm/moist: Entisols ! Andisols ! Inceptisols ! Alﬁsols/Ultisols Cool/moist: Entisols ! Andisols ! Spodosols Cold/moist: Entisols ! Spodosols Tropical Warm dry/moist: Entisols ! Mollisols Warm dry: Entisols ! Vertisols Warm/moist: Entisols ! Andisols ! Inceptisols ! Alﬁsols/Ultisols Warm/moist: Entisols ! Andisols ! Inceptisols ! Oxisols

Literature source

Dubroeucq et al., 1998 Dubroeucq et al., 1998 Arnalds and Kimble, 2001 Ping et al., 1988; 1989; Shoji et al., 1988a Ping et al., 1988; 1989; Shoji et al., 1988a Otsuka et al., 1988; Shoji et al., 1990a Numerous reports Southard and Southard, 1987; Takahashi et al., 1993; Chen et al., 2001 Shoji et al., 1988b; Takahashi et al. 1989 Shoji et al., 1988b; Ugolini et al., 1977; Parﬁtt and Saigusa, 1985 Wielemaker and Wakatsuki, 1984; Yerima et al., 1987; Chadwick et al., 1994 Yerima et al., 1987 Martini, 1976; Delvaux et al., 1989 Chadwick et al., 1994; Martini, 1976; Quantin, 1992; Van Ranst et al., 2002; Beinroth, 1982; Kimble and Eswaran, 1988

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 123

However, in regions with intermittent deposition of volcanic ash each addition of new material rejuvenates soil developmental processes so that Andisols may be maintained as a relatively stable soil condition. Under humid weathering conditions, criteria for andic soil properties (e.g., active Al and Fe) necessary for classiﬁcation as Andisols may be achieved in 200– 300 years (Dahlgren et al., 1997). In dry climates, soil development is stalled in the Entisols/Vitrands stage as weathering is limited by moisture (Dubroeucq et al., 1998; Arnalds and Kimble, 2001). Mollisols form in volcanic materials under dry conditions in temperate and tropical environments and under wetter conditions in the presence of basic volcanic ash (Shoji et al., 1990a; Chadwick et al., 1994). Warm/dry conditions promote formation of crystalline layer silicates rather than noncrystalline materials and leaching is limited leading to higher base saturation. Vertisols may also form in warm regions having a distinct dry season as volcanic ash weathers preferentially to smectite and vermiculite rather than noncrystalline materials (Yerima et al., 1987). In the semi-arid (xeric) moisture regime of California, Andisols are transformed to Inceptisols as noncrystalline materials are transformed to kaolins resulting in a decrease of the active Al and Fe fraction (Takahashi et al., 1993). Clay translocation is not a prevalent process in Andisols due to the difﬁculty in dispersing noncrystalline materials. However, once noncrystalline materials are transformed to layer silicate clays, translocation is greatly enhanced, especially in the xeric moisture regime where repeated wetting and drying leads to slaking and mobilization of clays. Formation of an argillic (clay-enriched) horizon results in transformation of Inceptisols to Alﬁsols/ Ultisols (Southard and Southard, 1987; Takahashi et al., 1993). Andisols are the typical product of weathering in both temperate and tropical environments with sufﬁcient moisture. As the degree of weathering increases, metastable noncrystalline materials are gradually consumed by transformation to more stable crystalline minerals (e.g., halloysite, kaolinite, gibbsite). This often leads to the alteration of Andisols to Inceptisols, Alﬁsols or Ultisols (Martini, 1976; Delvaux et al., 1989; Chen et al., 2001). Formation of Oxisols occurs in the perhumid tropics on highly weathered stable landscapes that have escaped rejuvenation by intermittent ashfall or erosion (Beinroth, 1982). Under cool to cold-humid conditions, Spodosols are often the dominant soil as cooler temperatures favor coniferous vegetation and predominance of the organic acid weathering/transport regime (Ugolini et al., 1977). Spodosols may form directly from fresh volcanic deposits or as an alteration product from Andisols.

IV. SOIL CLASSIFICATION At the international level, volcanic soils were ﬁrst recognized as a distinct category of soils, Andosols (from Japanese an, black and do, soil), in Japan in

124

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

1947 (Simonson, 1979). Andosols are currently deﬁned as a major soil group in the World Reference Base for Soil Resources (WRB) classiﬁcation system (FAO, ISRIC and ISSS, 1998). Volcanic soils were introduced as the Andept suborder of Inceptisols in the Seventh Approximation (Soil Survey Staff, 1960) and Soil Taxonomy (Soil Survey Staff, 1975). In 1978, Guy Smith proposed the reclassiﬁcation of Andepts into a new soil order of Soil Taxonomy termed Andisols (Soil Survey Staff, 1998). The classical expression of volcanic soils is roughly equivalent between Andosols (WRB) and Andisols (Soil Taxonomy) and both of these systems have been rigorously tested for the international classiﬁcation of volcanic soils. In addition to these international classiﬁcation systems, some national soil classiﬁcation systems [e.g., Classiﬁcation of Cultivated Soils in Japan, 3rd Approximation (Classiﬁcation Committee of Cultivated Soils, 1996); New Zealand Soil Classiﬁcation — Version 3.0 (Hewitt, 1998)], provide a more detailed separation of soils for speciﬁc conditions within a given country. Because of their international applicability and wide spread utilization, we will provide a brief discussion of Andosol (WRB) and Andisol (Soil Taxonomy) classiﬁcation criteria. The central concept of both Andosols and Andisols is soils, dominated by volcanic ejecta (having a large component of volcanic glass) that have a colloidal fraction dominated by amorphous and short-range-order materials (allophane, imogolite, ferrihydrite, and Al/Fe – humus complexes) resulting in pHdependent variable charge cation exchange capacity (CEC) and anion exchange capacity (AEC), high phosphate retention and low bulk density. Classiﬁcation is based on selected chemical, physical, and mineralogical properties acquired through rapid weathering of volcanic glass and is not based exclusively on parent material. As such, soils formed on volcanic ejecta may be classiﬁed in several other soil groups/orders in WRB (e.g., Regosols, Nitisols, Cambisols, Luvisols, and Vertisols) and Soil Taxonomy (e.g., Entisols, Inceptisols, Spodosols, Mollisols, Alﬁsols, Ultisols, and Oxisols). Similarly, there are a few reports of Andosols/Andisols formed in nonvolcanic parent materials in Spain (Garcia-Rodeja et al., 1987), USA (Hunter et al., 1987; Wilson et al., 1996), Nepal (Ba¨umler and Zech, 1994), and India (Caner et al., 2000). These latter soils have an active iron and aluminum fraction that is generally dominated by Al/Fe – humus complexes rather than noncrystalline inorganic materials, such as allophane and imogolite.

A. WORLD REFERENCE BASE FOR SOIL RESOURCES — ANDOSOLS WRB classiﬁcation divides soils into 30 soil groups based on soil properties deﬁned in terms of reference horizons (FAO, ISRIC and ISSS, 1998). Andosolsare deﬁned taxonomically by the presence of either an andic or vitric horizon. Andic horizons are characterized by the presence of high concentrations of active Al and Fe, deﬁned as those components extracted by acid oxalate (Alox, Feox, and

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 125

Siox from allophane, imogolite, ferrihydrite, and Al/Fe –humus complexes). Vitric horizons are distinguished from andic horizons by their lesser degree of weathering. As such, they are weakly weathered volcanic materials dominated by volcanic glass and other primary minerals derived from volcanic ejecta that display a weaker expression of andic properties. Andic horizons must have the following properties: 1. 2. 3. 4.

bulk density at ﬁeld-moisture water content , 0.9 g cm23, 10% or more clay and an Alox þ 12Feox value in the ﬁne-earth fraction $ 2%, phosphate retention $ 70%, a thickness of at least 30 cm.

Andic horizons are further subdivided into sil-andic horizons (dominated by allophane and/or imogolite) having Siox $ 0.6%, and alu-andic horizons (dominated by Al/Fe – humus complexes) having a Siox , 0.6%. These horizons may be alternatively deﬁned, as sil-andic horizons having an Alp/Alox , 0.5 and aluandic horizons having an Alp/Alox $0.5 (p ¼ pyrophosphate extractable). This subdivision is important for differentiating whether the andic properties result primarily from allophane/imogolite (sil-andic ¼ allophanic Andosols) or Al–humus complexes (alu-andic ¼ nonallophanic Andosols). Vitric horizons must have the following properties: 1. $ 10% volcanic glass and other primary minerals in the ﬁne-earth fraction; and either: 2. , 10% clay in the ﬁne-earth fraction; or 3. a bulk density . 0.9 g cm23; or 4. Alox þ 12Feox $ 0.4%; or 5. phosphate retention . 25%; and 6. a thickness of at least 30 cm. There are 25 lower-level soil units within Andosols that describe a wide range of soil properties, such as humus accumulation, forms of active Al/Fe, degree of weathering and leaching, drainage condition, presence of cemented layers, presence of buried horizons, and accumulation of carbonates or salts. These soil units provide important information regarding speciﬁc soil properties that strongly affect soil productivity, utilization, and management.

B. SOIL TAXONOMY — ANDISOLS Andisols are one of 12 soil orders in Soil Taxonomy and are deﬁned taxonomically by the presence of andic soil properties (Soil Survey Staff, 1998). Andic soil properties result primarily from the presence of allophane, imogolite,

126

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

ferrihydrite, and Al/Fe –humus complexes. Criteria for andic soil properties in Soil Taxonomy are nearly identical to andic horizons in WRB. Andic soil properties are deﬁned as soil materials with , 25% organic carbon that meet one or both of the following requirements: 1. In the ﬁne-earth fraction, all of the following: 1.1. Alox þ 12Feox $ 2.0; and 1.2. a bulk density, measured at 33 kPa water retention # 0.90 g cm23; and 1.3. a phosphate retention $ 85%; or 2. In the ﬁne-earth fraction, a phosphate retention $ 25%, $ 30% particles 0.02 to 2.0 mm in size, and one of the following: 2.1. Alox þ 12Feox $ 0.4% and, in the 0.02 –2.0 mm fraction, $ 30% volcanic glass; or 2.2. Alox þ 12Feox $ 2.0% and, $ 5% volcanic glass; or 2.3. Alox þ 12Feox totaling between 0.4 and 2.0%; and there is at least a proportional content of volcanic glass in the 0.02 to 2.0 mm fraction between 30 and 5% (this latter volcanic glass requirement is displayed graphical in Soil Taxonomy). Deﬁnition #1 is equivalent to the andic horizon in WRB while deﬁnition #2 is similar to the vitric horizon in WRB. Deﬁnition #2 assumes that there is sufﬁcient volcanic glass remaining in the 0.02– 2.0 mm fraction that continued weathering would result in Alox þ 12Feox values greater than 2%. There are seven subgroups of Andisols that are based on soil moisture regimes (aquic, aridic, xeric, ustic, udic), soil temperature regime (cyric), or low water retention characteristics due to a low degree of weathering (vitric). The global distribution of Andisols in these subgroups follows: vitrands (30.8%) < udands (30.6%) . cryands (28.0%) . ustands (6.9%) . xerands (3.5%) . torrands (0.2%) . aquands (, 0.1%) (Wilding, 2000). There are 12 unique elements identiﬁed at the great group level that are based on several factors including soil temperature and moisture regimes, cemented layers, degree of weathering, organic matter characteristics and accumulation, water retention properties, and drainage characteristics. Similarly, there are 24 unique elements identiﬁed at the subgroup level that identify various soil properties. In summary, Andosols (WRB) and Andisols (Soil Taxonomy) are differentiated on the basis of selected chemical, physical, and mineralogical properties rather than parent material alone (i.e., volcanic ejecta). As such, soils formed on volcanic materials may be classiﬁed in several other soil groups/orders in WRB and Soil Taxonomy and it is possible for soils formed on nonvolcanic parent materials to be classiﬁed as Andosols/Andisols. Both classiﬁcation systems have nearly identical requirements deﬁning andic/vitric horizons (WRB) and andic soil properties (Soil Taxonomy): acid oxalate extractable Al/Fe, phosphate retention, bulk density, and volcanic glass content. Both systems place strong

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 127

importance on the degree of weathering as reﬂected by vitric horizons in WRB and the vitric suborder in Soil Taxonomy for weakly weathered volcanic materials. WRB classiﬁcation places an emphasis on distinguishing between nonallophanic (alu-andic) and allophanic (sil-andic) Andosols as a high priority. Soil Taxonomy does not speciﬁcally distinguish between these two subdivisions of soils even though they indicate important differences in soil forming processes and soil properties. In contrast, Soil Taxonomy places much greater importance on soil temperature and moisture regimes by emphasizing these factors at the suborder and great group levels. Most of the lower-level units of Andosols are consistent with great group and subgroup units in Soil Taxonomy. Both systems have received rigorous international testing and review during their development. As a result, both classiﬁcation systems provide excellent classiﬁcation criteria for soils formed from volcanic materials. These classiﬁcation systems provide important information regarding soil properties, productivity, utilization, and management.

V.

CHEMICAL WEATHERING

A distinctive feature of soils derived from volcanic material is the dominance of noncrystalline and poorly crystalline components in the colloidal fraction. Preferential formation of these materials results, in part, from rapid weathering of volcanic glass, which shows low resistance to chemical weathering. In addition, the ﬁne particle size and vesicular nature (i.e., high surface area) and high porosity and permeability (i.e., effective leaching) of volcanic materials enhance weathering rates. Rapid weathering results in soil solutions becoming over saturated with respect to several noncrystalline components. It is presumed that rapid precipitation kinetics favor formation of metastable, noncrystalline constituents compared to their more stable crystalline mineral counterparts. Thus, rapid weathering is a key factor leading to the formation of andic soil properties in pyroclastic deposits. Volcanic materials are generally classiﬁed based upon their total silica content (SiO2%): rhyolite (70 – 100%), dacite (62 –70%), andesite (58 –62%), basaltic andesite (53.5 – 58%) and basalt (45 – 53.5%) (Shoji et al., 1975). The mineralogical composition of volcanic materials varies according to rock types; however, volcanic glass or glassy aggregates (i.e., glassy coated phenocrysts) are the dominant component of all types of volcanic ejecta (Fig. 5; Shoji, 1983). The chemical composition of volcanic glass is determined by the composition of the magma from which it forms. Because of the rapid cooling of extrusive magma, there is little segregation of crystals during the solidiﬁcation process. Volcanic glass is divided into colored and noncolored categories based on refractive indices (colored . 1.52, noncolored # 1.52). Colored glass is found in

128

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Figure 5 Relationship between rock types of volcanic materials and their mineralogy (adapted from Shoji, 1983; copyright Hakuyusha, with permission).

basaltic andesite and basalt rock types while noncolored glass is found in rhyolite, dacite, and andesite rock types. Colored glass has a low silica content and a high concentration of cations, such as Al, Fe, Ca, and Mg, which leads to substitution of these cations for silica in the glass structure. As a result, colored glass is more susceptible to weathering due to destabilization of the Si –O –Si framework in the glass by cations. The relative stability of mineralogical components with regard to chemical weathering in volcanic ejecta has been shown to follow (Aomine and Wada, 1962; Loughnan, 1969; Shoji et al., 1974; Mitchell, 1975; Yamada et al., 1978): Colored volcanic glass , noncolored glass < olivine , plagioclase , augite , hyperthene , hornblende , ferromagnetic minerals. Laboratory weathering of volcanic glass shows an initial period of rapid hydration and release of cations through surface exchange with aqueous hydrogen ions (White and Claassen, 1980; White, 1983). Incongruent dissolution forms a cation-depleted leached layer at the surface of the glass particle. The extent of cation depletion and the thickness of the leached layer both increase as the pH decreases (White, 1983). As weathering progresses, elemental release

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 129

rates are controlled by surface dissolution with concurrent diffusion of cations through the leached layer. A shift from incongruent to congruent dissolution presumably occurs when the increase in the diffusion length equals the rate of retreat of the solution– solid interface. Short-term (, 60 d) laboratory glass dissolution studies in the pH range of 5– 7 exhibit parabolic rate constants between 10214 and 10213 mol cm22 s20.5 for base cations, such as sodium, and linear rate constants between 10216 and 10215 mol cm22 s21 for silicon and aluminum (White and Claassen, 1980; White, 1983). Shoji et al. (1993b) measured rates of aluminum release from colored (basaltic-andesite composition) and noncolored (rhyolite composition) glass in the order of 10212 mol g21 s21 at pH 4.0. Release rates were 1.5 times greater for the colored glass reﬂecting its lower stability compared to noncolored glass. Weathering rates increased about 1.5 times for each 108C increase in temperature between 0 and 308C (Shoji et al., 1993b). Airfall tephra often undergo considerable chemical weathering within the eruption plume as the tephra interacts with acidic aerosols (H2SO4, HCl, HF, HNO3) during the eruption and subsequent transport. Once condensed on tephra particles, these acids are neutralized by hydrolytic reactions, which consume protons and liberate cations. These cations combine with the strong acid anions to form soluble salts that are quickly leached from the tephra. As a result, the initial leachates from the 1980 eruption of Mt. St. Helens tephra had a near neutral pH value with high concentrations of soluble salts composed primarily of base 2 cations (Ca2þ . Naþ .. Mg2þ . Kþ) and strong acid anions (SO22 4 .. Cl . 2 2 F .. NO3 ) (Dethier et al., 1981; Hinkley and Smith, 1987; Dahlgren and Ugolini, 1989a). In the immediate vicinity of volcanoes, where complete obliteration of the vegetation has occurred, a primary proton donor driving chemical weathering is carbonic acid (Ugolini et al., 1991). The primary source of carbonic acid in these systems is from the atmosphere because production of CO2 by soil organism respiration is low due to low organic matter concentrations in volcanic deposits. However, in those cases in which volcanic gas emanations continue from the volcanic vent, H2SO4 can be a prominent proton donor driving weathering reactions. In the immediate vicinity of Mt. St. Helens, the pH of precipitation ranged from 3.6 to 5.2 in the six years following the 1980 eruption with the acidity originating primarily from H2SO4. Emission of sulfur dioxide (SO2) averaged more than 1000 Mg d21 following the 1980 eruption (Casadevall et al., 1983) providing a considerable proton load to the atmosphere upon oxidation to H2SO4. Acidity originating from the precipitation in the vicinity of Mt. St. Helens was largely consumed by weathering reactions in the upper 5 cm of the pyroclastic ﬂow (Nuhn, 1987). Another example of chemical weathering on the barren landscape of the Mt. St. Helens pyroclastic ﬂows occurs in conjunction with lupine (Lupinus species, a nitrogen ﬁxer) establishment during primary succession. Protons are

130

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

generated as nitrogen ﬁxed by lupine becomes oxidized to nitrate and leaches along with base cations. Root and microbial respiration associated with lupine also enhance soil CO2 concentrations leading to higher carbonic acid concentrations in the soil solution. As a result, chemical weathering is enhanced in soils occupying the “islands of soil fertility” in the vicinity of the individual lupine plants (Nuhn, 1987). In areas where airfall tephra deposition does not kill the vegetation, chemical weathering is immediately inﬂuenced by biota that generate organic ligands and protons from organic acids, carbonic acid, and through uptake of cations in excess of anions. To examine the initial stages of chemical weathering in airfall tephra deposits, Dahlgren and coworkers (1989a; 1997; 1999) applied 5 and 15 cm of previously nonleached tephra from Mt. St. Helens to the surface of Spodosols in a subalpine forest ecosystem in the Washington Cascades. Soil solutions from the tephra layer were collected continuously over a four-year period and the elemental ﬂuxes were used to calculate chemical weathering rates. The primary proton donor in the fresh tephra layer was carbonic acid with a minor contribution from organic acids originating from leaching of the tree canopy and the incipient litter layer accumulating at the surface of the tephra layer (Dahlgren et al., 1999). The dominance of carbonic acid weathering was demonstrated by the near stoichiometric balance between base cations and the bicarbonate anion in tephra leachates. Solutions leached from the tephra layer indicated incongruent dissolution resulting in formation of a cation-depleted, silica-rich leached layer on the glass and mineral surfaces. Due to the near neutral pH values and low concentrations of complexing organic ligands, aluminum and iron were relatively insoluble and accumulated in the tephra layer rather than being leached. Calcium and sodium were released about 3– 4 times faster than silica based on their stoichiometry in the tephra. Field weathering rates, calculated as yearly averages, ranged between 10218 and 10217 mol cm22 s21 for sodium, calcium and silicon (Fig. 6). These rates are 1– 3 orders of magnitude less than those determined for glass and plagioclase minerals at steady state in laboratory dissolution experiments at pH 5 –7 and 258C. The principal factors contributing to the difference between ﬁeld and laboratory weathering rates are availability of water, temperature, and the extent of solution –solid mixing that serves to reduce the inﬂuence of solute diffusion on the rates. The ﬁeld weathering rates at this study site were substantially hindered by low soil temperatures (5.58C mean annual air temperature), which for at least six months each year are near freezing when the soils are buried by a snowpack. Similarly, when soil temperatures are higher during summer, there is a water deﬁcit that hinders chemical weathering. Thus, ﬁeld weathering rates are greatly attenuated because of the inverse relationship between optimal temperature and moisture conditions: (1) when moisture is most abundant, the soil temperatures are near freezing; and (2) when soil temperatures are highest, soil water availability and leaching are at their lowest.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 131

Figure 6 Weathering rates for 5 cm ash layer reported as a yearly average for the ﬁrst four years following addition of ash to the surface of a Spodosol in the western Cascade Range, Washington (Dahlgren et al., 1999).

Given these unfavorable ﬁeld conditions for chemical weathering, it is remarkable how fast dissolution rates are for the fresh tephra. Cation denudation rates from the 5 and 15 cm tephra layers (1.2 – 9.4 kmolc ha21 yr21) generally exceed cation denudation rates determined for entire watersheds (0.1 – 1.5 kmolc ha21 yr21) as reviewed by Sverdrup and Warfvinge (1988) and Sverdrup (1990). This comparison implies that the initial rates of tephra weathering are much more rapid compared to those of other silicate parent materials. Over the 4-year study period, weathering rates decreased by a factor of 3– 5 times. The largest fraction of the decrease occurred between the ﬁrst and the second year, while weathering rates were more similar in the third and the fourth year. The large initial decrease is probably attributable to the formation of the cation-depleted leached layer and to the rapid dissolution of easily weatherable materials (e.g., glass) and high surface area microparticles. Electron micrographs comparing the original, nonweathered tephra with the same tephra collected after 10 years of weathering show that there is a distinct loss of the sharp edges and vesicularity (high-energy sites) in the weathered tephra (Fig. 7). In a separate study, White et al. (1986) demonstrated the formation of a cation-depleted leached layer (. 10 nm) on tephra particles after two years of weathering in the soil environment using X-ray photoelectron spectroscopy (XPS). Weathering rates for the 5 cm tephra layer were about a factor of two greater than for the 15 cm tephra layer. The primary reason for this difference is believed to be due to a greater proton ﬂux of carbonic acid per unit of tephra in the 5 cm layer. The majority of the CO2 in the tephra layer originates from upward

132

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Figure 7 Volcanic ash from the 18 May 1980 eruption of Mt. St. Helens, Washington. Nonweathered ash (top) and the same ash weathered for 10 years (bottom) in a cool–humid environment in the Washington Cascades. Weathering results in a distinct loss of the sharp edges and vesicularity (Dahlgren et al., 1997; 1999).

transport of CO2 from the organic-rich soil horizons (~15 cm of O horizon) of the buried soil. Elevated concentrations of CO2 beneath the tephra layer originate from biological respiration (e.g., roots and microorganisms) and from protonation of HCO2 3 leaching from the overlying tephra layer. Cations released by weathering in the tephra layer (pH < 62 7) leach downward with bicarbonate to the acidic organic horizons (pH < 4) where H2CO3 reequilibrates with the high pCO2 (Fig. 8). At pH 4, H2CO3 decomposes to CO2 and H2O and the gaseous CO2 diffuses upwards to take part in another cycle of weathering and transport. If the ﬂux of CO2 transported upwards from the buried soil is similar

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 133

Figure 8 Schematic representation of the CO2 –H2CO3 –HCO2 3 weathering/transport cycle occurring between the ash layer at the surface of the soil and the buried organic soil layer (Dahlgren et al., 1999).

beneath the 5 and 15 cm tephra layers, the amount of H2CO3 available for weathering on a per unit volume or mass basis of tephra would be three times greater for the 5 cm layer. The fact that the measured weathering rates in the 5 cm treatment are only a factor of 2 greater may be related to kinetic limitations of weathering reactions or to more rapid consumption of easily weatherable minerals in the 5 cm layer. This example of weathering in airfall tephra deposits demonstrates a unique weathering pathway in which the buried organic-rich soil pumps protons upward to the tephra layer that acts as an alkaline trap for CO2. Thus, the overall process of tephra weathering appears to be controlled largely by solute/gas transport of the carbonate system (CO2 –H2CO3 – HCO3). In addition, the base cations leached from the tephra layer exchange with Hþ on the cation-exchange complex of the organic horizons resulting in an increased base saturation, somewhat higher pH levels, and enhanced plant availability of nutrient cations. An important variable controlling the mineralogy of the colloidal fraction in volcanic soils is the soil solution silicon activity. The initial rapid weathering of glassy volcanic materials provides an abundant source of silica to support silicarich minerals, such as opaline silica and halloysite. As weathering proceeds, the concentrations of highly weatherable materials, most importantly volcanic glass and glassy aggregates, are depleted and soil solution silica activities decrease. Studies examining long-term weathering of volcanic glass in tephra deposits indicate half-lives for volcanic glass ranging from 1650 to 7000 years (Kirkman and McHardy, 1980; Ruxton, 1988; Shoji et al., 1993b). It is important to

134

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

consider that weathering rates and volcanic glass half-lives will vary considerably depending on factors such as climate, vegetation, tephra depth, particle size and chemistry. In summary, rapid chemical weathering of volcanic ejecta is an important factor leading to the formation of a colloidal fraction dominated by noncrystalline and poorly crystalline constituents. These materials are in turn responsible for andic horizons/soil properties, which are the primary features distinguishing Andosols/Andisols as a distinctive soil category. Carbonic acid is an important proton donor driving weathering reactions in tephra. Carbonic acid weathering is an appreciable sink for carbon dioxide as gaseous carbon dioxide is sequestered as aqueous bicarbonate/carbonate. The rapid release of nutrient elements by weathering is also an important factor for rejuvenating soil forming processes and enhancing soil mineral nutrition.

VI. MINERALOGICAL CHARACTERISTICS Rapid weathering of volcanic materials often leads to a colloidal fraction dominated by noncrystalline or poorly crystalline materials. The rapid kinetics of nucleation for noncrystalline materials favor formation of these metastable solid phases relative to their more thermodynamically stable crystalline mineral equivalents (Ostwald step rule or the rule of steps; Stumm, 1992). Preferential precipitation of noncrystalline materials occurs because nucleation of a more soluble phase (i.e., noncrystalline materials) is kinetically favored over that of a less soluble phase (i.e., crystalline mineral) owing to the lower solid –solution interfacial tension of the more soluble phase (Stumm, 1992). Thus, nucleation of a noncrystalline phase will occur at a lower degree of supersaturation resulting in a higher nucleation rate than for crystalline minerals. Once nucleation is initiated, the rapid expansion in noncrystalline surface area that is nucleation-controlled will act to lower the degree of supersaturation, further reducing the probability of nucleation of the more crystalline phase (Robarge, 1999). Over time, crystalline minerals form at the expense of their metastable, noncrystalline precursors. The “aging” or “Ostwald Ripening” process (i.e., transformation of noncrystalline to crystalline phases via dissolution and reprecipitation) is typically a very slow process due to the small differences in free energy between the two phases. Climatic conditions play an important role in the formation of crystalline minerals as crystallization is promoted as the soil climate becomes warmer and dryer (Talibudeen, 1981; Schwertmann, 1985). Noncrystalline materials are more persistent under cooler soil conditions because crystallization is hindered by the low input of thermal energy. Seasonal soil desiccation has been shown to enhance transformation of noncrystalline

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 135

materials to crystalline minerals (Quantin, 1992). Fieldes (1966) proposed that a pronounced dry period led to crystallization via formation of cross-linkages within aluminosilicate gels during the period of dehydration. Thus, periods of wetting and drying promote formation of crystalline minerals relative to noncrystalline materials. The dominant colloidal assemblage of soils formed in volcanic materials varies widely depending on factors such as the composition of the parent material, stage of soil formation (i.e., degree of weathering), pH, soil temperature and moisture regimes, and the accumulation of organic matter (Lowe, 1986; Shoji et al., 1993a). The weathering environment deﬁned by these factors regulates Al and Si activities, which along with soil climate are the primary determinants of soil colloid genesis and transformation. Three colloidal assemblages tend to predominate in weakly to moderately weathered soils derived from volcanic materials: (1) Al– humus complexes, often in association with hydroxy-Al interlayered 2:1 layer silicates and opaline silica, (2) allophane and imogolite, and (3) halloysite. Ferrihydrite is often the predominant iron (hydr)oxide formed during the early stages of weathering in all volcanic materials (Childs et al., 1991). Authigenic formation of 2:1 layer silicates is limited in volcanic materials, however, formation of smectites and 1:1 –2:1 mixed-layer clays (e.g., halloysite –smectite) appears to be common in soils derived from volcanic materials of basaltic composition in tropical and subtropical areas (Quantin et al., 1988; Delvaux et al., 1990b, 1992; Quantin, 1992; Delvaux and Herbillon, 1995). Noncrystalline and poorly crystalline materials contribute to many of the unique physical and chemical properties of volcanic soils due to their high surface area, variable charge surfaces, high solubility, rapid reaction kinetics and unusual physical properties. In the following sections, we brieﬂy describe the major colloidal constituents found in soils derived from volcanic materials and discuss genesis and transformation pathways. For a more detailed discussion, refer to reviews by Wada and Harward (1974), Wada (1980; 1985; 1986; 1989), Yoshinaga (1988), Parﬁtt and Kimble (1989), Parﬁtt (1990), and Shoji et al. (1993a).

A. COMMON COLLOIDAL CONSTITUENTS 1.

OF

VOLCANIC SOILS

Allophane

Allophane is a group name given to a series of naturally occurring, noncrystalline, hydrous aluminosilicates with a varying chemical composition (van Olphen, 1971). Allophane consists of hollow, irregularly spherical particles with outside diameters of 3.5– 5.0 nm and a wall thickness of 0.7 –1 nm. There is evidence that micopores exist in the allophane spherule with diameters ranging between 0.3 and 2.0 nm (Paterson, 1977). The chemical composition generally ranges from an Al:Si atomic ratio of 1:1 – 2:1. Based on the structural

136

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

and chemical characterization of natural allophanes, three major types of allophanes have been identiﬁed: Al-rich (proto-imogolite allophane), Si-rich (halloysite-like or defect kaolin), and hydrous feldspathoid allophanes (Farmer and Russell, 1990). Al-rich allophane (proto-imogolite) has the same local atomic arrangement (short-range-order) and chemical composition (Al:Si ¼ 2:1) as imogolite; however, they differ morphologically (Parﬁtt and Henmi, 1980). The structure consists of a gibbsite-like sheet of octahedrally coordinated Al with Si bound to three Al– O of the gibbsite sheet and the remaining Si – OH group pointing toward the center of the sphere. It appears to form as fragments of the imogolite structure, depending on solution conditions they either form tubular (imogolite) or spherical morphology (Farmer and Russell, 1990). Si-rich allophanes have an Al:Si atomic ratio close to 1:1 with Al primarily in octahedral coordination (gibbsite sheet) and silica occurring as both polymerized silicate and orthosilicate groups (Parﬁtt, 1990). Spectroscopic studies reveal that Si-rich allophane contains both defect kaolin- and imogolite-structural units with the portion of polymerized Si increasing as the Al:Si atomic ratio decreases (Parﬁtt et al., 1980; Goodman et al., 1985; Shimizu et al., 1988). The isolated orthosilicate groups are associated with the gibbsite sheet of the 1:1 kaolin layer and the orthosilicate groups penetrate the inner sheet of silica tetrahedra through the holes formed by defects (MacKenzie et al., 1991). Thus, its structure is similar to a kaolin mineral with defects in the tetrahedral sheet. Reports of allophane with Al:Si atomic ratios between 2:1 and 1:1, based on selective dissolution analyses, may result from mixtures of the Al-rich and Si-rich allophane end-members. However, laboratory synthesis experiments utilizing a range of Al:Si atomic ratios in the matrix solutions are able to produce allophanes with a range of Al:Si atomic ratios from 1 to 2. Attempting to understand the chemical composition of allophanes based on selective dissolution analysis is not very reliable. Further research is needed to verify the nature of allophanes with apparent Al:Si atomic ratios between 1 and 2. Hydrous feldspathoid allophane contains no imogolite unit structures and has a signiﬁcant amount of tetrahedrally coordinated Al. The silicate layer is substantially polymerized with 1:3 Al for Si substitution (Childs et al., 1990). This is the same degree of substitution as in muscovite and suggests that this allophane phase has a structure based on a similar condensed silicate sheet. The proposed structure consists of a more or less complete tetrahedral sheet on the outer surface with an incomplete octahedral Al sheet on the inside. It preferentially forms at higher concentrations of silicon and at slightly acidic to neutral pH values, consistent with the formation of tetrahedrally coordinated Al species in solution. The most notable occurrence of natural feldspathoid allophane is as stream channel deposits below Silica Springs on the ﬂanks of Mt. Ruapehu, New Zealand (Wells et al., 1977; Childs et al., 1990).

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 137

2.

Imogolite

Imogolite was ﬁrst described by Yoshinaga and Aomine (1962) in a soil derived from glassy volcanic ash, known as “imogo.” It has a distinct tubular morphology consisting of bundles of well-deﬁned ﬁne tubes with inner and outer diameters of about 1.0 and 2.0 nm, respectively. Imogolite displays long-range order only along the length of the tubes, which may extend several microns. It has a ﬁxed chemical composition [(OH)3Al2O3SiOH] with an Al:Si atomic ratio of 2:1. The external surface consists of a gibbsite-like sheet with orthosilicate groups coordinated through oxygen with three aluminum atoms and one free OH group pointing toward the center of the tube (Cradwick et al., 1972). Aluminum is found exclusively in octahedral coordination. The precursor to imogolite is a sol consisting of small fragments of protoimogolite allophane sheets that are too small to show tubular morphology. The formation of imogolite from proto-imogolite sheets is a time dependent process akin to crystallization (Farmer and Fraser, 1979; Wilson et al., 2001). Crystallization is hindered by the presence of iron and complexing anions, such as weak organic acids. Thus, imogolite formation is favored under conditions of carbonic acid weathering (generally B horizons) as opposed to the organic acid weathering regime (often A horizons), which produces conjugate bases with complexing capacity.

3.

Halloysite

Halloysite is a common constituent in volcanic soils and is often the dominant colloidal constituent in Si-rich environments (Parﬁtt and Wilson, 1985). Halloysite is a 1:1 aluminosilicate mineral, which is characterized by a diversity of morphologies (e.g., spheroidal, tubular), chemical composition, speciﬁc surface area, structural disorder, and physiochemical properties (e.g., CEC, ion selectivity). Halloysite is hydrated (1.0 nm basal spacing) by two interlayer water molecules; however, they are susceptible to dehydration (0.7 nm basal spacing) under climatic conditions showing a distinct seasonal moisture deﬁcit (Churchman and Gilkes, 1989; Takahashi et al., 2001). In terms of nomenclature, the various forms of halloysite are often described as hydrated halloysite for those having a 1.0 nm basal reﬂection and dehydrated halloysite for those having a 0.7 nm basal reﬂection that is expandable to 1.0 nm by formamide treatment. In the xeric moisture regime of California, halloysite (1.0 nm with or without formamide treatment) and kaolinite (0.7 nm with formamide) concentrations show an inverse relationship with depth, with the content of halloysite increasing down the proﬁle (Takahashi et al., 2001). The kaolinite in the surface horizons has the same tubular morphology as the halloysite at depth suggesting that hydrated halloysite transforms to kaolinite upon dehydration. The term “tubular kaolinite”

138

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

has been used to describe kaolins with tubular morphology and a 0.7 nm basal reﬂection that does not expand with formamide (Churchman and Gilkes, 1989). Halloysite displays a wide range of structural disorder, which is primarily due to Al-vacancy displacements in the octahedral sheet and to random stacking of structural layers (Churchman and Theng, 1984; Soma et al., 1992; Newman et al., 1994). These vacancies may originate from nonstoichiometric substitution of Fe3þ for Al3þ in the octahedral sheet (Soma et al., 1992). A number of authors (e.g., Wada and Mizota, 1982; Delvaux et al., 1990b; Malucelli et al., 1999) have shown instances of halloysite incorporating appreciable concentrations of iron (up to 17% Fe). Variable effects of hydration might also add to the degree of disorder in halloysite. Vacancies in the octahedral sheet and isomorphous substitution of Al3þ for Si4þ in the tetrahedral position have been suggested as mechanisms by which halloysite may acquire permanent negative charge. The CEC values for halloysite generally range between 2 and 10 cmolc kg21, although values in excess of 20 cmolc kg21 clay have been reported (Bailey, 1990; Norrish, 1995).

4. 2:1 Layer Silicate Clays Layer silicate minerals of the 2:1 type are often found in soils derived from volcanic ejecta. Their occurrence in young volcanic materials is variously ascribed to in situ pedogenic origin, eolian addition, or inheritance from hydrothermally altered materials in the parent material. These layer silicates are the dominant clay mineral in the nonallophanic group of Andosols/Andisols. During the early stages of weathering, mica, vermiculite, and smectite are present as the dominant 2:1 layer silicates. As weathering advances, hydroxy-Al interlayering occurs in vermiculite and smectite resulting in a sink for Al released by weathering (Shoji and Fujiwara, 1984). Hydroxy-Al interlayering of 2:1 layer silicates reduces their susceptibility to alteration and increases their thermodynamic stability (Karathanasis et al., 1983). The authigenic formation of 2:1 layer silicates in volcanic soils has long been a topic of debate. Inheritance of 2:1 layer silicates formed by hydrothermal alteration in the volcanic cone prior to eruption and deposited with volcanic ejecta has been documented (e.g., Kondo et al., 1979; Pevear et al., 1982; LaManna and Ugolini, 1987; Jongmans et al., 1994). Recent isotopic evidence demonstrates the importance of eolian transport as a major source of 2:1 layer silicates in areas such as Japan, Canary Islands, Tanzania, and Hawaii (e.g., Dymond et al., 1974; Inoue, 1981; Inoue and Naruse, 1987; Mizota et al., 1988; Mizota and Matsuhisa, 1995). Eolian deposition explains the preferential accumulation of 2:1 layer silicates in surﬁcial horizons of older landscapes and those having higher precipitation. While the formation of 2:1 layer silicates in volcanic materials in the udic soil moisture regime cannot be ruled out, current evidence strongly suggests that

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 139

eolian deposition is the major source of 2:1 layer silicates, with lesser amounts inherited from pyroclastic deposits. Weathering of basalt results in the formation of smectite as the initial crystalline weathering product in seasonally dry environments (Glassmann and Simonson, 1985; Quantin, 1992). Under a tropical environment with a distinct dry season, Quantin (1992) identiﬁed a beidellite-phase with a crumpled morphology after 500 years and it transformed to a more crystalline form after 2000 years. In contrast, no 2:1 layer silicates were detected on similar parent materials under perudic conditions in the same region. 5.

1:1 – 2:1 Mixed-layer Clays (Halloysite-Smectite, Kaolinite-Smectite)

The presence of 1:1 – 2:1 mixed-layer clays as weathering products in volcanic materials has been demonstrated in basaltic volcanic ash, primarily in tropical and subtropical areas (Quantin et al., 1988; Delvaux et al., 1990 a, b, 1992; Delvaux and Herbillon, 1995). The occurrence of such clays explains the high CEC and Kþ selectivity of soils containing these clay minerals. Halloysite – smectite mixed-layer clays form in silica- and base-rich environments and are highly sensitive to the microenvironment. Weathering of basalt leads to smectite formation on the fresh surfaces of the pyroclastic materials while halloysite forms in the more open microenvironment subject to more leaching (Delvaux and Herbillon, 1995). It is therefore expected that 1:1 – 2:1 mixed-layer clays would form from primary minerals via a syngenetic (co-neoformation) pathway under ’intermediate’ microenvironmental conditions. In contrast, the formation pathway of kaolinite-smectite is a desilication process in which iron-rich 2:1 clays transform to 1:1 clays and iron (hydr)oxides. The intermediate step of such a transformation consists of 1:1 –2:1 mixed-layer clays in which the smectite content and layer charge both decrease as weathering proceeds (Delvaux et al., 1990b). Delvaux et al. (1990b) estimated a smectite content of 14 to 31% in halloysite – smectite mixed-layer clays from Western Cameroon. The smectite component contained octahedral iron and appeared to be a high-charge beidellite phase. Due to the difﬁculty of detecting 1:1 – 2:1 mixed-layer clays by standard mineralogical methods, the occurrence of these minerals may be appreciably underestimate in volcanic soils. Thus, special efforts should be made to verify the presence/absence of 1:1 – 2:1 mixed-layer clays in soils displaying halloysite-rich clays with high CEC and Kþ selectivity. 6.

Opaline Silica

Two types of opaline silica are common in young volcanic soils: pedogenic (commonly known as laminar opaline silica) and biogenic (plant opal and

140

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

diatoms). Laminar opaline silica is predominantly found in surface horizons during the early stages of weathering of volcanic materials (Shoji and Masui, 1971). It was also found in the E horizon of Spodosols derived from recent volcanic ash in the Washington Cascades (Ugolini et al., 1991) and as a cementing agent in indurated subsoil horizons (tepetate horizons) of volcanic soils in Mexico (Hidalgo et al., 1995). Laminar opaline silica is most often circular or elliptical in shape and is most abundant in the 0.4 to 2 mm fraction. Opaline silica forms in Si-rich environments from supersaturated solutions resulting from surface evaporation and possibly freezing. Preferential incorporation of Al into humus complexes and hydroxy-Al interlayers rather than aluminosilicate compounds may result in high silica activities that favor the formation of opaline silica. Opaline silica may persist in volcanic soils for more than 5000 years; however, it is progressively lost as chemical weathering rates decrease over time resulting in lower availability of silica (Shoji and Masui, 1971). Opaline silica does not have a signiﬁcant inﬂuence on soil properties other than buffering soil solution silica activities at a relatively high level during the early stages of soil development.

7.

Ferrihydrite

Iron-bearing weathering products occur primarily as noncrystalline (hydr)oxides in young volcanic soils. The dominant noncrystalline (hydr)oxide is ferrihydrite, a short-range-order Fe oxyhydroxide with a bulk composition of 5Fe2O3(9H2O) (Schwertmann and Taylor, 1989). Ferrihydrite resembles hematite structurally, except that some Fe positions are vacant and some O and OH groups are replaced by water molecules (Towe and Bradley, 1967). The degree of crystalline order varies widely, with “2-line” and “6-line” ferrihydrite differentiated by the number of diffraction peaks observed by X-ray diffraction (Schwertmann and Cornell, 1991). Ferrihydrite appears as individual spherical particles ranging in size between 2 and 5 nm. These particles become highly aggregated forming aggregates 100 – 300 nm in diameter. Ferrihydrite is metastable and with time converts to stable Fe (hydr)oxides, usually to goethite under temperate or cool, humid climate, and to hematite under warmer, dryer climate. The high afﬁnity of ferrihydrite for sorption of silica and humic substances may contribute to its formation by impeding the crystallization process (Schwertmann, 1988). Formation of siliceous iron (hydr)oxides from peralkaline volcanic ash was promoted by drier climatic condition in the Great Rift Valley of Kenya (Wakatsuki and Wielemaker, 1985). The SiO2/Fe2O3 molecular ratios of citrate –dithionite –bicarbonate extracts for these siliceous iron (hydr)oxides ranged from 1.8 to 4.2.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 141

8.

Hisingerite

An iron silicate mineral termed hisingerite has been proposed as an early product of weathering in iron-rich volcanic deposits (Sudo and Nakamura, 1952; Quantin, 1992). Hisingerite has been variously described as a noncrystalline iron silicate (silica-rich ferrihydrite), as a ferric allophane, as ferric halloysite, and as a poor crystalline nontronite (Eggleton and Tilley, 1998). A rigorous evaluation of hisingerite specimens by Eggleton and Tilley (1998) showed that pure specimens were a ferric analogue of spherical halloysite [Fe2Si2O5(OH)4(2H2O)]. Transmission electron microscopy shows a fabric of concentric spheres, with a diameter of about 14 nm, and shell walls 0.7 or 1.0 nm thick. The controversy concerning the nature of hisingerite is most likely the result of several impurities in soil samples.

9.

Aluminum – and Iron –Humus Complexes

Aluminum – and iron– humus complexes are often the dominant form of active Al and Fe in acidic (pH , 5), organic-rich horizons, especially in nonallophanic Andisols (Shoji and Fujiwara, 1984). This fraction is typically deﬁned as the Al and Fe fraction extracted by pyrophosphate reagent (McKeague, 1967). While the Al– humus fraction can be large in some soils, Fe– humus complexes are generally much lower because iron has a greater stability as Fe (hydr)oxides compared to humus complexes (Wada and Higashi, 1976). Metal –humus complexes are believed to form primarily by the interaction of metals with carboxylic functional groups. The complexing capacity of humus increases as the degree of humiﬁcation increases. The degree of metal complexation by humic substances can be evaluated by examining pyrophosphate extractable Al, Fe, and C using the ratio: (Alp þ Fep)/Cp. For most Andosols/Andisols, the ratio ranges between 0.1 and 0.2 (Inoue and Higashi, 1988). The degree of complexation is determined by the stability constants for metals with complexing functional groups, solution pH, the concentration and speciation of aqueous Al and Fe, and the concentration of competing aqueous species. It is suggested that the accumulation and stabilization of humus in volcanic soils is, in part, due to the formation of metal – humus complexes (Percival et al., 2000).

B. FORMATION

TRANSFORMATION VOLCANIC SOILS

AND

OF

COLLOIDS

IN

The composition of the colloidal fraction in volcanic soils is regulated by elemental activities, pH, soil temperature, and seasonal desiccation of the soil proﬁle (Table II). During the early stages of weathering, elements are rapidly

142

Colloidal constituent

pH

Al–humus complexes Hydroxy-Al interlayered 2:1 layer silicates Opaline silica Al-rich allophane Imogolite Si-rich allophane Halloysite 2:1 layer silicates

,5 5– 6 4– 6 5– 6 5– 6 6– 7 5– 6 6– 8

Ferrihydrite

4– 7

Goethite

Al activity

Si activity

Humus content

Base saturation

Seasonal desiccation

Soil temperature

Weathering regime

Promotes Promotes

Freezing Cool Cool Cool Warm Warm

Organic acid Organic acid or carbonic acid Organic acid Carbonic acid Carbonic acid Carbonic acid Carbonic acid Carbonic acid

high

Low Moderate Moderate Moderate Moderate Low

High Moderate Moderate High High High

High Low Low Low Low Low

Promotes

Moderate to high Low

Moderate

Hinders

Cool

Low

Promotes

Cool

Low Moderate to high

Organic acid or carbonic acid Carbonic acid

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Table II Soil Environmental Conditions Favoring the Formation of Colloidal Constituents in Soils Formed in Volcanic Materials

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 143

released from volcanic glass and primary minerals resulting in high elemental activities. Elemental activities are subsequently changed by complexation with organic matter, incorporation into colloidal constituents, consumption of easily weatherable constituents, and leaching or desiccation. Wetting and drying cycles coupled with soil temperature regimes also have a major inﬂuence on the colloidal fraction through their inﬂuence on crystallization. In addition, complexing and sorbed anions (e.g., organic acids, sulfate, silicate) can hinder the crystallization processes by blocking lattice sites (Huang and Violante, 1986). Humus and 2:1 layer silicates have a strong afﬁnity for aluminum resulting in their preferential formation until they become effectively saturated. Al– humus complexes with or without hydroxy-Al interlayered 2:1 layer silicates are the dominant colloidal constituents in nonallophanic Andosols. They form in areas with higher precipitation resulting in greater leaching and lower pH values (pH , 5). At low pH values, aluminum exists primarily as Al3þ which is strongly complexed by organic acids and does not undergo appreciable hydrolysis and polymerization, processes necessary for formation of aluminosilicate constituents. In addition to humus, 2:1 layer silicates act as a sink for Al released by weathering. Soil pH values in the range of 5 – 6 have been shown to favor polymerization of hydroxy-Al polymers and their incorporation into the interlayer of 2:1 layer silicates (Jackson, 1963b). Nonallophanic Andosols often have a considerable amount of 2:1 minerals showing a wide range in the degree of hydroxy-Al interlayer ﬁlling. Preferential incorporation of Al into Al– humus complexes and hydroxy-Al interlayers effectively competes for soluble Al so that little Al is available for formation of allophanic materials. This inhibition mechanism has been termed the “anti-allophanic effect ” (Shoji et al. 1993) following the “anti-gibbsitic effect ” coined by Jackson (1963b). As a result, there is an inverse relationship between the concentrations of Al– humus complexes plus 2:1 layer silicates and concentrations of allophane/imogolite (Fig. 4). Once humus and 2:1 layer silicates are saturated with respect to aqueous Al3þ activities, excess Al3þ is available for combining with silica to form allophane and imogolite. Formation of allophane/imogolite occurs at (Alp þ Fep)/Cp values greater than about 0.1 –0.2, corresponding to maximum Al-complexation by humus (Parﬁtt and Saigusa, 1985; Dahlgren and Ugolini, 1991). Allophane and imogolite formation is favored in the pH range 5 – 7, coinciding with carbonic acid as the major proton donor. Within this pH range, soluble Al undergoes hydrolysis and polymerization reactions. The polymerized Al may then combine with silica to form allophane and/or imogolite structures. Al-rich allophane is composed of proto-imogolite subunits and appears to form preferentially under conditions of greater supersaturation and in the presence of complexing or sorbed anions. More rapid precipitation kinetics and blockage of precipitation sites by sorbing anions lead to a more disordered proto-imogolite allophane structure in

144

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

which the proto-imogolite fragments form irregular edge-to-edge associations and a hollow spheroidal morphology. Slower precipitation kinetics coupled with the relative absence of complexing/sorbing anions allows formation of a more ordered arrangement of the proto-imogolite subunits into the thread-like imogolite strands. Proto-imogolite allophane cannot transform to imogolite without dissolution and reprecipitation (Farmer and Russell, 1990). In silica-rich environments, the Al:Si atomic ratio of the allophane approaches 1:1 and consists of a disordered spheroidal morphology. In the case of the hydrous feldspathoid allophane from Silica Springs, New Zealand, rapid neutralization of acidity by CO2 degassing in the stream channel results in the precipitation of allophane within a few hundred meters of the spring outlet. Thus, allophane precipitation kinetics can be extremely fast in nature. In the soil environment, allophane and imogolite are believed to form in situ through carbonic acid weathering in A and B horizons of Andisols (Ugolini et al., 1988) and in the B horizons of tephritic Spodosols (Ugolini and Dahlgren, 1987; Dahlgren and Ugolini, 1989b). The carbonic acid weathering regime favors allophane/imogolite formation because it occurs in the preferred pH range of 5 –7 (promoting hydrolysis and polymerization of Al) and the bicarbonate anion does not complex Al or appreciably sorbed to mineral surfaces to interfere with crystallization. Cooler soil temperatures and the lack of a distinct dry season also appear to favor formation of poorly ordered allophane and imogolite relative to their more crystalline counterparts (e.g., halloysite). Halloysite forms in silica-rich environments, such as regions with low rainfall and leaching, in buried soil layers receiving silica from above, and in zones with restricted drainage that prevent leaching. Halloysite may occur in the clay, silt, and sand fractions (Simonett and Bauleke, 1963; Takahashi et al., 2001). Halloysite may crystallize directly from the dissolution products of volcanic glass and primary minerals, alteration of feldspars, or as a transformation product of allophane/imogolite via dissolution and reprecipitation. Halloysite formation is favored by high silica activities (. 250 – 350 mM) and by warmer and drier conditions that tend to increase dissolved silica concentrations and promote crystallization (Parﬁtt and Wilson, 1985). Halloysite is found as the dominant mineral in volcanic materials where precipitation is generally less than about 1500 mm (Parﬁtt et al., 1983; Lowe, 1986; Mizota and van Reeuwijk, 1989; Takahashi et al., 1993). Seasonal desiccation of the upper soil horizons leads to dehydration of 1.0 nm halloysite to 0.7 nm halloysite/kaolinite that may or may not be expandable to 1.0 nm with formamide. Halloysite is also favored by high silica concentrations associated with poorly drained subsoil horizons (Lowe, 1986) and buried soil environments (Saigusa et al., 1978). There has been considerable controversy concerning the origin of 2:1 layer silicates in soils derived from volcanoclastic materials. Initially, 2:1 layer silicates were thought to form by an in situ synthesis process starting with amorphous precursors. More recently, this theory has been largely discarded in favor of eolian

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 145

sources and inheritance. However, the formation of 2:1 layer silicates in soils formed on basaltic parent materials in tropical environments having a pronounced dry season is well documented (Quantin, 1992). Smectite formation is favored by high pH, silica activities, and base saturation, along with warm soil temperatures and seasonal drying that promote crystallization. The leaching environment (a function of precipitation) strongly affects the stability of smectite in the tropical areas. Smectite is stable for only a short time (, 10,000 years) under humid conditions, while no smectite persists under perhumid conditions (Quantin, 1992). In contrast, smectite (beidellite) was a stable clay mineral along with halloysite (1.0 dehydrating to 0.7 nm) in tropical regions with a pronounced dry season. Similarly, the presence of 1:1 – 2:1 mixed-layer clays (halloysite-smectite and kaolinite-smectite) has been shown in basaltic volcanic ash of the tropical and subtropical areas (Quantin et al., 1988; Delvaux et al., 1990a, b, 1992; Delvaux and Herbillon, 1995).

VII. SELECTED CHEMICAL CHARACTERISTICS OF VOLCANIC SOILS A. ORGANIC MATTER ACCUMULATION Accumulation of organic matter is a characteristic property of Andisols. Soils formed in volcanic materials contain the largest accumulations of organic carbon among the mineral soil orders (Eswaran et al., 1993). Due to the high productivity of volcanic soils, there is typically a large annual input of detritus to the soil organic matter pool. In the case of grassland vegetation, a large component of detritus is incorporated directly into the mineral soil horizons (Shoji et al., 1990b). Once in the soil environment, organic matter preservation results from burial of soils by repeated additions of volcanic ash, chemical interactions with polyvalent cations and noncrystalline inorganic materials (e.g., allophane, imogolite, ferrihydrite), and physical protection from the microaggregation that these materials impart to soil structure. The 14C ages of humic acids extracted from A horizons of Andisols ranged from modern to 30,000 YBP, with the majority in the range 1000– 5000 YBP (Inoue and Higashi, 1988). It was concluded that mean residence times of organic carbon in Andisols were appreciably greater than for Mollisol A horizons and Spodosol Bh horizons. Burial of soils by addition of volcanic ash is an important factor contributing to subsurface accumulation of organic matter (Fig. 3). Burial by tephra brings organic materials into intimate association with mineral components, reduces aeration and decreases interactions with soil macro- and mesofauna whose processing greatly enhances decomposition rates. Organic-rich buried soil

146

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

horizons may be distinctly evident in polygenetic soils for several thousand years following burial (Saigusa et al., 1978; Shoji et al., 1993a). Organic matter stabilization may occur through the formation of Al/Fe – humus complexes and sorption to allophane, imogolite, and ferrihydrite. Torn et al. (1997) found that soil organic carbon accumulation was related to concentrations of noncrystalline materials along a 4-million year soil chronosequence formed on basaltic lava in Hawaii. Soil organic matter pools reached a maximum after 150,000 years and then decreased as noncrystalline minerals declined and more stable crystalline minerals accumulated. Speciﬁcally, allophane was shown to have a protective effect on soil organic matter as demonstrated by decreased organic C mineralization rates in volcanic soils (Zunino et al., 1982; Boudot et al., 1988, 1989). Noncrystalline inorganic components have a strong afﬁnity for sorption of degradation enzymes and organic matter substrate (Wada, 1977; Tate and Theng, 1980). It is generally accepted that sorption reactions provide a mechanism for stabilization of organic matter against biodegradation. The protective capacity of soil clays appears to be more a function of the surface area available for sorption of the organic matter than the actual amount of clay present (Baldock and Nelson, 2000). Thus, the high speciﬁc surface areas of noncrystalline constituents in volcanic soils would be expected to have a high capacity to stabilize soil organic matter. These constituents are also responsible for the high phosphorus sorption capacity that may further hinder decomposition by contributing to phosphorus deﬁciency of microorganisms (Brahim, 1987). Other studies show that organic carbon concentrations in Andisols are more strongly associated with metal – humus complexes than with concentrations of noncrystalline materials. Accumulation of organic C in the upper 35 cm of Andisols displays a strong linear relationship with pyrophosphate-extractable Al þ Fe (Inoue and Higashi, 1988). Similarly in New Zealand soils, allophane concentrations were unrelated to soil carbon, however, the content of pyrophosphate extractable Al correlated strongly with soil carbon concentrations (Percival et al., 2000). Complexation of multivalent cations (e.g., Al3þ and Fe3þ) by humic substances results in organic bearing functional groups becoming more condensed and less susceptible to biological attack (Baldock and Nelson, 2000). With increasing acidity, organic matter may be protected against biodegradation by Al toxicity to microorganisms (Tokashiki and Wada, 1975). Organic matter may be physically protected from microbial and enzyme attack by soil structural conditions that limit the accessibility of soil organic matter to decomposers (Baldock and Nelson, 2000; Gregorich and Janzen, 2000). Volcanic soils have an abundance of microaggregates resulting from the occurrence of noncrystalline materials with variable charge surfaces (Warkentin et al., 1988). In a review of the literature, Baldock and Nelson (2000) reported that physical protection might result from the ability of clays to encapsulate organic materials (Tisdall and Oades, 1982), the burial of organic C within aggregates

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 147

(Golchin et al., 1994, 1997), and the entrapment of organic C within small pores (Elliott and Coleman, 1988). For example, Kilbertus (1980) indicated that bacteria can only enter into pores with diameters greater than about 3 mm. Thus, a signiﬁcant fraction of soil organic matter in volcanic soils may be inaccessible to decomposing organisms. Due to the high water retention by microaggregates in volcanic soils, it is also plausible that anaerobic conditions may hinder decomposers in the aggregate interiors, even under otherwise well-aerated conditions. Breakdown of microaggregates, such as by tillage operations, may expose a large fraction of the physically protected organic matter to microorganisms resulting in a rapid loss of organic matter.

B. ALUMINUM DYNAMICS Processes regulating aluminum solubility and release/retention kinetics play an important role in determining soil mineralogy (i.e., formation of metastable noncrystalline materials) and crop productivity (i.e., Al toxicity). Studies examining Al solubility and Al dissolution kinetics in volcanic soils with a large fraction of active Al (i.e., oxalate-extractable Al) show fast reaction kinetics and rapid attainment of an apparent equilibrium (Dahlgren and Walker, 1993; Dahlgren and Saigusa, 1994; Takahashi et al., 1995; Takahashi and Dahlgren, 1998). In soils containing hydroxy-Al polymers (interlayered 2:1 layer silicates), allophane/imogolite, Al – humus complexes, and exchangeable Al, there appears to be a simultaneous equilibrium existing between all four phases (Fig. 9;Dahlgren and Walker, 1993). The rapid Al release/retention associated with exchangeable Al and Al –humus complexes results in apparent equilibrium from over- and under-saturation within a 20 min residence time (Dahlgren et al., 1989). It appears that hydroxy-Al polymers ultimately regulate Al solubility in this case and that they have a solubility similar to that of synthetic gibbsite,

Figure 9 The active Al fractions in many volcanic soils appear to form a simultaneous equilibrium between the various solid-phase constituents. Rapid Al dissolution kinetics for exchangeable Al and Al–humus complexes allows soil solutions to quickly attain an apparent equilibrium.

148

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

which has a log K *SO of 8.1 for the reaction: Al(OH)3 þ 3Hþ ¼ Al3þ þ 3H2O (May et al., 1979). Experimental data showed a slope of about 2.7, compared to the theoretical value of 3.0 for gibbsite. A slope less than 3.0 may indicate that the formula for hydroxy-Al polymers is AlðOHÞ0:3þ 2:7 , the positive charge satisfying the negative charge of the 2:1 layer silicate (Huang, 1988). Based on the assumption that hydroxy-Al polymers control Al3þ activities as a function of Hþ activities, the silica activities of soil solutions will be buffered by allophane/imogolite precipitation and dissolution: hskip26pt2AlðOHÞ3 þ H4 SiO4 ¼ Al2 SiO3 ðOHÞ4 þ 3H2 O hydroxy-Al polymer

Log K ¼ 4:1

imogolite

In fact, soil solutions collected from horizons having both hydroxy-Al interlayered 2:1 layer silicates and imogolite show remarkably consistent pH4SiO4 values between 4.0 and 4.1 (Farmer, 1987; Ugolini et al., 1988; Dahlgren and Ugolini, 1989b; Dahlgren et al., 1991). This suggests that imogolite formation may be limited by the supply of H4SiO4 in these soils because Al for synthesis of imogolite would be available from dissolution of interlayer Al(OH)3. In humus-rich horizons, the Al3þ activities are signiﬁcantly lower than that predicted for hydroxy-Al polymers (Dahlgren and Ugolini, 1989b; Takahashi et al., 1995). These studies suggest that exchange of Al3þ with humic substances (Al –humus complexes) controls the relationship between Al3þ and pH. In this case, the degree of Al3þ saturation of carboxyl groups on humic substances will determine the pAl versus pH solubility relationship (Cronin et al., 1986). The kinetics of Al release from soils containing a variety of active Al forms shows that Al release rates are rapid from both allophanic and nonallophanic Andisols (Fig. 10; Dahlgren and Saigusa, 1994; Takahashi et al., 1995). To determine the source of Al released, soils were chemically treated with KCl, pyrophosphate and acid oxalate to extract various pools of active Al. Removal of KCl extractable Al resulted in a large decrease in Al release rates for nonallophanic Andisols, but an increase in release rates for allophanic Andisols. Decreased release rates observed in the nonallophanic Andisols are explained by the removal of a large pool of exchangeable Al. In contrast, the increased Al release rates observed in the allophanic Andisols are believed to be due to the formation of an easily dissolvable, polymeric-Al surface precipitate following the mechanism of “induced hydrolysis” outlined by Wada (1987a, b). This mechanism suggests that adsorbed forms of Al are released to solution in the exchange process with Kþ. In the case of the allophanic Andosols, the most probable source of exchangeable Al is from weakly held Al – humus complexes because these soils do not contain detectable levels of 2:1 layer silicates. Following the release, Al undergoes hydrolysis and releases Hþ that is adsorbed on the surface of variable charge materials. The increased ionic strength due to

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 149

Figure 10 Aluminum release rates at pH 3.3 for nontreated soil samples and residues following selective dissolution treatment for a Bw horizon from allophanic and nonallophanic Andosols. Samples were treated with KCl, pyrophosphate, and acid oxalate to remove various Al pools for determining their affect on the overall Al dissolution rates (Dahlgren and Saigusa, 1994).

the 1 M extracting solution drives the reaction by increasing surface charge on variable charge constituents resulting in a proton sink for the Hþ released by Al hydrolysis, thereby buffering solution pH. Saturation indices were near equilibrium to slightly supersaturated with respect to synthetic gibbsite indicating that the precipitation of displaced Al could have occurred. These precipitates are most probably adsorbed to the soil surfaces making them readily available for dissolution by acidic solutions. Removal of Al – humus complexes with pyrophosphate reagent resulted in a substantial decrease in Al release rates for both allophanic and nonallophanic Andosols. This suggests that Al –humus complexes are a labile source of dissolved

150

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Al, especially for the nonallophanic Andosols. Further treatment with acid oxalate had little effect on Al release from nonallophanic Andisols; however, Al release from allophanic Andisols decreased to very low rates. This suggests that allophane and imogolite are an important source of dissolved Al in allophanic Andosols.

VIII.

PRODUCTIVITY AND MANAGEMENT OF VOLCANIC SOILS

Volcanic soils are among the most productive soils in the world as suggested by their high human carrying capacity and high agronomic potential (Leamy, 1984; Shoji et al., 1993a). In areas surrounding Mt. Merapi volcano, central Java, Indonesia, the most fertile soils are juvenile volcanic soils (800 to more than 1000 people per km2), followed by the youngest ash deposits (more than 400 people per km2), and then areas without ash deposits (245 people per km2) (Mohr, 1938). In contrast, volcanic soils in Japan are sometimes regarded as one of the most infertile agricultural soils. Differences in the agricultural productivity among Andisols are largely attributed to the colloidal composition in the rooting zone, namely allophanic versus nonallophanic. While nonallophanic Andisols share many common properties with allophanic Andisols, they display some distinctive properties due to the presence of 2:1 layer silicates (Saigusa et al., 1991; Saigusa, 1992). In particular, the strong acidity, high exchangeable Al content, and low exchangeable base concentrations of nonallophanic Andisols lead to severe aluminum toxicity in sensitive crops (Saigusa et al., 1980; Shoji et al., 1980a, b). Therefore, from a soil management perspective, Andisols should be divided into allophanic and nonallophanic categories.

A. CHARGE CHARACTERISTICS

AND

CHEMICAL FERTILITY

The different charge characteristics between allophanic and nonallophanic Andisols is the most important factor regulating chemical fertility attributes (Table III) (Shoji et al., 1985; Saigusa, 1991; 1992). Allophanic Andisols are dominated by variable charge constituents (allophane/imogolite), while nonallophanic Andisols display both variable charge and constant charge components (2:1 layer silicates and Al –humus complexes). The CEC and AEC of variable charge components depend on pH and the ionic strength of the soil solution (Okamura and Wada, 1983). The variable positive charge results from protonation of surface Al– OH, Fe –O and Fe – OH groups, while the variable negative charge results from dissociation of surface Si-OH and organic functional groups (e.g., carboxylic). Charge characteristics of Andisols in northern Japan were fully investigated with special reference to their colloidal

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 151 Table III Comparison of Selected Soil Properties Between Allophanic Andisols and Nonallophanic Andisols Property

Allophanic Andisols

Nonallophanic Andisols

Major clays

Allophane and imogolite

Source of active Al

Allophane and Al–humus complexes None Some Weakly acid Low Rare High None-slightly None-slightly Slight

Hydroxy-Al interlayered 2:1 layer silicates Al –humus complexes

Constant negative charge Variable positive charge Soil acidity KCl-extractable Al Al-toxicity Consistency–LL, PL, PIa Stickiness Plasticity Compaction potential

High None Strongly acid High Common Low Moderately Moderately Moderate

a

LL is liquid limit, PL is plastic limit; PI is plasticity index.

composition (allophanic and nonallophanic Andisols, with or without organic matter enrichment) (Saigusa et al., 1992b). The development of negative charge with increasing soil pH was common to all Andisols and was strongly related to the amount of soil organic matter. Soils with a large variable charge component required large additions of lime for pH amendment, and were susceptible to leaching of cations when the soil pH decreased. Some constant negative charge was observed in nonallophanic Andisols with or without organic matter and contributed to aluminum toxicity due to the presence of exchangeable Al. Positive variable charge increased with decreasing pH in Andisols having low organic matter, while Andisols with higher organic matter concentrations showed virtually no positive charge, even when soil pH decreased. AEC values for Andisols were generally less than 3 cmolc kg21 at pH values greater than 5.

1.

Nitrogen Dynamics

Nitrogen accumulation and mineralization. Nitrogen is the major nutrient limiting establishment of plants in thick deposits of volcanic materials following an eruption. Plants are able to rapidly regenerate in locations where their roots can tap nitrogen pools in the buried soil or where erosion exposes the buried soil. The thick pyroclastic ﬂows surrounding Mt. St. Helens were colonized by nitrogen ﬁxing lupine (Lupinus spp.). Individual lupine plants created “islands of soil fertility” by ﬁxing nitrogen, incorporating organic matter into the soil, and

152

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

trapping nutrient-rich eolian materials. Alder (Alnus spp.), another common nitrogen ﬁxer and pioneer species, has also been shown to contribute large amounts of nitrogen to disturbed ecosystems (Ugolini, 1968). These ecosystem pioneering species play an important role in revegetation of terrestrial ecosystems through their role in ﬁxing nitrogen, enhancing chemical weathering and incorporating organic matter into the soil to initiate nutrient cycling. Nitrogen is generally the most important nutrient determining yield and quality in agricultural ecosystems. Andisols often accumulate large amounts of nutrients in their organic-rich A horizons. For example, a 50 cm thick A horizon has about 15 Mg N ha21 and 5.5 Mg P ha21 (Shoji et al., 1993a). While the N pool greatly exceeds the requirement of commonly cultivated crops (~200 kg ha21), it is the mineralization rate during the growing season that determines the amount of soil nitrogen available to plants. Under intensive agricultural production in Japanese Andisols, about 80 to 110 kg N ha21 was made available to crops through nitrogen mineralization (Saigusa et al., 1983; Shoji et al., 1986). Therefore, mineralized nitrogen plays an important role in supplying N to crops and may be a source of nitrogen leaching to subsoil and ground water. Incubation methods are commonly used to estimate potentially mineralizable N in Andisols (Sugihara et al., 1986; Fridrik et al., 1996; Parﬁtt et al., 2001). These studies ﬁnd that the amount of nitrogen released by mineralization is highly correlated with the amounts of soil organic matter, especially the fraction deﬁned as the easily decomposable organic matter pool (Okano, 1990; Sakamoto and Oba, 1993). The following three equations, representing ﬁrst-order kinetic reactions, were proposed for estimating nitrogen release patterns in Andisols under aerobic (simple model) and anaerobic conditions (two source model) (Sugihara et al., 1986): 1. Simple type model (for upland soils): N ¼ N0 [1 2 exp(2 kt)] 2. Two-source type model (for paddy soils): N ¼ N01 [1 2 exp(2 k1t)] þ N02 [1 2 exp(2 k2t)] 3. Simple type model combined with immobilization: N ¼ Nim [1 2 exp(2 kimt)] þ N0 [1 2 exp(2 kt)] þ C where, N is the amount of mineralized nitrogen (mg N kg21 soil); N0 is the mineralization potential; N01 and N02 are the nitrogen mineralization potentials for easily and slowly decomposable fractions, respectively (mg N kg21 soil); Nim is the amount of immobilized nitrogen (mg N kg21 soil); k, k1, and k2 are the mineralization rate constants; kim is the immobilization rate constant (day21) and t is time (days). Stanford and Smith’s (1972) original kinetic equations were modiﬁed into Equation (1) for practical use under ﬁeld conditions, Equation (2) for the existence of two different nitrogen sources (easily and slowly decomposable nitrogen

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 153

sources), and Equation (3) for the coexistence of both mineralization and immobilization. N mineralization parameters for Andisols and nonandic soils from northeastern Japan were determined by Saito (1990) using Equation (1). The nitrogen mineralization potential (N0) in Andisols was considerably greater than for nonandic soils (176 versus 100 mg N kg21 soil); however, mineralization rate constants (k) were smaller in Andisols than in nonandic soils (0.0043 versus 0.0059 d21) incubated under aerobic conditions. As a result, the percentage of mineralizable N (N0/Nt values) from Andisols was less than 50% that of nonandic soils (3.5 versus 8.2%). Retardation of N mineralization by allophane and Al3þ ions in Andisols was suggested in comparison with nonandic soils (Parﬁtt et al., 2001). The incubation approach was very effective in predicting nitrogen mineralization under ﬁeld conditions of corn cultivation (Saito and Ishi, 1987). A similar kinetic analysis conducted for Typic Hapludands showed that the k value of reduced tillage was greater than that of conventional tillage (Nira and Nishimune, 1993). In the two-source model (paddy ﬁeld), the N mineralization potential for the easily decomposable component (N01) was inﬂuenced strongly by soil drying before puddling. The easily decomposable pool was important for supplying soil N for rice during the early growing stages. On the other hand, N mineralization from the slowly decomposable component (N02) was the primary source of N during the middle- and late-growth stages. The easily decomposable pool was much smaller than the slowly decomposable pool in Andisols compared to alluvial soils (Ando and Shoji, 1986; Shoji et al., 1993a). Nitrogen leaching in Andisols. Several investigations on adsorption and leaching of nitrate in Andisols conclude that nitrate mobility is affected by soil pH, organic matter, anion exchange capacity, competing anions, and the point of zero charge (Reynolds-Vargas et al., 1994; Babbar and Zak, 1995; Kamewada, 1996; Katou et al., 1996; Qafoku et al., 2000; Vaje et al., 2000; Ryan et al., 2001). The development of positive charge in both allophanic and nonallophanic Andisols is depressed by the accumulation of organic matter in surface horizons (Saigusa et al., 1992b). In addition, the pH of surface horizons is generally increased by addition of lime to obtain a recommended pH of 6 – 7 for most crops. Consequently, the nitrate adsorption capacity in the Ap layer of Andisols is very low and thus NO2 3 is freely leached to the subsoil and eventually to ground water after heavy rain (Saigusa, 1988; 1991). Exchangeable NO2 3 concentrations reported for volcanic soils receiving heavy nitrogen fertilization can be as high as 1– 2 cmol kg21 (Reynolds-Vargas et al., 1994; Kamewada, 1996); however, values less than 0.1 cmol kg21 are more common for soils receiving moderate N fertilization. A critical factor regulating N-use efﬁciency between allophanic and nonallophanic Andisols is the depth of the rooting zone. In ﬁelds of winter-planted

154

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

barley and wheat, nitrogen applied as (NH4)2SO4 was easily converted to NO2 3 within two weeks in both allophanic and nonallophanic Andisols. The downward movement of NO2 3 was 16 – 19 cm per 100 mm of rainfall in autumn when evapotranspiration was low (Saigusa et al., 1983; 1991). Because nonallophanic Andisols have strongly acidic subsoil horizons, roots of Al sensitive crops, such as barley, are restricted to the topsoil layer that has received lime. In contrast, rooting depth in allophanic Andisols was not restricted by Al toxicity leading to a much greater depth of rooting. Consequently, N fertilizer recovery was less than 10% on nonallophanic Andisols compared to 30– 50% on allophanic Andisols. Barley grown on nonallophanic Andisols showed nitrogen deﬁciency symptoms and yields were less than half of that obtained in allophanic Andisols. Evaluation of nitrate sorption potential of soils at local and landscape scales is useful for nitrogen management to reduce ground water pollution (Ryan et al., 2001). In the Central Valley of Costa Rica, the fertilizer used for coffee plantation was suspected to be a cause of ground water nitrate pollution. Differences in the nitrate adsorption characteristics and leaching were related to clay mineralogy. Andisols dominated by Al-rich allophane (Al:Si, 2:1) showed signiﬁcant nitrate adsorption (about 15 mg N g21) and thus retarded nitrate leaching compared to negligible nitrate adsorption in Andisols containing Si-rich allophane (Al:Si, 1:1) 21 and halloysite. They estimated that about 300 kg NO2 was retained on 3 -N ha the AEC in the 4 m soil proﬁle containing Al-rich allophane. Signiﬁcant nitrate adsorption in allophanic Andisols retards leaching of nitrate to ground water by retaining NO2 3 within the rooting zone allowing a greater time for plant uptake. However, the effectiveness of NO2 3 sorption in coffee plantations is limited because of high inputs of nitrogen fertilizer and liming (i.e., reduction of positive charges) (Reynolds-Vargas et al., 1994). Improvement of fertilizer nitrogen-use efﬁciency in Andisols. Because of the low NO2 3 retention characteristics of Andisols, N-use efﬁciencies of readily soluble fertilizers (e.g., (NH4)2SO4) are considered to be low, especially in nonallophanic Andisols because of their strongly acidic subsoil. To improve the nitrogen-use efﬁciency of applied fertilizer nitrogen, several approaches were examined in nonallophanic Andisols: top-dressing versus split application of basal nitrogen fertilizers, use of controlled release nitrogen fertilizers, and band versus nested placement of fertilizer (Saigusa et al., 1991; Saigusa, 1999). Among these methods, a split application of basal rapidly available nitrogen fertilizer and both basal and top dressings of controlled release nitrogen fertilizers were veriﬁed to be the most effective in improving the productivity of barley in strongly acidic nonallophanic Andisols (Saigusa et al., 1983; 1991). The N-use efﬁciencies of readily available nitrogen applied as top dressing, especially that applied at later growing periods were much higher than those applied as basal fertilizer nitrogen because of the intensive root development in the surface soil

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 155

and increased nutrient absorbing ability with age of the crops. Thus split application of basal readily available nitrogen fertilizer increased the N uptake of winter barley by 28% and grain yields by 38% in nonallophanic Andisols (Saigusa et al., 1991). The nutrient release pattern of some controlled release fertilizers may be synchronized with the growth rates of the crops to enhance N-use efﬁciencies. For example, polyoleﬁn coated ureas can be formulated to release N over a speciﬁed time period: POCU-70 releases . 80% of its nitrogen within 70 days at 25 8C. POCU-70 applied as a single basal application to silage-corn resulted in 49 – 66% N-use efﬁciency of applied N compared to 30– 44% for (NH4)2SO4 (Gandeza et al., 1991; Shoji et al., 1991). Greatly enhanced N-use efﬁciencies were also reported for sorghum (Saigusa et al., 1990), oats and barley (Ito et al., 1998), and tender green mustard (Ombodi and Saigusa, 2000). Similarly, a single basal application of POCU-100 (sigmoid-type) applied directly to nursery boxes with rice seeds (co-situs application) improved N-use efﬁciency by three times while also reducing the labor cost associated with fertilization (Saigusa et al., 1996). Mixtures of rapid and slower release POCU fertilizers may be formulated to provide a pulse of nitrogen during the early growth stages and sustained nitrogen availability throughout the later growth stages of crops. Therefore, the use of controlled release fertilizers for crop production in Andisols is an advantageous method that saves labor and maintains nutrient availability to plants during their late-growth stages. Additionally, the increased N-use efﬁciency reduces environmental impacts associated with NO2 3 leaching to ground water and N2O emissions, a potent greenhouse gas.

2.

Phosphorus Dynamics

Phosphorus is often a growth-limiting nutrient for agricultural crops grown on Andisols (Alvarez et al., 1999; Manske et al., 2000). Therefore, heavy application of phosphorus fertilizer to Andisols is commonly practiced in developed countries to maximize yields. Phosphorus availability decreases with increasing soil development due to incorporation of P into organic forms and P ﬁxation by active Al and Fe components (e.g., allophane, imogolite, Al –humus complexes). In contrast, P availability in young volcanic materials is relatively high due to rapid weathering of apatite, and low phosphate retention due to low concentrations of active Al and Fe. Fresh volcanic ash contains an appreciable amount of P extractable by dilute acid, Troug or Bray methods. Troug-extractable P in rhyolite and andesite tephras (0.1 – 0.6 g P2O5 kg21) was much higher compared to basaltic andesite and basalt tephras (, 0.1 g P2O5 kg21) (Nanzyo et al., 1997). The source of extractable P was attributed to apatite, the primary phosphorus-bearing mineral in volcanic

156

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

materials. Dissolution rates of apatite increased with decreasing pH and increasing concentrations of chelates (Nanzyo et al., 1999; Nakamaru et al., 2000a). Although, the content of apatite is usually less than 0.3% in fresh volcanic materials, it can provide an important source of P in natural ecosystems and agricultural production. For example, pigeon pea and chick pea were grown in fresh Mt. Pinatubo, Philippines volcanic ash containing 1.7 g P2O5 kg21, primarily as apatite. Dry matter production of pigeon pea was similar in plots with and without added P, and both pigeon pea and chick pea enhanced P availability through production of organic acids and chelates (Nakamaru et al., 2000b). While mature Andisols commonly have large amounts of active Al, the form of active Al is different between allophanic and nonallophanic Andisols (Table III). The major components responsible for phosphorus sorption in allophanic Andisols are allophane/imogolite and Al– humus, while Al –humus complexes are the dominant components in nonallophanic Andisols. The amount of phosphate adsorbed per mole of active Al was higher in nonallophanic Andisols (0.13 mol) compared to the allophanic Andisols (0.09 mol) (Saigusa et al., 1991; Matsuyama et al., 1999a). As a result, P availability is generally higher for allophanic Andisols. In fact, the amount of P absorbed by barley, buckwheat, sorghum, and dent corn was larger in allophanic than in nonallophanic Andisols, even though their active Al and Fe concentrations were similar (Mastuyama et al., 1994; Ito et al., 2000a). Phosphorus fertilizer applied to Andisols is rapidly sorbed to active Al and Fe components and changes to less available forms with time (Nanzyo, 1987). Many soil tests for estimating available soil P have been evaluated for volcanic soils. The Truog method is most popular in Japan, followed by Bray II, Mehlich No. 3, Olsen, and anion exchange resin methods. Each method has proved useful for estimating the amount of plant-available P, but only when comparisons are made among soils with similar colloidal composition (i.e., allophanic versus nonallophanic). Therefore, it is difﬁcult to compare results between allophanic and nonallophanic Andisols. Due to the large accumulation of organic matter in many volcanic soils, organically bound P is often a large pool (about 5.5 Mg ha21 in Andisols with thick A horizons; Shoji et al., 1993a). Organic P in uncultivated Andisols in Chile was strongly correlated with organic carbon content (Org. P [mg kg21] ¼ 59 organic C[%] þ 133; r 2 ¼ 0.83) (Escudey et al., 2001). Organic P is considered an important P source for crops through mineralization. However, measuring P mineralization rates by simple incubation methods is difﬁcult in Andisols because PO4 released by mineralization is rapidly sorbed by active Al and Fe. Adding anion exchange resins or membranes to incubating soil was shown to effectively capture P released from mineralization, preventing it from being sorbed to active Al and Fe components (Zou et al., 1992; Parﬁtt et al., 1994). P mineralization studies using anion exchange methods showed rates of 1.3 mg P kg21 d21 for

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 157

Hawaiian forest Andisols (Zou et al., 1992) and 0.2 –0.5 mg P kg21 d21 for Japanese Andisols (Ito et al., 2000b). These results conﬁrm that P mineralization is an important P source for plants growing in Andisols. Improvement of P uptake and plant growth by arbuscular mycorrhizae (AM) under conditions of low available phosphorus has been well documented. Improved P uptake and plant growth were demonstrated in Andisols following AM inoculation of onion (Botello et al., 1987; Palacios et al., 1987), onion and white clover (Tawaraya et al., 1995), and kidney bean (Isobe and Tsuboki, 1997). Inoculation increased both spore number and the grain weight of kidney bean at a soil-available P (Bray No. 2) level of 10 mg P kg21, while these effects were suppressed when soil-available P was 25 mg P kg21. Indigenous AM fungi predominate over inoculated species in most soils. The effectiveness of using indigenous AM fungi was investigated in several cropping systems (Arihara et al., 2000; Karasawa et al., 2000a). Soils contained more AM spores following cultivation of mycorrhizal crops (e.g., sunﬂower, maize, soybean, potato kidney bean, adzuki bean, and wheat) than after cultivation of nonmycorrhizal crops (mustard, radish, sugar beet and buckwheat). As a result there was increased growth of succeeding crops following mycorrhizal crops. This effect on maize growth and phosphorus uptake was evident in dry soil conditions and was less pronounced with increasing soil moisture (Karasawa et al., 2000b).

3.

Other Mineral Nutrients

Generally, volcanic ash soils are formed by intermittent tephra deposition and thus surface soils are often regenerated by addition of fresh volcanic ash. This means that Andisols are relatively young soils and the soil fertility is closely related to the properties of the parent tephra. The chemical composition of tephra varies depending on the chemical composition of the magma. There is generally a close relationship between the SiO2 content and concentrations of other elements (Shoji et al., 1975; Kobayashi et al., 1976; Yamada, 1988). In general, Al, Fe, Mg, Ca, Mn, Ti, Cu, and Co are most abundant in maﬁc (basic) tephras, whereas Na and K tend to be most abundant in felsic (acidic) tephras. Greenhouse studies examining fertility of Andisols formed from different tephra chemistries showed that the amounts of N, Ca, Mg, and Cu absorbed by plants were greater in maﬁc than felsic materials, and only K uptake was greater in felsic materials (Saigusa et al., 1976; Yamada and Shoji, 1980). Thus, the nutrient supplying power of elements, except K, was stronger in the Andisols derived from maﬁc materials. Intermittent deposition of tephra on soils has been shown to affect soil fertility (Dahlgren and Ugolini, 1989a; Cronin et al., 1997; 1998; Dahlgren et al., 1999). Under cool – humid weathering conditions in the Cascade Mountains, Washington, a 5 cm layer of fresh Mt. St. Helens tephra (1980 eruption) applied to the surface

158

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

of a Spodosol released appreciable concentrations of nutrient cations and sulfur over a 4 year period of investigation (Ca ¼ 217; Mg ¼ 31; K ¼ 50 and S ¼ 69 kg ha21 over 4 years; Dahlgren et al., 1999). Release rates showed a decrease with time as soluble salts and easily weatherable materials were depleted (Fig. 11). The base cations were preferentially retained on cation-exchange sites in the buried organic horizons as the HCO3 anion was lost to protonation (HCO2 3 þ Hþ ¼ H2CO3) (Fig. 8). Sulfur (as sulfate) was retained by sorption within B horizons (Fig. 12). The high anion sorption by noncrystalline materials in the B horizons resulted in virtually no enhanced losses of sulfate from the C horizons. Thus, release of nutrients from intermittent airfall tephra additions provides an important source of slowly available nutrients that helps to sustain forest productivity in these strongly acidic soils. The large pulse of base cations from the leaching of soluble salts contained in the fresh tephra resulted in displacement of high concentrations of Hþ and Al3þ that may result in short-term Al toxicity. The high Hþ and Al3þ concentrations returned to ambient levels within about 6 months of the tephra addition. Similarly, eruptions from Mt. Ruapehu, New Zealand in 1995 –1996 added between 30 and 1500 kg S ha21 (Cronin et al., 1997; 1998). As little as 0.25 mm of ash caused pale yellow S-deﬁcient ryegrass to “green” within two days of tephra application. Smaller, but agronomically signiﬁcant amounts of K, Mg, Na, Se, B, and Co were also leached from the ash deposits. Ash addition caused a short-term depression (, 50 d) of soil pH (, 4.5) due to the oxidation of elemental sulfur, which represented 55 –77% of the total sulfur content

Figure 11 Nutrient release rates for a 5 cm ash layer for the ﬁrst four years following addition of ash to the surface of a Spodosol in the western Cascade Range, Washington (Dahlgren et al., 1999).

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 159

Figure 12 Soil solution charge balance for soils receiving 0, 5, or 15 cm of Mt. St. Helens ash. Soil solutions represent the ﬁrst month of leaching of the fresh ash (September) after application to the surface of a Spodosol in the western Cascades Washington. The width of each compartment is in proportion to the ion’s charge contribution. The anion deﬁcit was assumed to be the contribution of dissociated organic ligands (Dahlgren and Ugolini, 1989a).

(Cronin et al., 1998). With respect to animal health, the most common cause of livestock poisoning results from ﬂuorosis (Oskarsson, 1980; Arya et al., 1990). Ingestion of tephra particles and/or forage with high ﬂuoride levels may result in acute or chronic ﬂuorosis. Concentrations of F, S, and Se from Mt. Ruapehu were high enough to reach potentially harmful levels within water and in the foliage of plants that accumulate these elements (Cronin et al., 1998).

160

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

4.

Silicon Nutrition

Silicon is generally not considered one of the essential elements for higher plants, but is a beneﬁcial element for certain plants such as rice, sugarcane, cucumber, and turf. In rice production, Si is considered an “agronomically essential element” because of its multiple functions in rice growth, such as improving the resistance to lodging, pest and disease, reducing excessive transpiration, and keeping leaves erect. Through these functions, silicon enhances water-use efﬁciency and the photosynthesis rate (Savant et al., 1997; Saigusa et al., 1998; 1999; Saigusa, 2000). The silicon content of sugarcane is strongly correlated with soluble Si concentrations (Fox et al., 1967), which in turn is strongly related to soil mineralogy. Silicon is the most abundant element in fresh tephra ranging from about 48 to 73% SiO2; it is highest in felsic ash and lowest in basaltic ash (Shoji et al., 1975). Generally, volcanic soils have a high potential for supplying Si when they are young. However, Si concentrations in Andisols decrease as weathering proceeds in the udic soil moisture regime because volcanic glass is consumed by weathering and Si is easily leached due to its high relative mobility (i.e., compared to Al and Fe). Silicon is incorporated in allophane/imogolite, opaline silica, and halloysite, and is easily adsorbed by noncrystalline materials in Andisols. As a result, the amount of Si available for some Si accumulating plants is not sufﬁcient. Therefore, the application of silicate fertilizer is a common practice to maintain the silicon supply in paddy ﬁelds, especially Andisols and peat soils (Saigusa et al., 1998; 1999; Saigusa, 2000).

5.

Soil Productivity and Acidity

Aluminum toxicity in Andisols. In acidic soils, Hþ, Mn2þ, and Al toxicities, P, Ca, Mg and micronutrient deﬁciencies, and suppression of some microorganisms have been shown to affect crop productivity. Among these factors, Al toxicity is the most important constraint on crop growth in acidic Andisols. Among Andisols, a signiﬁcant difference in Al toxicity potential was observed between allophanic and nonallophanic soils. Severe Al toxicity was observed in nonallophanic Andisols containing high amounts of KCl-extractable Al derived primarily from 2:1 layer silicates (strongly acidic constant charge), but not in allophanic Andisols whose major clay mineral was allophane (weakly acidic variable charge) (Saigusa et al., 1980; Shoji et al., 1980a, b). The primary manifestation of Al toxicity is impaired root development resulting in the formation of a shallow rooting system. The inhibition of root growth in Andisols is closely related to the amount of KCl-extractable Al rather than soil pH(H2O). Therefore, KCl-extractable Al is the most reliable measure of Al toxicity

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 161

potential in Andisols. Threshold values for Al toxicity were 2 cmolc kg21 for common agricultural crops and 1 cmolc kg21 for Al-sensitive crops. A value of 2 cmolc kg21 of KCl-extractable Al was adopted for the allic subgroup of Andisols to reﬂect their potential for Al toxicity (Soil Survey Staff, 1998). The origin of KCl-extractable Al has been attributed primarily to Al3þ adsorbed to 2:1 minerals, and to a lesser extent release from Al –humus complexes (Dahlgren and Saigusa, 1994; Takahashi et al., 1995). Allophanic Andisols generally contain very low KCl-extractable Al concentrations; however, these values may be underestimated due to “induced hydrolysis” of displaced Al and subsequent adsorption of polymeric Al to allophanic materials (Wada, 1987a, b). Plant nutritionists generally explain plant uptake of elements by active absorption from soil solution rather than by contact exchange. Therefore, it has been difﬁcult to explain Al toxicity in soils with pH . 4.5, where low concentrations of monomeric Al3þ should not contribute to Al toxicity. At pH . 4.5, Al3þ becomes polymerized. Polymerized Al was shown to be extremely toxic to higher plants compared to monomeric Al3þ in solution culture. Consequently, many plant nutritionists believed that the cause of Al toxicity at pH . 4.5 was attributably to polymerized Al, such as the Al13 polymer [AlO4Al12 (OH)24 (H2O)12]7þ (Bartlett and Riego, 1972; Wagatsuma and Ezoe, 1985; Shann and Bertsch, 1993). The presence of Al13 polymer was suggested in Spodosols with extremely low pH (pH ¼ 3.5) (Hunter and Ross, 1991) and in Andisols based on Al speciation calculations (Kato et al., 1995). A detailed study of phytotoxic forms of Al in an Andisol examined both root growth of rice plants under submerged soil culture and monomeric and polymeric sorption reactions to soil materials. Polymeric Al was selectively and irreversibly adsorbed by soil colloids and thus monomer Al3þ was concluded to be the main cause of Al toxicity (Saigusa et al., 1995). Aluminum speciation of KCl extracts from a variety of acidic soils using 27Al nuclear magnetic resonance (NMR) did not detected the presence of Al13 polymers (Hiradate et al., 1998). These results suggest that Al13 polymers were not formed in these soils, or alternatively Al13 polymers were removed from solution by adsorption and/or precipitation reactions. As a result, most of the recent work on the mechanism of Al toxicity and tolerance in higher plants, and the creation of Al tolerant plants by gene manipulation have focused on monomeric Al3þ. Allophanic Andisols generally have only trace levels of KCl-extractable Al and pH(H2O) values near 5. Aluminumphilic plants, such as tea, grow best under extremely acidic soil conditions. Therefore, farmers tended to apply huge amounts of nitrogen fertilizer (. 1500 kg N ha21 added as (NH4)2SO4) to maintain high quality of tea leaves (Okada et al., 1986; Tachibana et al., 1995; Pansombat et al., 1997). Similar acidiﬁcation of allophanic Andisols was observed in mulberry garden (Inamatsu et al., 1991) and some upland soils in Japan (Matsuyama et al., 1999b). Under heavy application of chemical fertilizer, allophanic Andisols become acidiﬁed to pH(H2O) , 4.0 and large amounts of KCl-extractable Al

162

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

develop. The extreme acidity breaks down allophane and contributes monomeric Al for uptake by tea plants (Pansombat et al., 1997). The origin of KCl-extractable Al in highly acidiﬁed allophanic Andisols was attributed to the dissolution of aluminum-sulfate compounds and/or Al –humus complexes. Subsoil acidity and acidity amelioration. Subsoil horizons play an important role in supplying both water and nutrients to crops. If subsoil acidity is severe enough to limit root development, as in many nonallophanic Andisols, crops cannot use water and nutrients stored in the subsoil and thus water and nutrient stress results in decreased yield and quality of crops (Saigusa et al., 1995; Shoji et al., 1986). Both yields and nitrogen uptake of Al-sensitive crops such as barley, sorghum, and alfalfa were remarkably reduced in nonallophanic Andisols with strongly acidic subsoils. In addition, nitrogen-use efﬁciency was decreased and nitrate leaching was increased due to the inability of the shallow rooted crops to extract nitrogen from subsoil horizons. The amount of KCl-extractable Al (phytotoxic Al) decreased with increasing soil pH(H2O), to less than 1 cmolc kg21 at pH(H2O) ¼ 6. Therefore, liming of the plow layer is commonly conducted using various materials, such as calcitic limestone and dolomite (Simono, 1990; Borie and Rubio, 1999; Mora et al., 1999a; 1999b; Toda et al., 1999). However, liming of acidic subsoil horizons is uncommon due to the difﬁculty and cost of delivering lime materials to subsoil horizons. Given the importance of subsoil horizons to crop production, amelioration of subsoil acidity in nonallophanic Andisols has been explored through surface application of gypsum, phosphogypsum, and organic calcium salts (Saigusa et al., 1994; Toma and Saigusa, 1997; Inoue et al., 2001). Surface application of gypsum was much more effective at ameliorating subsoil Al toxicity compared to application of CaCl2 or CaCO3, and this effect was especially prominent in subsoil horizons with low humus content. The mechanism responsible for reduction of exchangeable Al by gypsum additions was explained as follow: (1) release of exchangeable Al into soil solution from cation-exchange sites, (2) polymerization of monomeric Al in soil solution, and (3) selective and irreversible adsorption of polymerized Al on 2:1 clay minerals (Saigusa and Toma, 1997). The form of polymeric Al was not identiﬁed, but 27 Al-NMR showed that it was not the Al13 polymer (Toma et al., 1999). Suppression of soil borne plant pathogens by aluminum. Root rot of kidney beans (Phaseolus vurgaris L ) caused by Fusarium solani f. sp. phaseoli and potato scab caused by Streptomyces scabies have been devastating in Hokkaido, Japan where allophanic Andisols are widely distributed (Mizuno and Yoshida, 1993; Furuya et al., 1996). Suppression of these soil borne plant pathogens is reported in soils whose clay mineralogy is dominated by hydroxy-Al interlayered 2:1 minerals (i.e., nonallophanic Andisols). In the case of root rot of kidney bean, inhibition of macroconidal germination and disease incidence was observed in soils with

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 163

exchangeable aluminum contents of at least 0.4 cmol kg21 (Furuya et al., 1999). Common potato scab was suppressed in soils having relatively high amounts of exchangeable Al, or soluble Al concentrations of 0.3 mg kg21. As a result, a single basal band application of ammonium sulfate was recommended to acidify soils and suppress disease incidence in allophanic Andisols (Mizuno and Yoshida, 1993; Mizuno et al., 1998b; 2000). Rice seedlings are highly tolerant to Al toxicity and have been grown in soils having a pH range of 4.5 –5.5 to control damping-off disease and sudden withering of young seedlings (physiological disease) (Saigusa et al., 1992a). These examples indicate that the disease incidence of some soil borne plant pathogens is closely related to the clay mineralogy of Andisols, which regulates the contents of exchangeable and water soluble Al (Mizuno et al., 1998a; Furuya et al., 1999).

B. PHYSICAL PROPERTIES

AND

FERTILITY

Compared to chemical properties, Andisols generally display excellent physical properties for crop production. They typically have high water retention, high porosity, high permeability, high friability, low bulk density, and stable aggregates. These characteristics were originally attributed to the presence of allophane; however, the discovery of nonallophanic Andisols forced a reexamination of physical properties in Andisols with special reference to their colloidal composition. Consequently, the unique physical properties common to Andisols are attributed to the abundance of noncrystalline materials and/or soil organic matter. The favorable physical properties of Andisols contribute to a high rate of seedling emergence, excellent rooting characteristics, and strong resistance to drought stress. The fundamental physical properties of Andisols relating to soil productivity are well documented by Shoji et al. (1993a). Reviews concerning the physical and engineering properties of Andisols have been published by Yamanaka (1964), Maeda et al. (1977), Warkentin and Maeda (1980), Wada (1985), Maeda and Soma (1986), Warkentin et al. (1988) and Kimble et al. (2000).

1.

Soil Texture and Structure

Soil texture is an important index to describe soil physical conditions. However, particle-size analysis of volcanic soils is challenging due to poor colloidal dispersion resulting from the presence of extremely stable aggregates. Consequently, the clay content determined by standard methods is generally much smaller than those estimated by ﬁeld methods (Ping et al., 1988). Soil Taxonomy (Soil Survey Staff, 1998) uses 1500 kPa water retention to determine

164

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

particle-size class modiﬁers rather than mechanical analysis. There is a near linear relationship between the clay content determined by mechanical analysis with careful pretreatment for dispersion (ultra-sonication and pH adjustment) and their 1500 kPa water retention values (Shoji et al., 1993a). Generally speaking, Andisols have high physical fertility (tilth), such as high air and water retention, good drainage, and ease of root penetration. These physical properties are closely related to the development of soil structure. Soil structure in Andisols is strongly affected by soil organic matter and noncrystalline materials, which are generally abundant in Andisols (Wada, 1985). There is a strong correlation between dithionite –citrate –bicarbonate extractable Al and the degree of macroaggregation suggesting that active Al is important in the formation of water-stable macroaggregates (Ibrahim et al., 1998). The surface soils of uncultivated Andisols generally have moderate to strong granular structure. However, cultivation tends to degrade granular aggregates and promotes the formation of subangular blocky or angular blocky structure. This conversion is due to both compaction by machinery and reduction of soil organic matter. Subsurface horizons of Andisols tend to have mostly subangular and angular blocky structure (Shoji et al., 1993a). In contrast, young Andisols generally do not have any soil structure (i.e., single grained) due to low concentrations of noncrystalline materials and organic matter to bind particles.

2.

Bulk Density and Porosity

One of the most outstanding characteristics of Andisols is extremely low bulk density (, 0.9 Mg m23). Low bulk density is attributable to the large amounts of noncrystalline minerals and organic matter. Thus, bulk density of Andisols decreases with increasing concentrations of active Al and Fe. Andisols showing Alo þ 1/2Feo . 2% generally have bulk density values ranging from 0.2 to 0.9 Mg m23 (Shoji et al., 1996). Bulk density values decreased to , 0.9 Mg m23 when allophane concentrations were greater than about 5% (Shoji et al., 1993a). Organic matter also contributes to the low soil bulk density, and the bulk density of nonallophanic Andisols decreases with increasing organic carbon content. Bulk densities , 0.9 Mg m23 generally occur when organic C concentrations exceed approximately 3%. The low bulk density of Andisols mainly reﬂects the development of porous soil structure, while relatively low particle density of humus (1.4 –1.8 Mg m23) and volcanic glass (2.4 Mg m23) further contribute (Wada, 1985). The large pore space of allophanic Andisols is partially attributed to the intra- and interparticle micropores of allophane, while that of nonallophanic Andisols is primarily attributable to humus. The aggregate size of allophane clusters in perhumid Costa Rica Andisols was studied by laser diffraction grain-sizing and was between 2

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 165

and 20 mm (Buurman et al., 1997). The size of allophane aggregates increases with increasing allophane content. The positive charges displayed by many noncrystalline materials are believed to play an important role in aggregate formation (Wada, 1985). Macropores contain gravitational water or air, and the plant root distribution corresponds strongly with the distribution of the macropores (Hatano et al., 1988). These physical characteristics lead to soils that are very friable when moist resulting in ease of root penetration. While Andisols are considered relatively resilient to compaction, the pore structure can be broken down and bulk density increased by excessive compaction associated with crop production (Ito et al., 1997a; Dorel et al., 2000), forest harvest practices (Allbrook, 1986; Lenhard, 1986; Meurisse, 1988; Geist et al., 1989; Cullen et al., 1991), and grazing (Hofstede, 1995).

3.

Water Retention and Plant-Available Water

Andisols generally have high porosity and thus can retain large amounts of water. Micropores in noncrystalline materials and humus hold hygroscopic water, mesopores retain capillary water, and macropores contain gravitational water or air. The high water holding capacity of Andisols is attributed to large volumes of both mesopores and micropores. The 1500 kPa water content of allophanic Andisols was related to both the humus and noncrystalline material (N.M.) concentrations (Ito et al., 1991): 1500 kPa water ð%Þ ¼ 1:93 £ humus ð%Þ £ 1:47 £ N:M: ð%Þ 2 0:92 ðr ¼ 0:944; n ¼ 37Þ: Plants can absorb the water retained in mesopores and available water content is generally expressed as either 10 kPa or 33 kPa water content minus 1500 kPa water content. A matric potential of about 10 kPa is often considered as ﬁeld capacity because the natural moisture content of many Andisols is equivalent to this matric potential. Plant-available water contents can be effectively estimated from the 33 kPa water content (Saigusa et al., 1987): Surface soil: Available H2 O ð%Þ ¼ 0:586 £ 33 kPa water content ð%Þ þ 1:66 ðr ¼ 0:902; n ¼ 28Þ Subsoil: Available H2 O ð%Þ ¼ 0:486 £ 33 kPa water content ð%Þ þ 0:85 ðr ¼ 0:890; n ¼ 32Þ Thus, it may be concluded that Andisols retain large amounts of plantavailable water, about 60 and 50% of the 33 kPa water in surface and subsoil horizons, respectively.

166

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Following the 1980 eruption of Mt. St. Helens, wheat ﬁelds in eastern Washington (USA) recorded a bumper crop after deposition of a few centimeters of tephra that acted as a mulch by killing weeds and conserving moisture. Similarly, forest productivity throughout the semi-arid western United States is enhanced by surﬁcial deposits of volcanic ash that act as a mulch and weather to soils with high plant-available water retention (Meurisse, 1988). In areas receiving intermittent additions of pyroclastic materials, contrasting particle sizes (e.g., ﬁne ash over coarse pumice) within the proﬁle can result in textural discontinuities that impede water transmission. This condition often results in imperfect drainage that may limit root growth into subsoil horizons. Mechanical mixing of these layers by deep plowing is often employed to improve water movement and enhance productivity.

C. SOIL MANAGEMENT

AND CONSERVATION OF SOILS

VOLCANIC

Mature Andisols are typically considered to be resistant to water erosion because of their high inﬁltration rates and stable soil aggregates (Shoji et al., 1993a). However, immature soils developed on recent pyroclastic materials do not have stable aggregates and are highly susceptible to water erosion. Even welldeveloped Andisols are highly susceptible to both wind and water erosion when surface vegetation is removed because of their low bulk density (Kimble et al., 2000). Due to the serious consequences of soil degradation from erosion, several investigations have examined methods to suppress erosion in Andisols (Perret et al., 1996; Perret and Dorel, 1999; Poudel et al., 1999; Poulenard et al., 2001). Runoff and soil erosion on Andisols of northern Ecuador were investigated using rainfall simulations with special reference to soil maturity and land use. Noncultivated, mature Andisols showed high inﬁltration rates and low sediment loss, whereas cultivated, mature Andisols, Andisols after burning, and young Andisols were prone to crusting, surface runoff and erosion (Poulenard et al., 2001). Compared to conventional tillage practices, contouring, strip cropping, and contour hedgerow cultivation were effective in suppressing erosion under intensive farming systems on sloping topography in the Philippine (Poudel et al., 1999). Agro-ecological practices, such as maintaining plant cover and hedging along ﬁeld boundaries, were effective for soil rehabilitation and erosion control on Andisols in Reunion Island, Indian Ocean (Perret et al., 1996; Perret and Dorel, 1999). Thus, with proper soil conservation practices, it is possible to minimize soil erosion in Andisols under intensive cropping systems. No-tillage or reduced tillage practices are highly recommended to maintain surface vegetation or crop residues to suppress soil erosion in Andisols. In addition to erosion control, these cropping systems have many beneﬁcial aspects that contribute to environmental protection and sustainable agriculture:

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 167

† Suppressed soil erosion by maintaining surface cover (vegetation and residues). † Decreased cost for tillage, energy savings, and reduced CO2 emission. † Increased water holding capacity and inﬁltration, contributing to improved germination and drought resistance (Tsuji et al., 2000). † Recycles plant biomass, which enhances nutrients for future crops. † Accumulates organic matter in surface horizons, which maintains a stable temperature (mulching effect) and reduces bird damage and weed infestations. † Increased lodging resistance by improved rooting, and thus reducing harvest loss and improving crop yield and quality. † Maintains a ﬁrm soil surface, and thus allows timely ﬁeld operations, such as transplanting or harvesting even after rain. † Improves oxidative conditions (i.e., gas exchange) in paddy soils, and thus reduces hydrogen sulﬁde damage to rice roots. Some negative attributes of no-tillage and reduced tillage practices on Andisols are increased herbicide requirements, soil compaction, and decreased fertilizer-use efﬁciency. The most serious problem arises from low fertilizer-use efﬁciency, especially in paddy soils. Controlled release fertilizers with nutrient release patterns synchronized to crop nutrient demands resulted in greatly improved fertilizer-use efﬁciency for no-tillage cultivation of silage-corn and rice in Andisols (Saigusa et al., 1996; Ito et al., 1997b; Inoue et al., 2000). After ﬁve successive years of silage-corn production, Ito et al. (1997b) concluded that a notillage system with a single basal application of polyoleﬁn coated urea was highly effective in terms of fertilizer-use efﬁciency and yield optimization in nonallophanic Andisols. Similarly, no-tillage transplanting of rice in nonallophanic Andisols with a single basal application of polyoleﬁn-coated urea directly in the nursery box resulted in 80% nitrogen-use efﬁciency (Saigusa et al., 1996). This compares to less than 30% nitrogen-use efﬁciency using basal ammonium sulfate applied in a conventional tillage transplanting system. The recycling of rice straw under no-till greatly suppressed CH4 emissions from paddy soils and did not result in any reductive damage of rice (Saigusa et al., 1996). Therefore, no-tillage systems in combination with a single basal application of controlled release fertilizer in Andisols are very promising management strategies for both rice and upland crops.

REFERENCES Allbrook, R. F. (1986). Effect of skid trail compaction on a volcanic soils in central Oregon, USA. Soil Sci. Soc. Am. J. 50, 1344–1346. Alvarez, S. E., Etchevers, J. D., Ortiz, J., Nunez, R., Volke, V., Tijerina, L., and Martinez, A. (1999). Biomass production and phosphorus accumulation of potato as affected by phosphorus nutrition. J. Plant. Nutr. 22, 205 –217.

168

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Ando, H., and Shoji, S. (1986). Kinetic study on soil nitrogen mineralization and nitrogen immobilization of paddy soils. Jpn. J. Soil Sci Plant Nutr. 57, 1–7. Aomine, S., and Wada, K. (1962). Differential weathering of volcanic ash and pumice, resulting in formation of hydrated halloysite. Am. Miner. 47, 1024–1048. Arihara, J., and Karasawa, T. (2000). Effects of previous crops on arbuscular mycorrhizal formation and growth of succeeding maize. Soil Sci. Plant Nutr. 46, 43–51. Arnalds, O., and Kimble, J. (2001). Andisols of deserts in Iceland. Soil Sci. Soc. Am. J. 65, 1778–1786. Arya, O., Wittwer, F., Villa, A., and Ducon, C. (1990). Bovine ﬂuorosis following volcanic activity in the Southern Andes. Vet. Rec. 26, 641– 642. Babbar, L. I., and Zak, D. R. (1995). Nitrogen loss from coffee agroecosystems in Costa Rica — leaching and denitriﬁcation in the presence and absence of shade trees. J. Environ. Qual. 24, 227–233. Bailey, S. W. (1990). Halloysite — a critical assessment. In “Proceedings 9th Int. Clay Conf., Strasbourg, France 1989” (V. C. Farmer and Y. Tardy, Eds.), pp. 89 –98. Sciences Ge´ologiques, Me´moires No. 86. Baldock, J. A., and Nelson, P. N. (2000). Soil organic matter. In “Handbook of Soil Science” (M. E. Sumner, Ed.), pp. B25–84. CRC Press, Boca Raton, FL. Bartlett, R. J., and Riego, D. C. (1972). Toxicity of hydroxy aluminum in relation to pH and phosphorous. Soil Sci. 114, 194–200. Ba¨umler, R., and Zech, W. (1994). Characterization of Andisols developed from nonvolcanic material in eastern Nepal. Soil Sci. 158, 211–217. Bautista-Tulin, A. T., and Inoue, K. (1997). Hydroxy-interlayered minerals in Japanese soils inﬂuenced by eolian deposition. Soil Sci. Soc. Am. J. 61, 631–640. Beinroth, R. H. (1982). Some highly weathered soils of Puerto Rico. 1. Morphology, formation and classiﬁcation. Geoderma. 27, 1–73. Blong, R. J. (1984). “Volcanic Hazards: A Sourcebook on the Effects of Eruptions”, p. 424. Academic Press, Orlando, FL. Borie, F., and Rubio, R. (1999). Effects of arbuscular mycorrhizae and liming on growth and mineral acquisition of aluminum-tolerant and aluminum-sensitive barley cultivars. J. Plant Nutr. 22, 121–137. Botello, G. J., and Ferrera, C. R. (1987). Effect of VA mycorrhiza on onion allium strain in a Mexican Andosol. Rev. Latinoam. Microbiol. 29, 97–102. Boudot, J. P., Bel Hadi Brahim, A., and Chone, T. (1988). Dependence of carbon and nitrogen mineralization rates upon amorphous metallic constituents and allophanes in highland soils. Geoderma. 42, 245–260. Boudot, J. P., Bel Hadi Brahim, A., Steiman, R., and Seigle-Murandi, F. (1989). Biodegradation of synthetic organo-metallic complexes of iron and aluminum with selected metal to carbon ratios. Soil Biol Biochem. 21, 961– 966. Brahim, B. H. (1987). “Inﬂuence des constituants alumineux et ferriques non cristallins sur les cycles du carbone et de l’azote dans les sols montegands acids”, p. 99, These Universite de Nancy I. Buurman, P., deBoer, K., and Pape, T. (1997). Laser diffraction grain-size characteristics of Andisols in perhumid Costa Rica: the aggregate size of allophane. Geoderma. 78, 71–91. Caner, L., Bourgeon, G., Toutain, F., and Herbillon, A. J. (2000). Characteristics of non-allophanic Andisols derived from low-activity clay regoliths in the Nilgiri Hills (Southern India). Eur. J. Soil Sci. 51, 553–563. Casadevall, T., Rose, W., Gerlach, T., Greenland, L. P., Ewert, J., Wunderman, R., and Symonds, R. (1983). Gas emissions and the eruptions of Mount St. Helens through 1982. Science. 221, 1383–1385. Chadwick, O. A., Olson, C. G., Hendricks, D. M., Kelly, E. F., and Gavenda, R. T. (1994). Quantifying climatic effects on mineral weathering and neoformation in Hawaii “Transactions of the 15th World Congress of Soil Science, Acapulco, Mexico, 1994”, pp. 94–105.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 169 Chen, Z. S., Asio, V. B., and Yi, D. F. (1999). Characteristics and genesis of volcanic soils along a toposequence under a subtropical climate in Taiwan. Soil Sci. 164, 510 –525. Chen, Z. S., Tsou, T. C., Asio, V. B., and Tsai, C. C. (2001). Genesis of Inceptisols on a volcanic landscape in Taiwan. Soil Sci. 166, 255–266. Childs, C. W., Parﬁtt, R. L., and Newman, R. H. (1990). Structural studies of Silica Springs allophane. Clay Min. 25, 329 –341. Childs, C. W., Matsue, N., and Yoshinaga, N. (1991). Ferrihydrite in volcanic ash soils of Japan. Soil Sci. Plant Nutr. 37, 299–311. Churchman, G. J., and Gilkes, R. J. (1989). Recognition of intermediates in the possible transformation of halloysite to kaolinite in weathering proﬁles. Clay Miner. 24, 579–590. Churchman, G. J., and Theng, B. K. G. (1984). Interactions of halloysites with amides: mineralogical factors affecting complex formation. Clay Miner. 19, 161–175. Cradwick, P. D. G., Farmer, V. C., Russell, J. D., Masson, C. R., Wada, K., and Yoshinaga, N. (1972). Imogolite, a hydrated aluminum silicate of tubular structure. Nature (London) Phys. Sci. 240, 187 –189. Cronan, C. S., Walker, W. J., and Bloom, P. R. (1986). Predicting aqueous aluminum concentrations in natural waters. Nature. 324, 140–143. Cronin, S. J., Hedley, M. J., Neall, V. E., and Smith, R. G. (1998). Agronomic impact of tephra fallout from the 1995 and 1996 Ruapehu volcano eruptions, New Zealand. Environ. Geol. 34, 21–30. Cronin, S. J., Hedley, M. J., Smith, R. G., and Neall, V. E. (1997). Impact of Ruapehu ash fall on soil and pasture nutrient status. 1. October 1995 eruptions. N.Z. J. Agric. Res. 40, 383–395. Cullen, S. J., Montagne, C., and Ferguson, H. (1991). Timber harvest trafﬁcking and soil compaction in western Montana. Soil Sci. Soc. Am. J. 55, 1416–1421. Classiﬁcation Committee of Cultivated Soils. (1996). “Classiﬁcation of Cultivated Soils in Japan, Third Approximation”, p. 82. National Institute of Agro-Environmental Sciences, Tsukuba, Japan. Dahlgren, R. A., and Saigusa, M. (1994). Aluminum release rates from allophanic and nonallophanic Andosols. Soil Sci. Plant Nutr. 40, 125–136. Dahlgren, R. A., and Ugolini, F. C. (1989a). Effects of tephra addition on soil processes in Spodosols in the Cascade Range, Washington, U.S.A. Geoderma. 45, 331 –355. Dahlgren, R. A., and Ugolini, F. C. (1989b). Formation and stability of imogolite in a tephritic Spodosol, Cascade Range, Washington, USA. Geochim. Cosmochim. Acta. 53, 1897–1904. Dahlgren, R. A., and Ugolini, F. C. (1991). Distribution and characterization of short-range-order minerals in Spodosols from the Washington Cascades. Geoderma. 48, 391 –413. Dahlgren, R. A., and Walker, W. J. (1993). Aluminum release rates from selected Spodosol Bs horizons: effect of pH and solid-phase aluminum pools. Geochim. Cosmochim. Acta. 57, 57–66. Dahlgren, R. A., Driscoll, C. T., and McAvoy, D. C. (1989). Aluminum precipitation and dissolution rates in Spodosol Bs horizons in the northeastern USA. Soil Sci. Soc. Am. J. 53, 1045–1052. Dahlgren, R. A., Ugolini, F. C., Shoji, S., Ito, T., and Sletten, R. S. (1991). Soil-forming processes in Alic Melanudands under Japanese pampas grass and oak. Soil Sci. Soc. Am. J. 55, 1049–1056. Dahlgren, R. A., Dragoo, J. P., and Ugolini, F. C. (1997). Weathering of Mt. St. Helens tephra under a cryic-udic climatic regime. Soil Sci. Soc. Am. J. 61, 1519–1525. Dahlgren, R. A., Ugolini, F. C., and Casey, W. H. (1999). Field weathering rates of Mt. St. Helens tephra. Geochim. Cosmochim. Acta. 63, 587 –598. Delvaux, B., and Herbillon, A. J. (1995). Pathways of mixed-layer kaolin–smectite formation in soils. In “Clays – Controlling the Environment” (G. J. Churchman, R. W. Fitzpatrick, and R. A. Eggleton, Eds.), pp. 457– 461. Proceedings 10th Int. Clay Conf., Adelaide, Australia, 1993, CSIRO Publ., Melbourne, Australia. Delvaux, B., Herbillon, A. J., and Vielvoye, L. (1989). Characterization of a weathering sequence of soils derived from volcanic ash in Cameroon– taxonomic, mineralogical and agronomic implications. Geoderma. 45, 375– 388.

170

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Delvaux, B., Herbillon, A. J., Dufey, J. E., and Vielvoye, L. (1990a). Surface properties and clay mineralogy of hydrated halloysitic soil clays. I. Existence of interlayer Kþ speciﬁc sites. Clay Miner. 25, 129–139. Delvaux, B., Herbillon, A. J., Vielvoye, L., and Mestdagh, M. M. (1990b). Surface properties and clay mineralogy of hydrated halloysitic soil clays. II. Evidence for the presence of halloysite/smectite (H/SM) mixed-layer clays. Clay Miner. 25, 141–160. Delvaux, B., Tessier, D., Herbillon, A. J., Gurtin, G., Jaunet, A. M., and Vielvoye, L. (1992). Morphology, texture, and microstructure of halloysitic soil clays as related to weathering and exchangeable cation. Clays Clay Miner. 40, 446– 456. Dethier, D. P., Pevear, D. R., and Frank, D. (1981). Alteration of new volcanic deposits. In “The 1980 Eruption of Mount Mt. Helens, Washington”. (P. W. Lipman and D. R. Mullineaux, Eds.). U.S. Geol. Surv. Prof. Pap, 1250, pp. 649 –665. Dorel, M., Roger-Estrade, J., Manichon, H., and Delvaux, B. (2000). Porosity and soil water properties of Caribbean volcanic ash soils. Soil Use Manage. 16, 133–140. Dubroeucq, D., Geissert, D., and Quantin, P. (1998). Weathering and soil forming processes under semi-arid conditions in two Mexican volcanic ash soils. Geoderma. 86, 99–122. Duchaufour, P. (1977). “Pe´dogene´se et Classiﬁcation”, p. 477. Masson, Paris. Dymond, J., Biscaye, P. E., and Rex, R. W. (1974). Eolian origin of mica in Hawaiian soils. Geol. Soc. Am. Bull. 85, 37–40. Eggleton, R. A., and Tilley, D. B. (1998). Hisingerite: A ferric kaolin mineral with curved morphology. Clays Clay Miner. 46, 400– 413. Elliot, E. T., and Coleman, D. C. (1988). Let the soil do the work for us. Ecol. Bull. 39, 23–32. Escudey, M., Galindo, G., Forster, J. E., Briceno, M., Diaz, P., and Chang, A. (2001). Chemical forms of phosphorus of volcanic ash-derived soils in Chile. Commun. Soil Sci. Plant Nutr. 32, 601– 616. Eswaran, H., Vandenberg, E., and Reich, P. (1993). Organic carbon in soils of the world. Soil Sci. Soc. Am. J. 57, 192 –194. FAO, ISRIC, and ISSS. (1998). “World Reference Base for Soil Resources” FAO, Rome. Farmer, V. C. (1987). The role of inorganic species in the transport of aluminum in Podzols. In “Podzols et Podzolization” (D. Righi and A. Chauvel, Eds.), pp. 187 –194. AFES et INRA, Paris. Farmer, V. C., and Fraser, A. R. (1979). Synthetic imogolite, a tubular hydroxy-aluminium silicate. In “Proceedings, International Clay Conference (Oxford 1978)” (M. M. Mortland and V. C. Farmer, Eds.), pp. 547–553. Elsevier, Amsterdam. Farmer, V. C., and Russell, J. D. (1990). The structure and genesis of allophanes and imogolite: their distribution in nonvolcanic soils. In “Soil Colloids and their Association in Soil Aggregates” (M. F. DeBoot, M. H. B. Hayes, and A. Herbillon, Eds.), pp. 165–178. Proc. NATO Advanced Studies Workshop, Ghent 1985, Plenum, New York. Fieldes, M. (1966). The nature of allophane in soils. Part 1. Signiﬁcance of structural randomness in pedogenesis. N.Z. J. Sci. 9, 599 –607. Fox, R. L., Silva, J. A., Younge, O. R., Plucknett, D. L., and Sherman, G. D. (1967). Soil and plant silicon and silicate response by sugarcane. Soil Sci. Soc. Am. Proc. 31, 775–779. Fridrik, P., Halldor, T., Holmfridur, S., Holmgeir, B., and Olafur, A. (1996). Nitrogen mineralization in soil [Icelandic]. Buvisindi. 10, 185 –208. Furuya, H., Wakai, A., Takahashi, T., and Ui, T. (1996). Inhibition of macroconidial germination of Fusarium solani f. sp. Phaseoli by soil aluminum. Ann. Phytopathol. Soc. Jpn. 62, 69–74. Furuya, H., Takahashi, T., and Matsumoto, T. (1999). Suppression of Fusarium solani f. sp. Phaseoli on bean by aluminum in acid soils. Phytopathology. 89, 47– 52. Gandeza, A. T., Shoji, S., and Yamada, I. (1991). Simulation of crop response to polyoleﬁn-coated urea. I. Field dissolution. Soil Sci. Soc. Am. J. 55, 1462–1467.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 171 Garcia-Rodeja, E., Silva, B. M., and Macı´as, F. (1987). Andosols developed from non-volcanic materials in Galicia, NW Spain. J. Soil Sci. 38, 573–591. Geist, J. M., Hazard, J. W., and Seidel, K. W. (1989). Assessing physical conditions of some Paciﬁc Northwest volcanic ash soils after forest harvest. Soil Sci. Soc. Am. J. 53, 946–950. Glassmann, J. R., and Simonson, G. M. (1985). Alteration of basalt in soils of Western Oregon. Soil Sci. Soc. Am. J. 49, 262–273. Global Volcanism Program. (2003). Dept. of Mineral Sciences, Smithsonian Institution, Washington DC. http://www.volcano.si.edu/gvp/. Goldchin, A., Oades, J. M., Skjemstad, J. O., and Clarke, P. (1994). Soil structure and carbon cycling. Aust. J. Soil Res. 32, 1043–1068. Goldchin, A., Baldock, J. A., Clarke, P., Higashi, T., and Oades, J. M. (1997). The effects of vegetation and burning on the chemical composition of soil organic matter in a volcanic ash soil. II. Density fractions. Geoderma. 76, 175–192. Goodman, B. A., Russell, J. D., Montez, B., Oldﬁeld, E., and Kirkpatrick, R. J. (1985). Structural studies of imogolite and allophane by aluminum-27 and silicon-29 nuclear magnetic resonance spectroscopy. Phys. Chem. Miner. 12, 342–346. Gregorich, E. G., and Janzen, H. H. (2000). Decomposition. In “Handbook of Soil Science” (M. E. Sumner, Ed.), pp. C107–120. CRC Press, Boca Raton, FL. Harris, S. L. (1988). “Fire Mountains of the West — The Cascade and Mono Lake Volcanoes”, p. 379. Mountain Press Pub., Missoula, MT. Hatano, R., Iwanaga, K., Okajima, H., and Sakuma, T. (1988). Relationship between the distribution soil macropores and root elongation. Soil Sci. Plant Nutr. 34, 535 –546. Hewitt, A. E. (1998). “New Zealand Soil Classiﬁcation”, 2nd edition. “Landcare Research Science Series No. 1”. Manaaki Whenua Press, Lincoln, New Zealand. Hidalgo, C., Bertaux, J., and Quantin, P. (1995). Forms of silica in indurated volcanic soils of the Mexico Valley. In “Clays–Controlling the Environment” (G. J. Churchman, R. W. Fitzpatrick, and R. A. Eggleton, Eds.), pp. 487– 493. Proceedings 10th Int. Clay Conf., Adelaide, Australia. CSIRO Publ., Melbourne, Australia. Hinkley, T. K., and Smith, K. S. (1987). Leachate chemistry of ash from the May 18, 1980 eruption of Mount St. Helens. In “Chemistry of Ash and Leachates from May 18, 1980 Eruption of Mount St. Helens, Washington” (T. K. Hinkley, Ed.), U.S. Geological Survey Professional Paper, 1397-B, pp. 27–55. Hiradate, S., Taniguchi, S., and Sakurai, K. (1998). Aluminum speciation in aluminum-silica solutions and potassium chloride extracts of acidic soils. Soil Sci. Soc. Am. J. 62, 630 –636. Hofstede, R. G. M. (1995). The effects of grazing and burning on soil and plant nutrient concentrations in Colombian paramo grasslands. Plant Soil. 173, 111–132. Huang, P. M. (1988). Ionic factors affecting aluminum transformations and the impact on soil and environmental sciences. Adv. Soil Sci. 8, 1–78. Huang, P. M., and Violante, A. (1986). Inﬂuence of organic acids on crystallization and surface properties of precipitation products of aluminum. In “The Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 159–221. Soil Science Society of America Special Publication 17, Madison, WI. Hunter, C. R., Frazier, B. E., and Busacca, A. J. (1987). Lytell series: a nonvolcanic Andisol. Soil Sci. Soc. Am. J. 51, 376– 383. Hunter, D., and Ross, D. S. (1991). Evidence for a phytotoxic hydroxyaluminum polymer in organic soil horizons. Science. 251, 1056–1058. Ibrahim, S. M., Inoue, Y., and Shindo, H. (1998). Role of active aluminum in the formation of waterstable macroaggregates. Soil Sci. Plant Nutr. 44, 685–689. Inamatsu, K., Kiuchi, M., and Shibuya, K. (1991). Acidiﬁcation of mulberry ﬁelds soil in case of volcanic ash soils on Tsukuba upland. Jpn. J. Soil Sci. Plant Nutr. 62, 351–356.

172

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

˚ minerals in the volcanic ash soils in Japan. Inoue, K. (1981). Implications of eolian dusts to 14A Pedologist. 25, 97–118. Inoue, K., and Naruse, T. (1987). Physical chemical and mineralogical characteristics of modern eolian dust in Japan and rate of dust deposition. Soil Sci. Plant Nutr. 33, 327 –345. Inoue, K., and Higashi, T. (1988). Al– and Fe–humus complexes in Andisols. In “Proceedings 9th Int. Soil Classiﬁcation Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 535– 546. Japanese Committee for 9th International Soil Classiﬁcation Workshop and Soil Management Support Services, Washington, D.C. Inoue, H., Ito, T., and Saigusa, M. (2000). Yield and nutrient uptake of dent corn (Zea mays L.) in notillage cultivation with single basal co-situs fertilization. Jpn. J. Soil Sci. Plant Nutr. 71, 674–681. Inoue, K., Kondo, S., Tamano, Y., and Yokota, H. (2001). Amelioration of subsoil acidity in nonallophanic Andosol by surface application of organic calcium salts. Soil Sci. Plant Nutr. 47, 113 –122. Isobe, K., and Tsuboki, Y. (1997). Study on utilization of arbuscular mycorrhizal fungi in kidney bean cultivation — effects of available phosphorus content on the infection of arbuscular mycorrhizal fungi and two species of arbuscular mycorrhizal on the growth of kidney bean (Phaseolus vulgaris L.). Jpn. J. Crop Sci. 66, 374– 380. Ito, T., Shoji, S., and Saigusa, M. (1991). Classiﬁcation of volcanic ash soils from Konsen District, Hokkaido, according to the last Keys to Soil Taxonomy (1990). Jpn. J. Soil Sci. Plant Nutr. 62, 237–247. Ito, T., Inoue, H., and Saigusa, M. (1997a). Effects of no-tillage system on silage-corn yields and characteristics in Andisol. In “Plant Nutrition for Sustainable Food Production and Environment” (T. Ando, K. Fujita, T. Mae, H. Matsumoto, S. Mori, and J. Sekiya, Eds.), pp. 653–654. Kluwer Academic Publishers, Dordrecht. Ito, T., Ide, T., and Saigusa, M. (1997b). Change of physical properties with agricultural machine trafﬁc in cultivated Andisols with different colloidal composition. Bull. Kawatabi Exp. Farm. 13, 17–21. Ito, T., Yamada, D., and Saigusa, M. (1998). Effects of subsoil and polyoleﬁn-coated urea application on growth and nitrogen uptake of oats and barley in Andisols. Tohoku J. Agric. Res. 49, 25 –32. Ito, T., Kikawa, N., and Saigusa, M. (2000a). Phosphorus availability difference between allophanic and nonallophanic Andisols. In “Abstracts — Int. Symp. Productivity and Charge Characteristics of Soils Derived from Volcanic Ash”, p. 18. Meirin Printing Co, Sendai, Japan. Ito, T., Yokota, T., and Saigusa, M. (2000b). Measurement of organic phosphorus mineralization in Andisols using anion exchange resin In “Abstracts — Int. Symp. Productivity and Charge Characteristics of Soils Derived from Volcanic Ash”, p. 19. Meirin Printing Co, Sendai, Japan. Jackson, M. L. (1963a). Aluminum bonding in soils: a unifying principle in soil science. Soil Sci. Soc. Am. Proc. 37, 1–10. Jackson, M. L. (1963b). Interlayering of expansible layer silicates in soils by chemical weathering. Clays Clay Mineral. 11, 29–46. Johnson-Maynard, J. L., McDaniel, P. A., Ferguson, D. E., and Falen, A. L. (1997). Chemical and mineralogical conversion of Andisols following invasion by bracken fern. Soil Sci. Soc. Am. J. 61, 549–555. Jongmans, A. G., van Oort, F., Nieuwenhuyse, A., Buurman, P., Jaunet, A. M., and van Doesburg, J. D. J. (1994). Inheritance of 2:1 phyllosilicates in Costa Rican Andisols. Soil Sci. Soc. Am. J. 58, 494–501. 22 Kamewada, K. (1996). Application of “four-plane model” to the adsorption of Kþ, NO2 3 , and SO4 from a mixed solution of KNO3 and K2SO4 on Andisols. Soil Sci. Plant Nutr. 42, 801–808. Karasawa, T., Arihara, J., and Kasahara, Y. (2000a). Effects of previous crops on arbuscular mycorrhizal formation and growth of maize under various soil moisture conditions. Soil Sci. Plant Nutr. 46, 53–60.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 173 Karasawa, T., Takebe, M., and Kasahara, Y. (2000b). Arbuscular mycorrhizal (AM) effects on maize growth and AM colonization of root under various soil conditions. Soil Sci. Plant Nutr. 46, 61–67. Karathanasis, A. D., Adams, F., and Hajek, B. F. (1983). Stability relationships in kaolinite, gibbsite, and Al-hydroxy interlayered vermiculite soil systems. Soil Sci. Soc. Am. J. 47, 1247–1251. Kato, H., Shirai, M., and Matsukawa, S. (1995). Ionic species of aluminium in acid soils in terms of chemical equilibrium. Jpn. J. Soil Sci. Plant Nutr. 66, 39 –47. Katou, H., Clothier, B. E., and Green, S. R. (1996). Anion transport involving competitive adsorption during transient water ﬂow in an Andisol. Soil Sci. Soc. Am. J. 60, 1368–1375. Kilbertus, G. (1980). Microhabitats in soil aggregates. Their relationship with bacterial biomass and the size of procaryotes present. Rev. Ecol. Biol. Sol. 17, 543 –557. Kimble, J. M., and Eswaran, H. (1988). The Andisol-Oxisol transition. In “Proceedings 9th Int. Soil Classiﬁcation Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 219–232. Japanese Committee for 9th International Soil Classiﬁcation Workshop and Soil Management Support Services, Washington, DC. Kimble, J. M., Ping, C. L., Sumner, M. E., and Wilding, L. P. (2000). Andisols. In “Handbook of Soil Science” (M. E. Sumner, Ed.), pp. E209–224. CRC Press, Boca Raton, FL. Kirkman, J. H., and McHardy, W. J. (1980). A comparative study of the morphology, chemical composition and weathering of rhyolitic and andesitic glass. Clay Miner. 15, 165–173. Kobayashi, S., Shoji, S., Yamada, I., and Masui, J. (1976). Chemical and mineralogical studies on volcanic ashes. III. Some mineralogical and chemical properties of volcanic glasses with special reference to the rock types of volcanic ashes. Soil Sci. Plant Nutr. 22, 7–13. Kondo, Y., Fujitani, T., Katsui, Y., and Niida, K. (1979). Nature of the 1977–1978 volcanic ash from Usu volcano, Hokkaido, Japan. Kazan. 24, 223– 238. LaManna, J. M., and Ugolini, F. C. (1987). Trioctahedral vermiculite in the 1980 pyroclastic ﬂow, Mt. St. Helens, Washington. Soil Sci. 143, 162–167. Leamy, M. L. (1984). “International Committee on the Classiﬁcation of Andisols (ICOMAND) Circular letter No. 6.” New Zealand Soil Bureau, DSIR, Lower Hutt. Leamy, M. L., Clayden, B., Parﬁtt, RL., Kinloch, D. I., and Childs, C. W. (1988). “The Andisol Proposal 1988, Final proposal of the International Committee on the Classiﬁcation of Andisols (ICOMAND).” New Zealand Soil Bureau, Lower Hutt, New Zealand. Lenhard, R. J. (1986). Changes in void distribution and volume during compaction of a forest soil. Soil Sci. Soc. Am J. 50, 462–464. Loughnan, F. C. (1969). “Chemical Weathering of the Silicate Minerals”, p. 154. Elsevier, New York. Lowe, D. J. (1986). Controls on the rates of weathering and clay mineral genesis in airfall tephras: a review and New Zealand case study. In “Rates of Chemical Weathering of Rocks and Minerals” (S. M. Colman and D. P. Dethier, Eds.), pp. 265 –330. Academic Press, Orlando. MacKenzie, K. J. D., Bowden, M. E., and Meinhold, R. H. (1991). The structure and thermal transformations of allophanes studied by 29Si and 27Al high-resolution solid-state NMR. Clays Clay Miner. 39, 337– 346. Madeira, M., Furtado, A., Jeanroy, E., and Herbillon, A. J. (1994). Andisols of Madeira Island (Portugal) — characteristics and classiﬁcation. Geoderma. 62, 363–383. Maeda, T., and Soma, K. (1986). Physical properties. In “Ando Soils in Japan” (K. Wada, Ed.), pp. 99–111. Kyushu University Press, Fukuoka, Japan. Maeda, T., Takenaka, H., and Warkentin, B. P. (1977). Physical properties of allophanic soils. Adv. Agron. 29, 229 –264. Malucelli, F., Terribile, F., and Colombo, C. (1999). Mineralogy, micromorphology and chemical analysis of Andosols on the Island of Sa˜o Miguel (Azores). Geoderma. 88, 73–98. Manske, G. G. B., Ortiz-Monasterio, J. I., Van Ginkel, M., Gonzalez, R. M., Rajaram, S., Molina, E., and Vlek, P. L. G. (2000). Trait associated with improved P-uptake efﬁciency in CIMMYT’s semidwarf spring bread wheat grown on an acid Andisol in Mexico. Plant Soil. 221, 189–204.

174

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Martini, J. A. (1976). The evolution of soil properties as it relates to the genesis of volcanic ash soils in Costa Rica. Soil Sci. Soc. Am. J. 40, 895–900. Matsuyama, N., Saigusa, M., and Kudo, K. (1994). Phosphorus response of dent corn young plants and evaluation of phosphorus availability in Andosols. Tohoku J. Crop Sci. 37, 57–58. Matsuyama, N., Kudo, K., and Saigusa, M. (1999a). Active aluminum of cultivated Andosols and related soil chemical properties in Japan. Bull. Fac. Agric. Life Sci. Hirosaki Univ. 1, 30 –36. Matsuyama, N., Saigusa, M., and Kudo, K. (1999b). Acidity of Japanese cultivated Andosols and signiﬁcance of exchange acidity Y1 in their classiﬁcation. J. Soil Sci. Plant Nutr. 70, 754– 761. May, H. M., Helmke, P., and Jackson, M. L. (1979). Gibbsite solubility and thermodynamic properties of hydroxyaluminum ions in aqueous solutions at 258C. Geochim. Cosmochim. Acta. 50, 1667–1677. McKeague, J. A. (1967). An evaluation of 0.1 M pyrophosphate and pyrophosphate-dithionite in comparison with oxalate as extractants of the accumulation products in podzols and some other soils. Can. J. Soil Sci. 47, 95–99. Meurisse, R. T. (1988). Forest productivity and the management of U.S. Andisols. In “Proceedings 9th Int. Soil Classiﬁcation Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 245 –257. Japanese Committee for 9th International Soil Classiﬁcation Workshop and Soil Management Support Services, Washington, DC. Mitchell, W. A. (1975). Heavy minerals. In “Soil Components, Vol. 2. Inorganic Components” (J. E. Gieseking, Ed.), pp. 449 –480. Springer-Verlag, Berlin-Heidelberg-New York. Mizota, C., and Inoue, K. (1988). Eolian dust contribution to soil development on volcanic ashes in Japan. In “Proceedings 9th Int. Soil Classiﬁcation Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 547–557. Japanese Committee for 9th International Soil Classiﬁcation Workshop and Soil Management Support Services, Washington, DC. Mizota, C., and Matsuhisa, Y. (1995). Isotopic evidence for the eolian origin of quartz and mica in soils developed on volcanic materials in the Canary Archipelago. Geoderma. 66, 313– 320. Mizota, C., and van Reeuwijk, L. P. (1989). “Clay Mineralogy and Chemistry of Soils formed in Volcanic Material in Diverse Climatic Regions”, p. 186. Soil Monograph 2, ISRIC, Wageningen, the Netherlands. Mizota, C., Kawasaki, I., and Wakatsuki, T. (1988). Clay mineralogy and chemistry of seven pedons formed in volcanic ash, Tanzania. Geoderma. 43, 131–141. Mizota, C., Kusakabe, M., and Noto, M. (1990). Oxygen isotope composition of quartz in soils developed on late Quaternary volcanic ashes in Japan. Geoderma. 46, 319–327. Mizuno, N., and Yoshida, H. (1993). Effect of exchangeable aluminum on the reduction of potato scab. Plant Soil. 156, 505–508. Mizuno, N., Nishimidin, K., Yoshida, H., Nanzyo, M., and Tadano, T. (1998a). Effects of deep plowing and “cutting and banking” practice on the concentration of water-soluble aluminum and allophane content in Andosols: Implication for recent incidence of common potato scab in Abashiri area. Soil Sci. Plant Nutr. 44, 571 –578. Mizuno, N., Yoshida, H., Nanzyo, M., and Tadano, T. (1998b). Chemical characterization of conducive and suppressive soils for potato scab in Hokkaido, Japan. Soil Sci. Plant Nutr. 44, 289–295. Mizuno, N. K., Yoshida, H., and Tadano, T. (2000). Efﬁcacy of single application of ammonium sulfate in suppressing potato scab. Soil Sci. Plant Nutr. 46, 611 –616. Mohr, E. C. J. (1938). The relation between soil and population density in the Netherlands Indies. In “Compies Rendus du Congres International de Geogrphic”, pp. 478 –493. Amsterdam. Tome Deuxiem, Section IIIc. Mora, M. L., Baeza, G., Pizarro, C., and Demanet, R. (1999a). Effect of calcitic and dolomitic lime on physicochemical properties of a Chilean Andisol. Commun. Soil Sci. Plant Anal. 30, 427–439.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 175 Mora, M. L., Schnettler, B., and Demanet, R. (1999b). Effect of liming and gypsum on soil chemistry, yield, and mineral composition of rye grass grown in an acidic Andisol. Commun. Soil Sci. Plant Anal. 30, 1251–1266. Nakamaru, Y., Nanzyo, M., and Yamasaki, S. (2000a). Short-term reaction of apatite in fresh Pinatubo volcanic ash with reactive components in the soil-plant system. Jpn. J. Soil Sci. Plant Nutr. 71, 55 –62. Nakamaru, Y., Nanzyo, M., and Yamasaki, S. (2000b). Utilization of apatite in fresh volcanic ash by pigeonpea and chickpea. Soil Sci. Plant Nutr. 46, 591–600. Nanzyo, M. (1987). Formation of noncrystalline aluminum phosphate through phosphate sorption on allophanic Andosols. Commun. Soil Sci. Plant Anal. 18, 735 –742. Nanzyo, M., Takahashi, T., Sato, A., Shoji, S., and Yamada, I. (1997). Dilute acid-soluble phosphorus in fresh air-borne tephras and ﬁxation with an increase in active aluminum and iron. Soil Sci. Plant Nutr. 43, 839 –848. Nanzyo, M., Nakamaru, Y., and Yamasaki, S. (1999). Inhibition of apatite dissolution due to formation of calcium oxalate coating. Phosphorus Res. Bull. 9, 17–22. Newman, R. H., Childs, C. W., and Churchman, G. J. (1994). Aluminum coordination and structural disorder in halloysite and kaolinite by 27Al NMR spectroscopy. Clay Miner. 29, 305 –312. Nira, R., and Nishimune, A. (1993). Studies on nitrogen mineralization properties of Tokachi soils by kinetic analysis. Soil Sci. Plant Nutr. 39, 321 –329. Norrish, K. (1995). An unusual ﬁbrous halloysite. In “Clays–Controlling the Environment” (G. J. Churchman, R. W. Fitzpatrick, and R. A. Eggleton, Eds.), pp. 275– 284. Proceedings 10th Int. Clay Conf., Adelaide, Australia. 1993. CSIRO Publ., Melbourne, Australia. Nuhn, W. W. (1987). “Soil Genesis on the 1980 Pyroclastic Flow of Mount St. Helens”, p. 179. M.S. thesis, College of Forest Resources, University of Washington, Seattle, WA. Okada, T., Hosokawa, K., and Kawaguchi, K. (1986). Inﬂuences of excessive fertilizing on tea plants and soils. Bull. Tokushima Exp. Agric. Instit. 23, 23–31. Okamura, Y., and Wada, K. (1983). Electrical charge characteristics of Ando(B) and Red-Yellow B soils and weathered pumices. J. Soil Sci. 34, 287 –295. Okano, S. (1990). Availability of mineralized nitrogen from microbial biomass and organic matter after drying and heating of grassland soils. Plant Soil. 129, 219 –225. Ombodi, A., and Saigusa, M. (2000). Improvement of nitrogen use efﬁciency of tendergreen mustard grown in the ﬁeld during rainy seasons by using polyoleﬁn-coated urea. Tohoku J. Agric. Res. 51, 1 –6. Oskarsson, N. (1980). The interaction between volcanic gases and tephra: ﬂuorine adhering to tephra of the 1970 Hekla eruption. J. Volcanol. Geotherm. Res. 8, 251 –256. Otsuka, H., Briones, A. A., Daquiado, N. P., and Evangelio, F. A. (1988). Characteristics and genesis of volcanic ash soils in the Philippines. Technical Bulletin of Tropical Agricultural Research Center, Japan. 24, 1–122. Palacios, M. S., Shimada, M. K., and Salinas, C. C. (1987). Effect of inoculation in two varieties of onion (Allium cepa L.) with four VA endomycorrhizal fungi in phosphate deﬁcient soil. Revista Latinoamericana de Microbiologia. 29, 329–336. Pansombat, K., Kanazawa, S., and Horiguchi, T. (1997). Microbial ecology in tea soils. I. Soil properties and microbial populations. Soil Sci. Plant Nutr. 43, 317–327. Parﬁtt, R. L. (1990). Allophane in New Zealand — A review. Aust. J. Soil Res. 28, 343–360. Parﬁtt, R. L., and Henmi, T. (1980). Structure of some allophanes from New Zealand. Clays Clay Miner. 28, 285– 294. Parﬁtt, R. L., and Kimble, J. M. (1989). Conditions for formation of allophane in soils. Soil Sci. Soc. Am. J. 53, 971–977. Parﬁtt, R. L., and Saigusa, M. (1985). Allophane and humus–aluminum in Spodosols and Andepts formed from the same volcanic ash beds in New Zealand. Soil Sci. 139, 149– 155.

176

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Parﬁtt, R. L., and Wilson, A. D. (1985). Estimation of allophane and halloysite in three sequences of volcanic soils, New Zealand. In “Volcanic Soils” (E. Fernandez Caldas and D. H. Yaalon, Eds.), Catena Suppl. 7, pp. 1–8. Parﬁtt, R. L., Furkert, F. J., and Henmi, T. (1980). Identiﬁcation and structure of two types of allophane from volcanic ash soils and tephra. Clays Clay Min. 28, 328–334. Parﬁtt, R. L., Russell, M., and Orbell, G. E. (1983). Weathering sequence of soils from volcanic ash involving allophane and halloysite. Geoderma. 29, 41–57. Parﬁtt, R. L., Tate, K. R., and McKercher, R. B. (1994). Measurement of phosphorus mineralisation using an anion exchange membrane. Commun. Soil Sci. Plant Anal. 25, 3209–3219. Parﬁtt, R. L., Salt, G. J., and Saggar, S. (2001). Effect of leaching and clay content on carbon and nitrogen mineralisation in maize and pasture soils. Aust. J. Soil Res. 39, 535 –542. Paterson, E. (1977). Speciﬁc surface area and pore structure of allophanic soil clays. Clay Miner. 12, 1–9. Percival, H. J., Parﬁtt, R. L., and Scott, N. A. (2000). Factors controlling soil carbon levels in New Zealand grasslands: is clay content important? Soil Sci. Soc. Am. J. 64, 1623–1630. Perret, S., and Dorel, M. (1999). Relationship between land use, fertility and Andisol behaviour: examples from volcanic islands. Soil Use Manage. 15, 144– 149. Perret, S., Michellon, R., Boyer, J., and Tassin, J. (1996). Soil rehabilitation and erosion control through agro-ecological practices on Reunion Island. Agric. Ecosyst. Environ. 59, 149 –157. Pevear, D. R., Dethier, D. P., and Frank, D. (1982). Clay minerals in the 1980 deposits from Mount St. Helens. Clays Clay Miner. 30, 241–252. Ping, C. L., Shoji, S., and Ito, T. (1988). Properties and classiﬁcation of three volcanic ash-derived pedons from Aleutian Islands and Alaska Peninsula, Alaska. Soil Sci. Soc. Am. J. 52, 455– 462. Ping, C. L., Shoji, S., Ito, T., Takahashi, T., and Moore, J. P. (1989). Characteristics and classiﬁcation of volcanic-ash-derived soils in Alaska. Soil Sci. 148, 8–28. Poudel, D. D., Midmore, D. J., and West, L. T. (1999). Erosion and productivity of vegetable systems on sloping volcanic ash-derived Philippine soils. Soil Sci. Soc. Am. J. 63, 1366–1376. Poulenard, J., Podwojewski, P., Janeau, J. L., and Collinet, J. (2001). Runoff and soil erosion under rainfall simulation of Andisols from the Ecuadorian Paramo: effect of tillage and burning. Catena. 45, 185 –207. Qafoku, N. P., Sumner, M. E., and Radcliffe, D. E. (2000). Anion transport in columns of variable charge subsoils: nitrate and chloride. J. Environ. Qual. 29, 484–493. Quantin, P. (1992). “Les Soils de l’ Archipel Volcanique des Nouvelles-Hebrides (Vanuatu)”, p. 498. ORSTOM, Paris. Quantin, P., Gautheyrou, J., and Lorenzoni, P. (1988). Halloysite formation through in situ weathering of volcanic glass from trachytic pumices, Vico’s Volcano, Italy. Clay Miner. 23, 423–437. Reynolds-Vargas, J. S., Richter, D. D., and Bornemisza, E. (1994). Environmental impacts of nitriﬁcation and nitrate adsorption in fertilized Andisols in the Valle-Central of Costa Rica. Soil Sci. 157, 289–299. Robarge, W. P. (1999). Precipitation/dissolution reactions in soils. In “Soil Physical Chemistry”, 2nd edn. (D. L. Sparks, Ed.), pp. 193– 238. CRC Press, Boca Raton. Ruxton, B. P. (1988). Towards a weathering model of Mount Lamington Ash, Papua New Guinea. Earth Sci. Rev. 25, 387–397. Ryan, M. C., Graham, G. R., and Rudolph, D. L. (2001). Contrasting nitrate adsorption in Andisols of two coffee plantation in Costa Rica. J. Environ. Qual. 30, 1848–1852. Saigusa, M. (1988). Agronomy of Aludands in northeastern Japan. In “Proceedings 9th Int. Soil Classiﬁcation Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 593 –602. Japanese Committee for 9th International Soil Classiﬁcation Workshop and Soil Management Support Services, Washington, DC. Saigusa, M. (1991). Plant growth on acid soils with special reference to phytotoxic Al and subsoil acidity (review). Jpn. J. Soil Sci. Plant Nutr. 62, 451 –459.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 177 Saigusa, M. (1992). Characterization and soil management of acid Andisols. In “Plant-Soil Interaction at Low pH” (T. Sakuma, Ed.), pp. 36–41. Faculty of Agric., Hokkaido Univ, Sapporo, Japan. Saigusa, M. (1999). Slow-release nitrogen fertilizers and plant nutrition. In “Nitrogen Nutrition and Plant Growth” (H. S. Srivastava and R. P. Singh, Eds.), pp. 305–335. Science Publishers, Enﬁeld, N.H. Saigusa, M. (2000). Growth of rice and its silicon nutrition. Kagaku To Seibutsu. 38, 514–519. Saigusa, M., and Matsuyama, N. (1998). Distribution of allophanic Andosols and non-allophanic Andosols in Japan. Tohoku J. Agric. Sci. 48, 75– 83. Saigusa, M., and Toma, M. (1997). Mechanism of reduction of exchangeable aluminum by gypsum application in acid Andosols. Soil Sci. Plant Nutr. 43, 343 –349. Saigusa, M., Saito, K., and Shoji, S. (1976). Occurrence of Cu and Zn within the proﬁles of soils formed from different parent ashes and corn response to these elements. Tohoku J. Agric. Res. 27, 12 –19. Saigusa, M., Shoji, S., and Kato, T. (1978). Origin and nature of halloysite in Ando soils from Towada tephra, Japan. Geoderma. 20, 115–129. Saigusa, M., Shoji, S., and Takahashi, T. (1980). Plant root growth in acid Andosols from northeastern Japan. Exchange acidity Y1 as a realistic measure of aluminum toxicity potential. Soil Sci. 130, 242 –250. Saigusa, M., Shoji, S., and Sakai, H. (1983). The effects of subsoil acidity of Andosols on the growth and nitrogen uptake of barley and wheat. Jpn. J. Soil Sci. Plant Nutr. 54, 460– 466. Saigusa, M., Shoji, S., and Nakaminami, H. (1987). Measurement of water retention at 15 bar tension by pressure membrane method and available moisture of Andosols. Jpn. J. Soil Sci. Plant Nutr. 58, 374–377. Saigusa, M., Shoji, S., Goto, J., Kodama, H., and Abe, T. (1990). Nitrogen fertility of acid Andisols and nitrogen application to upland crops. Trans. 14th Inter. Cong. Soil Sci. IV, 638– 639. Saigusa, M., Matsuyama, N., Honnna, T., and Abe, T. (1991). Chemistry and fertility of acid Andisols with special reference to subsoil acidity. In “Plant– Soil Interactions at Low pH” (R. J. Wright, V. C. Baligar, and R. P. Murrmann, Eds.), pp. 73–80. Kluwer Academic Publishers, Dordrecht. Saigusa, M., Hossain, M. Z., Toma, M., Shibuya, K., and Abe, T. (1992a). Utilization of nonallophanic Andosols as nursery bed soil of rice seedlings. Bull. Kawatabi Farm. 8, 1–5. Saigusa, M., Matsuyama, N., and Abe, T. (1992b). Electric charge characteristics of Andosols and its problem on soil management. Jpn. J. Soil Sci. Plant Nutr. 63, 196–201. Saigusa, M., Toma, M., and Abe, T. (1994). Effects of phosphogypsum application in topsoil on amelioration of subsoil acidity of nonallophanic Andosols. J. Jpn. Soc. Grass Sci. 39, 397 –404. Saigusa, M., Matsumoto, T., and Abe, T. (1995). Phytotoxicity of monomer aluminum ions and hydroxy-aluminum polymer ions in an Andosol. In “Plant–Soil Interaction at Low pH” (R. A. Date, N. J. Grundon, G. E. Rayment, and M. E. Probert, Eds.), pp. 367 –370. Kluwer Academic Publishers, Dordrecht. Saigusa, M., Hossain, M. Z., Tashiro, T., and Shibuya, K. (1996). Maximizing rice yield with controlled availability fertilizer and no-tillage culture in controlling environmental degradation. In “Proc. Int. Symp. Maximizing Sustainable Rice Yields through Improved Soil and Environmental Management” Khonkaen, Thailand”, 1, pp. 75 –85. Saigusa, M., Yamamoto, A., and Shibuya, K. (1998). Agricultural use of porous hydrate calcium silicate. I. Evaluation of porous hydrate calcium silicate as a silicate fertilizer. Jpn. J. Soil Sci. Plant Nutr. 69, 576 –581. Saigusa, M., Yamamoto, A., and Shibuya, K. (1999). Agricultural use of porous hydrate calcium silicate-Effect of porous hydrated calcium silicate on resistance of rice plant (Oryza sativa L.) to rice blast (Pyricularia olyzae). Plant Prod. Sci. 3, 51 –54. Saito, M. (1990). Nitrogen mineralization parameters and its availability indices of soils in Tohoku District, Japan: their relationship. Jpn. J. Soil Sci. Plant Nutr. 61, 265– 272.

178

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Saito, M., and Ishi, K. (1987). Estimation of soil nitrogen mineralization in corn-grown ﬁelds based on mineralization parameters. Soil Sci. Plant Nutr. 33, 555 –566. Sakamoto, K., and Oba, Y. (1993). Relationship between available N and soil biomass in upland ﬁeld soils. Jpn. J. Soil Sci. Plant Nutr. 64, 42–48. Savant, N. K., Snyder, G. H., and Datnoff, L. E. (1997). Silicon management and sustainable rice production. Adv. Agron. 58, 151 –199. Schwertmann, U. (1985). The effect of pedogenic environments on iron oxide minerals. Adv. Soil Sci. 1, 171– 200. Schwertmann, U. (1988). Occurrence and formation of iron oxides in various pedoenvironments. In “Iron in Soils and Clay Minerals” (J. W. Stucki, B. A. Goodman, and U. Schwertmann, Eds.), pp. 267 –308. Reidel, Bordrecht. Schwertmann, U., and Cornell, R. M. (1991). “Iron Oxides in the Laboratory — Preparation and Characterization.”, p. 137. Weinheim, New York. VCH. Schwertmann, U., and Taylor, R. M. (1989). Iron oxides. In “Minerals in Soil Environments”, 2nd edn. (J. B. Dixon and S. B. Weed, Eds.), pp. 379–438. Soil Sci. Soc. Am., Madison, WI. Shann, J. R., and Bertsch, P. M. (1993). Differential cultivar response to polynuclear hydroxyoaluminum complexes. Soil Sci. Soc. Am. J. 57, 116 –120. Shimizu, H., Watanabe, T., Henmi, T., Masuda, A., and Saito, H. (1988). Studies on allophane and imogolite by high-resolution solid-state 29Si- and 27Al-NMR and ESR. Geochem. J. 22, 23 –31. Shimono, K. (1990). Effects of soil pH on the growth and yield of ﬁeld crops in wet Andosols. Jpn. J. Soil Sci. Plant Nutr. 61, 8–15. Shoji, S. (1983). Mineralogical properties of volcanic ash soils. In “Volcanic Ash Soil — Genesis, Properties, Classiﬁcation” (N. Yoshinaga, Ed.), pp. 31 –72. Hakuyusha, Tokyo. Shoji, S. (1985). Genesis and properties of non-allophanic Andisols in Japan. Applied Clay Sci. 1, 83–88. Shoji, S., and Fujiwara, T. (1984). Active aluminum and iron in the humus horizons of Andosols from northeastern Japan: their forms, properties, and signiﬁcance in clay weathering. Soil Sci. 137, 216–226. Shoji, S., and Masui, J. (1971). Opaline silica of recent volcanic ash soils in Japan. J. Soil Sci. 22, 101–112. Shoji, S., and Ono, T. (1978). Physical and chemical properties and clay mineralogy of Andosols from Kitakami, Japan. Soil Sci. 126, 297–312. Shoji, S., Yamada, I., and Masui, J. (1974). Soils formed from the andesitic and basaltic ashes. I. The nature of the parent ashes and soil formation. Tohoku, J. Agric. Res. 25, 104–112. Shoji, S., Kobayashi, S., Yamada, I., and Masui, J. (1975). Chemical and mineralogical studies on volcanic ashes. I. Chemical composition of volcanic ashes and their classiﬁcation. Soil Sci. Plant Nutr. 21, 311 –318. Shoji, S., Saigusa, M., and Ebihara, M. (1980a). Cobalt content in volcanic ashes. Jpn. J. Soil Sci. Plant Nutr. 51, 335–336. Shoji, S., Saigusa, M., and Takahashi, T. (1980b). Plant root growth in acid Andosols from northeastern Japan. I. Soil properties and root growth of burdock, barley, and orchard grass. Soil Sci. 130, 124–130. Shoji, S., Fujiwara, Y., Yamada, I., and Saigusa, M. (1982). Chemistry and clay mineralogy of ando soils, brown forest soils, and podzolic soils formed from recent Towada ashes, N.E. Japan. Soil Sci. 133, 69–86. Shoji, S., Ito, T., Saigusa, M., and Yamada, I. (1985). Properties of nonallophanic Andosols from Japan. Soil Sci. 140, 264– 277. Shoji, S., Saigusa, M., and Goto, J. (1986). Acidity to subsoil of Andisols and nitrogen uptake and growth of sorghum. Jpn. J. Soil Sci. Plant Nutr. 57, 264–271. Shoji, S., Suzuki, Y., and Saigusa, M. (1987). Clay mineralogical and chemical properties of nonallophanic Andepts (Andisols) from Oregon, USA. Soil Sci. Soc. Am. J. 51, 986–990.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 179 Shoji, S., Hakamada, T., and Tomioka, E. (1990a). Properties and classiﬁcation of selected volcanic ash soils from Abashiri, northern Japan–transition of Andisols to Mollisols. Soil Sci. Plant Nutr. 36, 409–423. Shoji, S., Kurebayashi, T., and Yamada, I. (1990b). Growth and chemical composition of Japanese pampas grass (Miscanthus sinensis) with special reference to the formation of dark-colored Andisols in northeastern Japan. Soil Sci. Plant Nutr. 36, 105– 120. Shoji, S., Gandeza, A. T., and Kimura, K. (1991). Simulation of crop response to polyoleﬁn-coated urea. II. Nitrogen uptake by corn. Soil Sci. Soc. Am. J. 55, 1468–1473. Shoji, S., Nanzyo, M., and Dahlgren, R. A. (1993a). “Volcanic Ash Soils — Genesis, Properties and Utilization.”, p. 288. Elsevier, Amsterdam, the Netherlands. Shoji, S., Nanzyo, M., Shirato, Y., and Ito, T. (1993b). Chemical kinetics of weathering in young Andisols from northeastern Japan using 108C-normalized soil age. Soil Sci. 155, 53–60. Shoji, S., Nanzyo, M., Dahlgren, R. A., and Quantin, P. (1996). Evaluation and proposed revisions of criteria for Andosols in the World Reference Base for Soil Resources. Soil Sci. 161, 604–615. Shoji, S., Takahashi, T., Ito, T., and Ping, C. L. (1988a). Properties and classiﬁcation of selected volcanic ash soils from Kenai Peninsula, Alaska. Soil Sci. 145, 395 –413. Shoji, S., Takahashi, T., Saigusa, M., Yamada, I., and Ugolini, F. C. (1988b). Properties of Spodosols and Andisols showing climosequential and biosequential relations in southern Hakkoda, northeastern Japan. Soil Sci. 145, 135–150. Simkin, T., and Siebert, L. (1994). “Volcanoes of the world: a regional directory, gazetteer, and chronology of volcanism during the last 10,000 years.” 2nd edn. p. 349. Geoscience Press, Tucson, AZ. Simonett, D. A., and Bauleke, M. P. (1963). Mineralogy of soils on basalt in North Queensland. Soil Sci. Soc. Am. Proc. 27, 205–212. Simonson, R. W. (1979). Origin of the name “Ando soils”. Geoderma. 22, 333– 335. Soil Survey Staff, (1960). “Soil Classiﬁcation, a Comprehensive System, 7th Approximation.” U.S. Dept. Agric., Soil Conservation Service, Washington D.C. Soil Survey Staff, (1975). “Soil Taxonomy: A Basic System of Soil Classiﬁcation for Making and Interpreting Soil Surveys,”. USDA Agric. Handb. No. 436. U.S. Govt. Printing Ofﬁce, Washington, DC. Soil Survey Staff, (1998). “Keys to Soil Taxonomy”, 8th ed. USDA Natural Resources Conservation Service, U.S. Govt. Printing Ofﬁce, Washington DC. Soma, M., Churchman, G. J., and Theng, B. K. G. (1992). X-ray photoelectron spectroscopic analysis of halloysites with different composition and particle morphology. Clay Miner. 27, 413–421. Southard, R. J., and Southard, S. B. (1987). Sand-sized kaolinized feldspar pseudomorphs in a California Humult. Soil Sci. Soc. Am. J. 51, 1666–1672. Stanford, G., and Smith, S. J. (1972). Nitrogen mineralization potentials of soils. Soil Sci. Soc. Am. Proc. 36, 465– 472. Stumm, W. (1992). “Chemistry of the Solid–Water Interface.”, p. 428. Wiley, New York. Sudo, T., and Nakamura, T. (1952). Hisingerite from Japan. Am. Mineral. 37, 618 –621. Sugihara, S., Konno, T., and Ishii, K. (1986). Kinetics of mineralization of organic nitrogen in soil. Bull. Nat. Inst. Agro-Environ. Sci. 1, 127 –166. Sverdrup, H. U. (1990). “The Kinetics of Base Cation Release Due To Chemical Weathering”, p. 246. Lund University Press, Lund, Sweden. Sverdrup, H., and Warfvinge, P. (1988). Weathering of primary silicate minerals in the natural soil environment in relation to a chemical weathering model. Water, Air, Soil Pollut. 38, 387–408. Tachibana, N., Yoshikawa, S., and Ikeda, K. (1995). Inﬂuences of heavy application of nitrogen on soil acidiﬁcation and root growth in tea ﬁelds. Jpn. J. Crop Sci. 64, 516–522. Takahashi, T., and Dahlgren, R. A. (1998). Possible control of aluminum solubility by 1 M KCl treatment in some soils dominated by aluminum– humus complexes. Soil Sci. Plant Nutr. 44, 43 –51.

180

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Takahashi, T., Shoji, S., and Sato, A. (1989). Clayey Spodosols and Andisols showing a biosequential relation from Shimokita Peninsula, northeastern Japan. Soil Sci. 148, 204 –218. Takahashi, T., Dahlgren, R. A., and van Susteren, P. (1993). Clay mineralogy and chemistry of soils formed in volcanic materials in the xeric moisture regime of northern California. Geoderma. 59, 131–150. Takahashi, T., Fukuoka, T., and Dahlgren, R. A. (1995). Aluminum solubility and release rates from soil horizons dominated by aluminum–humus complexes. Soil Sci. Plant Nutr. 41, 119–131. Takahashi, T., Dahlgren, R. A., Theng, B. K. G., Whitton, J. S., and Soma, M. (2001). Potassiumselective, halloysite-rich soils formed in volcanic materials from northern California. Soil Sci. Soc. Am. J. 65, 516–526. Talibudeen, O. (1981). Precipitation. In “The Chemistry of Soil Processes” (D. J. Greenland and M. H. B. Hayes, Eds.), pp. 81–114. Wiley, New York. Tate, K. R., and Theng, B. K. G. (1980). Organic matter and its interactions with inorganic soil constituents. In “Soils with Variable Charge” (B. K. G. Theng, Ed.), pp. 225 –249. Soil Bureau, Lower Hutt, New Zealand. Tawaraya, K., Kunii, Y., and Wagatsuma, T. (1995). Effects of arbuscular mycorrhizal inoculation and phosphate application on phosphorus uptake and growth of white clover (Trifolium repens L.) and onion (Allium cepa L.) in Andosols. Jpn. J. Soil Sci. Plant Nutr. 66, 48–53. Theng, B. K. G. (1980). “Soils with Variable Charge”, p. 448. Soil Bureau, Lower Hutt, New Zealand. Tisdall, J. M., and Oades, J. M. (1982). Organic matter and water-stable aggregates in soils. J. Soil Sci. 33, 141 –163. Toda, T., Koyama, H., Hori, T., and Hara, T. (1999). Aluminum tolerance of Arabidopsis thaliana under hydroponic and soil culture conditions. Soil Sci. Plant Nutr. 45, 419–425. Tokashiki, T., and Wada, K. (1975). Weathering implications of the mineralogy of clay fractions of two Ando soils, Kyushu. Geoderma. 14, 47–62. Toma, M., and Saigusa, M. (1997). Effects of phosphogypsum on amelioration of strongly acid nonallophanic Andosols. Plant Soil. 192, 49– 55. Toma, M., Hiradate, S., and Saigusa, M. (1999). Chemical species of Al in a gypsum-treated Kitakami Andosol. Soil Sci. Plant Nutr. 45, 279 –285. Torn, M. S., Trumbore, S. E., Chadwick, O. A., Vitousek, P. M., and Hendricks, D. M. (1997). Mineral control of soil organic carbon storage and turnover. Nature. 389, 170–173. Towe, K. M., and Bradley, W. F. (1967). Mineralogical constitution of colloidal “hydrous ferric oxides.” J. Colloid Interface Sci. 24, 384–392. Tsuji, H., Yamamoto, H., Matsuo, K., and Usuki, K. (2000). Effect of nontilled cropping on drought injury of upland rice. Jpn. J. Crop Sci. 69, 20–23. Ugolini, F. C. (1968). Soil development and alder invasion in a recently deglaciated area of Glacier Bay, Alaska. In “Biology of the Alder” (G. M. Trappe, J. F. Franklin, R. F. Tarrant, and G. M. Hansen, Eds.), pp. 115–140. U.S. Forest Service, Paciﬁc Northwest Forest and Range Experiment Station, Portland, Oregon. Ugolini, F. C., and Dahlgren, R. A. (1987). The mechanism of podzolization as revealed by soil solution studies. In “Podzols et Podzolization” (D. Righi and A. Chauvel, Eds.), pp. 195– 203. Assoc. Fr. Estude Sol, Plasisier, France. Ugolini, F. C., and Dahlgren, R. A. (1991). Weathering environments and occurrence of imogolite/ allophane in selected Andisols and Spodosols. Soil Sci. Soc. Am. J. 55, 1166–1171. Ugolini, F. C., and Dahlgren, R. A. (2002). Soil development in volcanic ash. Global Environ. Res. 6, 69–81. Ugolini, F. C., and Zasoski, R. J. (1979). Soils derived from tephra. In “Volcanic Activity and Human Ecology” (P. D. Sheets and K. K. Grayson, Eds.), pp. 83–124. Academic Press, New York. Ugolini, F. C., Dahlgren, R. A., Shoji, S., and Ito, T. (1988). An example of andosolization and podzolization as revealed by soil solution studies, S. Hakkoda, N.E. Japan. Soil Sci. 145, 111–125.

NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 181 Ugolini, F. C., Dahlgren, R. A., LaManna, J., Nuhn, W., and Zachara, J. (1991). Mineralogy and weathering processes in recent and Holocene tephra deposits of the Paciﬁc Northwest, USA. Geoderma. 51, 277 –299. Ugolini, F. C., Minden, R., Dawson, H., and Zachara, J. (1977). An example of soil processes in the Abies amabilis zone of central Cascades, Washington. Soil Sci. 124, 291– 302. Vaje, P. I., Singh, B. R., and Lal, R. (2000). Leaching and plant uptake of nitrogen from a volcanic ash soil in Kilimanjaro Region, Tanzania. J. Sustain. Agric. 16, 95–112. Van Olphen, H. (1971). Amorphous clay materials. Science. 171, 90–91. Van Ranst, E., Utami, S. R., and Shamshuddin, J. (2002). Andisols on volcanic ash from Java Island, Indonesia: physico-chemical properties and classiﬁcation. Soil Sci. 167, 68–79. Wada, K. (1977). Active aluminum in Kuroboku soils and non- and para-crystalline clay minerals. Nendo Kagaku. 17, 143 –151. Wada, K. (1980). Mineralogical characteristics of Andisols. In “Soils with Variable Charge” (B. K. G. Theng, Ed.), pp. 87 –107. New Zealand Society of Soil Science, Lower Hutt. Wada, K. (1985). The distinctive properties of Andosols. Adv. Soil Sci. 2, 173 –229. Wada, K. (1986). “Ando Soils in Japan”, p. 276. Kyushu University Press, Fukuoka, Japan. Wada, K. (1989). Allophane and imogolite. In “Minerals in Soil Environments, 2nd edn.” (J. B. Dixon and S. B. Weed, Eds.), pp. 1051–1087. Soil Sci. Soc. Am., Madison, WI. Wada, K., and Harward, M. E. (1974). Amorphous clay constituents of soils. Adv. Agron. 26, 211 –260. Wada, K., and Higashi, T. (1976). The categories of aluminum– and iron –humus complexes in Ando soils determined by selective dissolution. J. Soil Sci. 27, 357– 368. Wada, S.-I. (1987a). Critical evaluation of 1 M KCl-extraction method for determining exchangeable Al ions in variable charge soils. Soil Sci. Plant Nutr. 33, 153–160. Wada, S.-I. (1987b). Adsorption of Al(III) on allophane, imogolite, goethite, and noncrystalline silica and the extractability of the adsorbed Al(III) in 1 M KCl solution. Soil Sci. Plant Nutr. 33, 487 –491. ˚ ) with crumpled lamellar morphology Wada, S.-I., and Mizota, C. (1982). Iron-rich halloysite (10A from Hokkaido, Japan. Clays Clay Miner. 30, 315–317. Wagatsuma, T., and Ezoe, Y. (1985). Effect of pH on ionic species of aluminum in medium and on aluminum toxicity under solution culture. Soil Sci. Plant Nutr. 31, 547 –561. Wakatsuki, T., and Wielemaker, W. G. (1985). Clay mineralogy of a soil formed in peralkaline volcanic ash from the Great Rift Valley in Kenya. Soil Sci. Plant Nutr. 31, 475 –480. Warkentin, B. P., and Maeda, T. (1980). Physical and mechanical characteristics of Andosols. In “Soils with Variable Charge” (B. K. G. Theng, Ed.), pp. 97 –107. N.Z. Soil Bureau, Lower Hutt, New Zealand. Warkentin, B. P., Maeda, T., and Soma, K. (1988). Physical characteristics for classiﬁcation of Andisols. In “Proceedings 9th Int. Soil Classiﬁcation Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 97 –107. Japanese Committee for 9th International Soil Classiﬁcation Workshop and Soil Management Support Services, Washington, DC. Wells, N., Childs, C. W., and Downes, C. J. (1977). Silica Springs, Tongariro National Park, New Zealand: analyses of the spring water and characterization of the alumio-silicate deposit. Geochim. Cosmochim. Acta. 41, 1497– 1506. White, A. F. (1983). Surface chemistry and dissolution kinetics of glassy rocks at 258C. Geochim. Cosmochim. Acta. 47, 805– 815. White, A. F., and Claassen, H. C. (1980). Kinetic model for short-term dissolution of a rhyolitic glass. Chem. Geol. 29, 91 –109. White, A. F., Benson, L. V., and Yee, A. (1986). Chemical weathering of the May 18, 1980 Mount St. Helens ash fall and the effect on the Iron Creek Watershed, Washington. In “Rates of Chemical Weathering of Rocks and Minerals” (S. M. Colman and D. P. Detheir, Eds.), pp. 351–375. Academic Press, Orlando.

182

R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI

Wielemaker, W. G., and Wakatsuki, T. (1984). Properties, weathering and classiﬁcation of some soils formed in peralkaline volcanic ash in Kenya. Geoderma. 32, 21 –33. Wilding, L. P. (2000). Classiﬁcation of soils. In “Handbook of Soil Science” (M. E. Sumner, Ed.), pp. E175–183. CRC Press, Boca Raton, FL. Wilson, M. A., Burt, R., Sobecki, T. M., Engel, R. J., and Hipple, K. (1996). Soil properties and genesis of pans in till-derived Andisols, Olympic Peninsula, Washington. Soil Sci. Soc. Am. J. 60, 206–218. Wilson, M. A., Lee, G. S. H., and Taylor, R. C. (2001). Tetrahedral rehydration during imogolite formation. J. Non-Crystall. Solids. 296, 172– 181. Yamada, I. (1988). Tephra as parent material. In “Proceedings 9th Int. Soil Classiﬁcation Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 509–519. Japanese Committee for 9th International Soil Classiﬁcation Workshop and Soil Management Support Services, Washington, DC. Yamada, I., and Shoji, S. (1980). Fertility of Andosols formed from basaltic and dacitic volcanic ashes. Tohoku J. Agri. Res. 31, 149– 155. Yamada, I., Saigusa, M., and Shoji, S. (1978). Clay mineralogy of Hijiori and Numazawa Ando soils. Soil Sci. Plant Nutr. 24, 75–89. Yamanaka, K. (1964). Physical properties. In “Volcanic Ash Soils in Japan” (Ministry of Agriculture and Forestry, Ed.), pp. 69–91. Japanese Govt, Tokyo. Yerima, B. P. K., Wilding, L. P., Calhoun, F. G., and Hallmark, C. T. (1987). Volcanic ash-inﬂuenced Vertisols and associated Mollisols of El Salvador–physical, chemical, and morphological properties. Soil Sci. Soc. Am. J. 51, 699–708. Yoshinaga, N. (1983). “Volcanic Ash Soil — Genesis, Properties, Classiﬁcation”, Hakuyuisha, Tokyo. Yoshinaga, N. (1988). Mineralogy of Andisols. In “Proceedings 9th Int. Soil Classiﬁcation Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 45–59. Japanese Committee for 9th International Soil Classiﬁcation Workshop and Soil Management Support Services, Washington, DC. Yoshinaga, N., and Aomine, S. (1962). Imogolite in some Ando soils. Soil Sci. Plant Nutr. 8, 22 –29. Zou, X. M., Binkley, D., and Doxtader, K. G. (1992). A new method for estimating gross phosphorus mineralization and immobilization rates in soils. Plant Soil. 147, 243–250. Zunino, H., Borie, F., Aguilera, S., Martin, J. P., and Haider, K. (1982). Decomposition of 14C-labeled glucose, plant and microbial products and phenols in volcanic ash-derived soils of Chile. Soil Biol. Biochem. 14, 37–43.

MICROARRAY TECHNOLOGY AND APPLICATIONS IN ENVIRONMENTAL MICROBIOLOGY Jizhong Zhou and Dorothea K. Thompson Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA

I. Introduction II. Microarray Types and Advantages A. Types of Microarrays B. Advantages of Microarrays III. Microarray Fabrication A. Microarray Substrates B. Surface Modiﬁcation for the Attachment of Nucleic Acids C. Arraying Technology D. Critical Issues for Microarray Fabrication IV. Microarray Hybridization and Detection A. Probe Design and Synthesis B. Target Labeling and Quality C. Hybridization D. Detection E. Critical Issues in Hybridization and Detection V. Microarray Image Processing A. Data Acquisition B. Assessment of Spot Quality and Background Subtraction VI. Microarray Data Analysis A. Data Normalization B. Data Transformation C. Methods for Identifying Differentially Expressed Genes D. Microarray Data Analysis VII. Using Microarrays to Monitor Genomic Expression A. General Approaches to Revealing Differences in Gene Expression B. Experimental Design for Microarray-based Monitoring of Gene Expression C. Microarray-based Functional Analysis of Environmental Microorganisms VIII. Application of Microarrays to Environmental Studies A. Functional Gene Arrays B. Phylogenetic Oligonucleotide Arrays C. Community Genome Arrays 183 Advances in Agronomy, Volume 82 Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved. 0065-2113/03 $35.00

184

J. ZHOU AND D. K. THOMPSON D. Whole-genome Open Reading Frame Arrays for Revealing Genome Differences and Relatedness E. Other Types of Microarrays for Microbial Detection and Characterization IX. Concluding Remarks References

I. INTRODUCTION Microarrays are miniaturized arrays of hundreds to thousands of discrete DNA fragments or synthetic oligonucleotides that have been attached to a solid substrate (e.g., glass) using automated printing equipment such that each spot (element) in a ﬁxed position on the array corresponds to a unique DNA (Schena et al., 1998; Schena, 2003). Microarrays are variously referred to as microchips, biochips, DNA chips, or gene chips and have emerged as a widely accepted functional genomics technology for large-scale genomic analysis. In particular, DNA or oligonucleotide arrays have been used to monitor messenger RNA (mRNA or transcript) abundance levels of differentially expressed genes under different cell growth conditions or in response to environmental perturbations or genetic mutations (c.f., Lockhart et al., 1996; Schena et al., 1996; DeRisi et al., 1997; Wodicka et al., 1997; Richmond et al., 1999; Ye et al., 2000; Thompson et al., 2002; Liu et al., 2003) and to detect speciﬁc mutations in DNA sequences (Hacia, 1999; Broude et al., 2001). Recently, the potential research applications of microarray technology to studies in microbial ecology have been explored (Zhou and Thompson, 2002; Zhou, 2003). In principle and practice, microarrays are extensions of conventional membrane-based Southern and Northern hybridization blots, which have been used for decades to detect and characterize nucleic acids in diverse biological samples. Microarray hybridization is based on the association of a singlestranded molecule labeled with a ﬂuorescent tag, or ﬂuorescein, with its complementary molecule, which is covalently attached or immobilized to a solid support, usually glass. In such an assay, the speciﬁc hybridization pattern or gene expression proﬁle generated by an unknown (experimental) sample is typically compared with a control (reference) pattern. In microarray terminology, the ﬂuorescein-labeled DNA in solution is generally termed the target, and the DNA strand immobilized on the microarray surface is referred to as the probe. Because the sequence of the arrayed molecule is usually known, it is used to ‘probe’ or investigate the unknown target molecule in solution. This is directly opposite to the convention established with the development of Southern blot hybridization,

MICROARRAY TECHNOLOGY

185

in which target molecules ﬁxed to a porous membrane are interrogated by known solution-phase probes. The concept of microarrays was ﬁrst proposed in the late 1980s. One of the ﬁrst descriptions of DNA microarrays in the literature was provided by Augenlicht and his colleagues, who spotted 4000 complementary DNA (cDNA) sequences on nitrocellulose and used radioactive labeling to analyze differences in gene expression patterns among different types of colon tumors in various stages of malignancy (Augenlicht et al., 1987, 1991). At the same time, four separate research groups simultaneously developed the concept of determining a DNA sequence by hybridization to a comprehensive set of oligonucleotides, i.e., sequencing by hybridization or SBH (Bains and Smith, 1988; Drmanac et al., 1989; Khrapko et al., 1989; Southern et al., 1992). Although SBH is an extremely elegant alternative to conventional DNA sequencing, various inherent problems associated with repeated sequences and the imperfect speciﬁcity of hybridization limit the practicality of using SBH for routine sequence determination. Such technical challenges, therefore, have led researchers to focus on the more readily addressable applications of microarray technology, such as gene expression proﬁling. By the mid-1990s, the reverse dot-blot scheme for monitoring genomic expression was reformulated by several different groups. Both DNA fragments and synthetic oligonucleotides were arrayed on various substrates, including nylon membranes, plastic and glass (Schena et al., 1995; Lockhart et al., 1996). All of them depended on sequence-speciﬁc hybridization between the arrayed DNA and the labeled nucleic acids from cellular mRNA. Later studies using the simple eukaryote, yeast, clearly demonstrated that DNA and oligonucleotide arrays are powerful tools for monitoring global gene expression (DeRisi et al., 1997; Wodicka et al., 1997). Microarray-based genomic technology has greatly beneﬁted from many parallel advances in other ﬁelds. Without such advancements, the development of high-density microarrays and the various applications that we see today would not be possible (Eisen and Brown, 1999; Schena and Davis, 2000). For example, large-scale genome sequencing projects have produced the raw sequence information needed for microarray expression proﬁling, and the development of robotic printers or arrayers has made it possible to fabricate high-density microarrays in a very small area. In addition, recent advances in methods of ﬂuorescent labeling and detection offer signiﬁcant advantages in speed, data quality, and user safety for microarray-based assays. Together, these technical advancements have enabled microarray-based genomic technologies to revolutionize genetic analyses of biological systems. The widespread, routine use of such genomic technologies will shed light on a wide range of important research questions associated with the genetic programs controlling cell growth and differentiation, bacterial pathogenesis and the host response to infection, antibiotic resistance, specialized metabolic capabilities of microorganisms of bioremediation potential, as well as agricultural and pharmaceutical applications.

186

J. ZHOU AND D. K. THOMPSON

In this chapter, we review the technical underpinnings of microarray hybridization and its applications in environmental microbiology, with emphasis on the most important issues of microarray-based assays as outlined in Fig. 1. We ﬁrst describe the technologies used for microarray fabrication, followed by a discussion on microarray hybridization, ﬂuorescence detection technologies, image processing, and data analysis. In addition, we provide an overview of the recent applications of microarray technologies to study gene expression patterns in environmentally important microorganisms. Finally, we describe various types of microarrays speciﬁcally developed for analyzing microbial community composition and function in natural environments. Because glass-based DNA microarrays are currently preferred by most basic research laboratories, our discussion of microarray technologies will focus on this type of array, while other types will be mentioned only brieﬂy. It should be noted that the goal of this chapter is to provide an in-depth description of the basis and principles of

Figure 1

A ﬂow chart of a microarray experiment.

MICROARRAY TECHNOLOGY

187

microarray-based technologies rather than an exhaustive review of current microarray technology.

II. MICROARRAY TYPES AND ADVANTAGES A. TYPES

OF

MICROARRAYS

Microarrays can be divided into two major formats based on the type of immobilized probes: (1) DNA microarrays constructed with DNA fragments typically generated using the polymerase chain reaction (PCR) (Schena et al., 1995; DeRisi et al., 1997; Marshall and Hodgson, 1998); and (2) oligonucleotide microarrays constructed with shorter (10- to 40-mer) or longer (up to 120-mer) oligonucleotide sequences that are designed to be complementary to speciﬁc coding regions of interest. DNA microarrays have certain advantages over oligonucleotide microarrays, especially for monitoring gene expression patterns. While oligonucleotide microarrays are limited to array elements of low sequence complexity, the speciﬁcity of hybridization for a complex probe is improved with arrays containing DNA fragments that are substantially longer than oligonucleotides (Shalon et al., 1996). In addition, oligonucleotide synthesis requires prior sequence knowledge, whereas DNA arrays do not because DNA fragments of unknown sequences can be ampliﬁed from clones using vector-speciﬁc primers. For microarrays constructed with PCR-ampliﬁed DNA elements, nucleic acids of virtually any length, composition, or origin can be arrayed (Shalon et al., 1996). However, oligonucleotide-based microarrays have the advantage of minimizing the potentially confounding effects of occasional cross-hybridization (Wodicka et al., 1997) and are uniquely suited for detecting genetic mutations and polymorphisms. Since oligonucleotide probes can be commercially synthesized, the handling and tracking of oligonucleotide array elements, unlike PCR products, is generally easier. Amplifying all of the probes with a desired minimum quantity for printing is labor intensive and time-consuming. Based on probe immobilization and fabrication strategies, there are two general types of oligonucleotide microarrays: 1. Direct parallel synthesis on solid substrates by light-directed or photoactivatable chemistries (Pease et al., 1994; Lipshultz et al., 1999) or standard phosphoramidite chemistries (Southern et al., 1994); or 2. Chemical attachment of pre-made oligonucleotides to solid supports (Khrapko et al., 1989; Beattie et al., 1992, 1995; Eggers et al., 1994; Lamture et al., 1994; Fotin et al., 1998; Guschin et al., 1997a, b; Rehman et al., 1999; Rogers et al., 1999).

188

J. ZHOU AND D. K. THOMPSON

Each strategy for oligonucleotide immobilization has its own set of speciﬁc advantages and disadvantages (Schena et al., 1998; Hoheisel, 1997). The direct or in situ synthesis approach has two major advantages. First, the photoprotected versions of the four DNA bases allow microarrays to be manufactured directly from DNA sequence databases, thereby removing the uncertain and burdensome aspects of sample handling and tracking. Second, the use of synthetic reagents minimizes variations between arrays by ensuring a high degree of precision in each coupling cycle. Costliness is a major disadvantage of the photolithographic approach; photomasks, which direct light to speciﬁc areas on the array for localized chemical synthesis, are very expensive and timeconsuming to design and build. Also, the yield and length of the synthesized oligonucleotides are subject to wide variation and uncertainty, which could lead to unpredictable effects on hybridization across the microarray. A major advantage of the attachment of pre-synthesized probes is that the concentration and length of each oligonucleotide on the array can be controlled prior to immobilization. Standard synthesis chemistry is also well established for many nucleotide derivatives for which no light-inducible monomer equivalents are available. In addition, the post-synthesis approach is less complicated and can be customized according to the speciﬁc needs of the laboratory. The critical drawback of the post-synthesis approach, however, is still the need for the external synthesis and storage of different oligonucleotides prior to array fabrication.

B. ADVANTAGES

OF

MICROARRAYS

Microarrays offer the following advantages over conventional nucleic acidbased approaches.

1. High-throughput and Parallel Analysis Microarray technology allows thousands to hundreds of thousands of array elements or probes to be uniformly deposited in a very small area on the surface of a non-porous substrate. Consequently, the high-density capacity of microarrays permits parallel analysis, in which the expression of the entire gene content of a genome of interest can be monitored, or many constituents of a microbial community can be simultaneously assessed in a single assay using the same microarray. Genomic expression data allows researchers to begin to build a comprehensive, integrated view of a complex biological system.

MICROARRAY TECHNOLOGY

2.

189

High Sensitivity

High sensitivity can be achieved in probe-target hybridization, because microarray hybridization uses a very small volume of probe and the target nucleic acid is restricted to a small area (Shalon et al., 1996; Guschin et al., 1997a). This feature enables high sample concentrations and rapid hybridization kinetics.

3.

Differential Display

Multi-ﬂuorescence detection schemes allow differential display of different biological samples. Different target samples, for example, can be labeled with different ﬂuorescent tags and then hybridized in parallel to the same microarray, allowing the simultaneous analysis of two or more biological samples in a single assay. Multi-color hybridization detection minimizes variations resulting from inconsistent experimental conditions and allows direct and quantitative comparison of target sequence abundance among different biological samples (Shalon et al., 1996; Ramsay, 1998).

4.

Low Background Signal Noise

Non-porous surfaces substantially reduce the amount of non-speciﬁc hybridization; as a result, organic and ﬂuorescent compounds that attach to microarrays during fabrication and hybridization procedures can be rapidly removed by post-hybridization washing, resulting in considerably less background signal noise than is typically encountered with porous membranes (Shalon et al., 1996).

5. Real-time Data Analysis Once the microarrays are constructed, hybridization and detection are relatively simple and rapid, allowing real-time data analysis in ﬁeld-scale heterogeneous environments.

6. Automation Microarray technology is amenable to automation and therefore, has the potential of being cost-effective compared to traditional hybridization methods (Shalon et al., 1996).

190

J. ZHOU AND D. K. THOMPSON

III. MICROARRAY FABRICATION A. MICROARRAY SUBSTRATES The substrate used for printing microarrays has a large impact ultimately on the quality of the data obtained from microarray hybridizations. Poor surface treatment may result in poor attachment of the DNA probes to the slide, and a non-uniform surface will cause variations in the amount of the attached DNA. Furthermore, residual substances deposited on the slide surface during a microarray experiment can lead to high background ﬂuorescence or noise. Thus, the selection of appropriate substrates for microarray experiments is of critical importance. The substrates used for fabricating microarrays fall into two categories: porous and non-porous.

1.

Non-porous Substrate

At present, a non-porous solid surface is the most common type of substrate used for printing arrays. Several non-porous materials, such as glass and polypropylene, are suitable for microarray fabrication. Glass slides are the most widely used substrates, because they are inexpensive, possess physical characteristics advantageous to hybridization, and are easily modiﬁed for nucleic acid attachment and synthesis (Southern, 2001). In general, non-porous substrates offer a number of advantages over porous substrates (Schena and Davis, 2000). First, small amounts of molecules may be deposited at precise, predeﬁned positions on the substrate surface with little diffusion, thus enabling the highdensity capacity of microarrays. Second, hybridization between target and probe molecules occurs at a much faster rate on non-porous solid surfaces than on porous substrates. This is because molecules do not have to diffuse in and out of the pores and as a result, steric inhibition of hybridization is not a problem (Southern, 2001). Third, because small sample volumes can be applied to a nonporous surface under a coverslip, high probe concentrations, rapid hybridization kinetics, and high sensitivity can be achieved. Fourth, non-porous substrates prevent the absorption of reagents and samples into pores, thus allowing unbound labeled materials to be easily removed. This expedites the procedure, improves reproducibility and reduces background. Fifth, a non-porous substrate has low intrinsic ﬂuorescence and thus allows the use of ﬂuorescence detection. Finally, a solid substrate offers a homogeneous attachment surface, and its inherent uniform ﬂatness permits true parallel analysis. However, a major drawback of non-porous substrates, such as glass, is the susceptibility to dust and other airborne particle contamination, which can cause a scanned slide to appear “dirty”. Poor modiﬁcation of microarray slides is another common cause of poor hybridization

MICROARRAY TECHNOLOGY

191

results (Southern, 2001). In addition, because of the planar surface, the capacity for immobilization is limited, and consequently, the sensitivity of the assay is relatively low compared to that of the porous substrates (Afanassiev et al., 2000). 2.

Porous Substrates

Porous substrates such as nitrocellulose and nylon membranes have also been used for microarray fabrication (Englert, 2000). The principal advantage of membranes is that larger volumes and concentrations of samples can be immobilized on a small area, because the pores of the substrates provide a larger total surface area for binding. As a result, a relatively higher sensitivity and better dynamic range for quantitative comparison can be achieved. Homogeneous spots are also more readily obtained, because deposited samples are able to distribute immediately into the membrane through capillary ﬂow. In addition, membranebased microarrays can be reused several times (Beier and Hoheisel, 1999). However, there are some important disadvantages in using porous membranes for microarray fabrication. The boundaries and shapes of the spots are poorly deﬁned, and membranes swell in solvent, and shrink and distort when dried. Such fragility and ﬂexibility make it difﬁcult to precisely locate probe positions during spotting and image analysis. Also, many membranes have high intrinsic ﬂuorescence and thus higher background noise compared to non-porous substrates. In addition, because the spot sizes on a membrane cannot be reduced to a level comparable with glass slides or other non-porous substrates, much more DNA is required for producing a membrane-based microarray (Beier and Hoheisel, 1999). Overall, non-porous substrates are preferred for microarray experiments, even though porous substrates have some advantages. This is because the unique physical and chemical characteristics of glass slides (e.g., little diffusion, low intrinsic ﬂuorescence) allow miniaturization and use of ﬂuorescent labeling and detection, which are the most critical requirements for large-scale genomic analysis. The miniaturized microarray format coupled with ﬂuorescent detection represents a fundamental revolution in biological analysis (Schena and Davis, 2000).

B. SURFACE MODIFICATION FOR THE ATTACHMENT OF NUCLEIC ACIDS 1.

Attachment Strategies

Appropriate attachment and retainment of nucleic acid probes to an array surface is very important for microarray analysis. For reliable microarray

192

J. ZHOU AND D. K. THOMPSON

hybridization, the attachment chemistry must meet the following criteria: (i) nucleic acids must be tightly (covalently) bound to the array surface; (ii) the surface-bound molecules must be accessible for hybridization; and (iii) the attachment chemistry must be reproducible. Both ionic interaction and covalent bonding, which are described in more detail below, are used for attaching nucleic acids to solid surfaces, depending on the size of the nucleic acid molecules. Electrostatic interaction. Long DNA fragments (in the order of several hundred bases in length) can be immobilized on a glass surface through ionic interaction between the negatively charged phosphodiester backbone and the positively charged slide surface (Fig. 2A). Recent studies showed that synthetic oligonucleotides more than 70 bp in length can also be bound to glass surfaces through ionic interaction (Hughes et al., 2001). Generally, an amine or lysine coating is used to adsorb DNA to glass slides. Because amines have a positive charge at neutral pH, they allow attachment of native DNA molecules through the formation of ionic bonds with the negatively charged phosphate backbone. Electrostatic attachment can be enhanced by exposing the fabricated arrays to ultraviolet light or heat, which induces free radical-based coupling between thymidine residues in the DNA and carbons on the alkyl amine. The combination of electrostatic bonding and non-speciﬁc covalent attachment, links native DNA to the substrate surface in a stable manner. Although ionic interaction-mediated attachment is less expensive and more versatile than covalent bonding (Worley et al., 2000), the immobilized DNA molecules are susceptible to removal under high salt and/or high temperature conditions. Therefore, covalent binding methods are preferred. Covalent bonding. DNA can also be covalently attached to glass surfaces using different attachment chemistries (Fig. 2B). Although long DNA molecules can be attached covalently to the microarray surface by different methods, immobilization of aminated DNA to an aldehyde-coated slide is the usual method of choice (Zammatteo et al., 2000). Because oligonucleotides are typically short, covalent bonding is generally required for attachment of such molecules to a glass surface. Usually, oligonucleotides are ﬁxed covalently onto solid surfaces at one end of the molecule using a variety of methods. The attachment of biomolecules to a solid phase presents some problems that are unique to homogeneous solutions. Because the bound probe is not free to diffuse, a lower reaction rate is expected. In addition, target molecules in solution may not be able to effectively interact with the bound probes due to steric hindrance from the solid support and the close proximity of other bound probes (Shchepinov et al., 1997). Additional molecules, termed linkers or spacers, are used to tether the probe to the substrate surface, thereby providing a sufﬁcient amount of distance between the

MICROARRAY TECHNOLOGY

193

Figure 2 Attachment strategies. (A) Attachment of nucleic acids to solid surfaces through electrostatic interactions. The microarray substrates contain primary amine groups (NHþ 3 ) attached covalently to the glass surface. The amines carry a positive charge at neutral pH, which permits attachment of native DNA through the formation of ionic bonds with the negatively charged phosphate backbone. Covalent attachment of DNA to the surface can be further achieved by treatment with ultraviolet light or heat. (B) Attachment of nucleic acids to solid surfaces through covalent bonding. The microarray substrates contain primary aldehyde groups attached covalently to the glass surface. Primary amino linkers (NH2) on the DNA attack the aldehyde groups to form covalent bonds. Such attachment is stabilized with a dehydration reaction by drying in low humidity, which leads to Schiff base covalent bond formation.

oligonucleotide probe and the support to minimize steric interference. To serve as an effective linker, the molecule must meet several criteria (Guo et al., 1994). First, the linkage must be chemically stable under the hybridization conditions used and must be sufﬁciently long to minimize steric interference. Second, the linker should be sufﬁciently hydrophilic to be freely soluble in aqueous solution. Third, there should be no non-speciﬁc binding of the linker to the support. Shchepinov et al. (1997) showed that the optimal linker for immobilizing oligonucleotides at either the 50 or 30 terminus should have low negative charge density and a length of 30– 60 atoms.

194

J. ZHOU AND D. K. THOMPSON

Generally, a linker is coupled to the probe during oligonucleotide synthesis. Various types of linkers have been used, including poly-dT (Guo et al., 1994; Rehman et al., 1999; Afanassiev et al., 2000), poly carbon atoms (Afanassiev et al., 2000), and oligodeoxyribonucleotides with hairpin stem-loop structures (Zhao et al., 2001). Studies have shown that the length of the spacer has a signiﬁcant impact on the success of hybridization. While virtually no hybridization signal was observed for poly-dT spacers (15 bp), about a 20-fold enhancement of hybridization was obtained for a poly-dT spacer of 15 nucleotides (Guo et al., 1994). The effect of linkers composed of multiple carbon atoms (e.g., C36, C18, C12, and C6) on microarray hybridization was also examined (Afanassiev et al., 2000). Overall, the signal intensity was improved with a longer C linker. A spacer can also be added by chemically modifying the slide surface (Guo et al., 1994; Beier and Hoheisel, 1999). Oligonucleotides can be immobilized onto solid supports through homobifunctional or hetero-biofunctional cross-linkers. For example, amino-modiﬁed oligonucleotides can be covalently attached to glass surfaces containing amine functional groups through homo-biofunctional cross-linkers (Guo et al., 1994) and to glass surfaces containing aldehyde and epoxide through heterobiofunctional cross-linkers (Lamture et al., 1994; Schena et al., 1996). Thiolmodiﬁed or disulﬁde-modiﬁed oligonucleotides can be attached to the glass surface containing amine functional groups via hetero-biofunctional crosslinkers (Chrisey et al., 1996). A hetero-biofunctional cross-linker is generally preferred over a homo-biofunctional cross-linker to prevent surfaceto-surface linkages and probe-to-probe linkages as opposed to the desired surface-to-probe linkages (Steel et al., 2000). When using a hetero-biofunctional cross-linker, the probe should have a different modiﬁcation chemistry from the array surface. While microscope slides made from low-ﬂuorescence glass are suitable substrates for microarray construction, the glass surface must be modiﬁed with a chemical coating and cleaned (e.g., free of dust particles) before use. The glass surface must have suitable functional groups for the attachment of target DNA molecules, because DNA does not inherently bind to untreated glass. A hydrophobic surface is essential for achieving high-density spots, because spotted hydrophilic samples will spread less on a hydrophobic surface than on an untreated hydrophilic glass surface (Rose, 2000). The quality of slide coating ultimately impacts the quality of the microarray data. A poor surface coating can result in poor probe retention. For spot-to-spot consistency, the coating must be uniform and homogeneous without untreated patches. In addition, the coating must be non-ﬂuorescent and capable of resisting harsh physical conditions, such as boiling, baking and soaking. Silanization, dendrimeric linker coating, gel coating, and nitrocellulose coating are types of glass surface modiﬁcations that are discussed in greater detail in the following sections.

MICROARRAY TECHNOLOGY

195

2. Silanization Silanes are most commonly used for slide surface modiﬁcation to provide organic functional groups for the covalent attachment of biomolecules (Shriver-Lake, 1998), and silanized glass slides are commercially available from a number of companies. Glass slides can be modiﬁed to contain surface hydroxyls that react with methoxy or ethoxy residues of a silane molecule. Many different commercially available silanes contain various functional reactive groups such as amino, epoxide, carboxylic acid and aldehyde, which are suitable for covalent bonding with appropriately modiﬁed biomolecules (Schena, 2003). Most microarray analyses are performed with slide surfaces that contain reactive amine and aldehyde groups (Schena, 2003), which allow attachment of biomolecules via electrostatic interactions or covalent bonding. Silanization can be accomplished simply by immersing the slides into a silane-containing solution or by vapor deposition (Steel et al., 2000; Worley et al., 2000). Vapor-phase coating is most effective at uniformly depositing a monolayer of silane on the slide surface (Chrisey et al., 1996; Worley et al., 2000).

3.

Dendrimeric Linker Coating

To increase the binding capacity of arrayed probes, a more elaborate chemistry was developed for synthesizing dendrimeric linkers on silanized glass slides to allow covalent attachment of aminated DNA molecules (Beier and Hoheisel, 1999). Such a linker system multiplies the coupling sites by introducing additional reactive groups through branched linker molecules. There are several advantages of this linker system. First, it allows covalent immobilization of both pre-synthesized and in situ synthesized oligonucleotides on glass slides. Second, the dendrimeric linker system increases the loading capacity by a factor of 10. Third, it eliminates non-speciﬁc attachment of hybridization probes and provides a low ﬂuorescent background. Fourth, covalent bonding is stable, and the microarrays can be reused many times. Finally, bonding through a terminal group of the attached molecules produces no apparent negative effects on hybridization efﬁciency.

4.

Gel Coating

Recently developed attachment strategies that use polyacrylamide or agarose for surface modiﬁcation combine the advantages of porous and non-porous substrates (Afanassiev et al., 2000; Zlatanova and Mirzabekov, 2001). In this approach, polyacrylamide gel elements or pads, ranging in size from 10 £ 10 £ 5

196

J. ZHOU AND D. K. THOMPSON

to 100 £ 100 £ 20 mm3 with volumes varying from picoliters to nanoliters, are afﬁxed to the glass surface. Because each gel-pad is surrounded by a hydrophobic surface that prevents solution diffusion among the elements, they can function independently. The probe molecules are immobilized in the gel-pads by robotic application. Compared to the direct attachment of probes to solid supports, the use of polyacrylamide gel-pad as an immobilization support offers some important advantages (Drobyshev et al., 1999). Three-dimensional immobilization of probes in gel-pads provides a greater density capacity and a more homogeneous environment than heterophase immobilization on glass, leading to higher sensitivity and a faster hybridization rate (Vasiliskov et al., 1999). However, like nylon membrane-based supports, gel-pads can yield higher background levels (Beier and Hoheisel, 1999). In addition, there are restrictions on the size of the molecules that can diffuse into the gel, such that fragmentation of the probe and target DNA may be required to generate molecules of the appropriate size (Englert, 2000). The convenience of using the gel-pad attachment strategy is limited by the fact that the method requires activation of gels and probes with labile reactive chemicals (Rehman et al., 1999). A more ﬂexible attachment method using copolymerization of 50 -terminal modiﬁed oligonucleotides with acrylamide monomers was developed (Rehman et al., 1999). The advantages of this method are that probes can be prepared easily using standard DNA synthesis chemistries and probes can be speciﬁcally and efﬁciently immobilized in the absence of highly reactive and unstable chemical crosslinking agents. Agarose was also examined as a coating material for probe attachment (Afanassiev et al., 2000). Agarose ﬁlm is activated to produce reactive sites that permit covalent immobilization of molecules with amino groups. Agarose has a higher binding capacity compared to a glass-based planar surface and does not interfere with ﬂuorescent detection. In contrast to acrylamide gels (Zlatanova and Mirzabekov, 2001) and dendrimeric branched systems (Beier and Hoheisel, 1999), this method does not require complex preparation technology.

5.

Nitrocellulose Coating

Glass slides coated with a proprietary nitrocellulose-based polymer have also been examined as an immobilization support (Stillman and Tonkinson, 2000). The nitrocellulose-based polymer can bind biomolecules (both DNA and proteins) noncovalently but irreversibly, providing better spot-to-spot consistency, higher binding capacity, and greater dynamic range compared to other glass slide modiﬁcations. Nitrocellulose-coated slides are also suitable for ﬂuorescent detection due to their relatively low-light scattering capacity. Although they

MICROARRAY TECHNOLOGY

197

have a higher ﬂuorescent background, the background-subtracted signal is signiﬁcantly higher on this support than the glass support coated with polylysine. However, the potential for miniaturizing the array dimensions of such slides remains to be determined.

C. ARRAYING TECHNOLOGY Microarray fabrication involves the printing and stable attachment of DNA probes on the array (Fig. 1). The microarray format is compatible with many advanced printing technologies, of which the most widely used are photolithography, mechanical microspotting, and ink-jet ejection. Each fabrication technology possesses advantages and disadvantages (Schena et al., 1998; Schena and Davis, 2000). All three technologies allow the manufacture of microarrays with sufﬁcient density for genetic mutation detection and gene expression proﬁling applications. The key considerations in selecting a fabrication technology include microarray density and design, biochemical composition and versatility, reproducibility, high-throughput capacity, and cost. Because of its versatility, affordability, and wide applications, microspotting is likely to become the printing technology of choice for the basic research laboratory. Thus, our discussion in this section will focus primarily on microspotting technology. 1.

Light-directed Synthesis

In photolithography, oligonucleotides are synthesized in situ on a solid surface in a predeﬁned spatial pattern by using a combination of chemistry and photolithographic methods borrowed from the semiconductor industry (Fodor et al., 1991; McGall and Fidanza, 2001) (Fig. 3). Brieﬂy, a glass or fused silica substrate is covalently modiﬁed with a silane reagent to obtain a surface containing reactive amine groups, which are then modiﬁed with a speciﬁc photoprotecting group, namely methylnitropoperonyloxycarbonyl (MeNPOC). Then the speciﬁc regions of the surface are activated through exposure to light, and a single base is added to the hydroxyl groups of these exposed surface regions using a standard phosphoramidite DNA synthesis method. The process of photodeprotection and nucleotide addition is iterated until the desired sequences are generated. Typically, the probes synthesized in situ on the arrays are 20– 25 bp in length. Since the average stepwise efﬁciency of oligonucleotide synthesis ranges from 90 – 95%, the proportion of the full-length sequences for 20-mer probes is approximately 10%. However, this should have a relatively minor effect on the performance of microarray hybridization because

198

J. ZHOU AND D. K. THOMPSON

Figure 3 In situ light-directed oligonucleotide probe array synthesis. The solid surface contains linkers with a photolabile protecting group X (black box) (e.g., MeNPOC). MeNPOC is resistant to many chemical reagents but it can be removed selectively by using ultraviolet light for a short time. When MeNPOC is removed, the deprotected region on the surface can form chemical bonds with DNA bases containing a MeNPOC photoprotecting group at its 50 hydroxyl position. In this illustration, light is directed through a photolithographic mask to speciﬁc areas of the array surface, which are activated for chemical coupling. The ﬁrst chemical building block A containing a photolabile protecting group X is then attached. Next, light is directed to a different region of the array surface through a new mask. The second chemical building block T containing a photolabile protecting group X is also added. This process is repeated until the desired product is obtained.

of the high absolute amount of full-length probes on the support (McGall and Fidanza, 2001). Another emerging light-directed synthesis approach for constructing highthroughput oligonucleotide arrays is to use a digital light processor (DLP), i.e., micromirror (Nuwaysir et al., 2002). This maskless array synthesizer (MAS) technology uses DLP to create “virtual” masks that direct an ultraviolet light beam to discrete locations on a glass substrate for DNA synthesis. Similar to the Affymetrix photolithographic approach, MAS is capable of constructing highdensity microarrays containing any desired nucleotide sequence. In contrast, MAS does not require photomasks, which are very expensive and timeconsuming to manufacture. The MAS technology makes photolithography much more ﬂexible and user-friendly, although it is still in the early development stages (Nuwyasir et al., 2002). Photolithographic parallel synthesis offers a very efﬁcient approach to highdensity array fabrication, in which the maximum achievable density is ultimately dependent on the spatial resolution of the photolithographic process. Due to steric and/or electrostatic repulsive effects, there is an optimum probe density for maximum hybridization signal. Affymetrix chips currently contain ~250,000

MICROARRAY TECHNOLOGY

199

oligonucleotides in an area of 1 cm2. One of the main advantages of this approach is that microarrays of extremely high-density can be constructed (Ramsay, 1998), but only oligonucleotides can be used in photolithography.

2.

Contact Printing

The most commonly used microarray fabrication technology is mechanical microspotting, which uses direct contact of computer-controlled multiple pins, tweezers, or capillaries to deliver picoliter volumes of pre-made biochemical reagents (e.g., oligonucleotides, cDNA, genomic DNA, antibodies, or small molecules) to a solid surface. Currently, more than 1,000 individual cDNA molecules can be deposited in an area of 1 cm2 using this technology (Rose, 2000). The advantages of microspotting include ease of implementation, low cost, and versatility, while a major disadvantage is that each sample to be arrayed must be prepared, puriﬁed, and stored prior to microarray fabrication. In addition, microspotting rarely produces the densities that can be achieved with photolithography. The various pin technologies for microspotting are described below. Solid pins. Solid pins have either ﬂat or concave tips. Because such tip can accommodate only a relatively small volume of the sample, only one microarray can be printed generally with a single sampling load. Consequently, the overall printing process is slow, making this technology suitable only for constructing low-density arrays. Additionally, loss of sample due to evaporation is a signiﬁcant problem due to the large surface-to-sample volume ratio of solid pins. Under standard laboratory conditions, about half of a 250 pl volume is lost in 1 s (Mace et al., 2000). To minimize evaporation loss, a highly humid environment is absolutely necessary. However, high humidity may prevent the sample from drying sufﬁciently on the slide, resulting in sample migration or spreading. Split pins. Split pins have a ﬁne slot at the end of the pin for sample holding. When the slit pin is dipped into the sample solution, the sample is loaded into the slot, which generally holds 0.2 –1.0 ml of sample solution. A small volume of sample (0.5 – 2.5 nl) is deposited on the microarray by tapping the pins onto the slide surface with sufﬁcient force (Rose, 2000) or touching the pins lightly on the surface like an ink stamp (Martinsky and Haje, 2000). The company, TeleChem International, manufactures a split pin by using digital control, so that there are virtually no variations in mechanical quality from pin to pin (Martinsky and Haje, 2000). Thus, TeleChem pins provide very high printing consistency under conditions of good sample preparation, proper motion control and homogeneous printing substrate. Because the split pins hold a larger sample volume than

200

J. ZHOU AND D. K. THOMPSON

the solid pins, more than one microarray can be printed from a single sampling event. Although split pin technology has been used successfully to print microarrays, one of the drawbacks is that dust, particulates, evaporated buffer crystals and/or other contaminants can clog the pin slot. Tapping the pins on the substrate surface, however, is not recommended for various reasons (Martinsky and Haje, 2000). Physical tapping leads to bulk transfer of the sample from the pins and hence causes non-uniformly large spots and spot merging. Also, tapping on the slide surface may lead to deformation of the pin tip, causing larger spots, irregular spot shapes, a larger amount of sample deposition, and poor printing quality (Mace et al., 2000). In addition, tapping may fracture the surface coating and cause irregular spot shapes such as doughnut shapes, in which the center of the spot lacks probe material (Martinsky and Haje, 2000). Pin and Ring. This is a variation of the pin-based printing process. The sample is taken by dipping the ring into the sample well and then a small volume of sample solution is deposited onto the slide surface by pushing the sample captured in the ring using solid pins (Mace et al., 2000). Different sized rings can be selected to hold 0.5– 3.0 ml of sample. Many different spot sizes can be obtained by simply using pins of different diameter. In addition, loss of sample due to evaporation is alleviated by minimizing ﬂuid exposure through speciﬁc ring conﬁguration. The pin-and-ring arraying technology offers a number of advantages. Since the pin is used only for spotting and the sample ﬂuid captured by the ring is relatively large, the deposition of samples on different slides is accomplished in an identical manner for each printing cycle, yielding a microarray fabrication quality that is consistent and reproducible (Mace et al., 2000). In addition, the ring geometry is capable of handling a wide variety of volumes and ﬂuid viscosities. Unlike the split pin, the pin and ring conﬁguration is not susceptible to clogging by the accumulation of dust, particulate matter, high-viscosity ﬂuids, debris, buffer or salts, and other materials. Finally, the pin and ring can deposit samples on soft substrates such as agar, gels and membranes. However, one drawback of this technology is sample loss. The majority of samples captured by the ring cannot be used for spotting and is lost through washing for the next sampling cycle.

3.

Non-contact Ink-jet Printing

Ink-jet ejection technologies provide another means of fabricating microarrays. In this approach, the sample is taken from the source plate, and a droplet of sample is ejected from the print head onto the surface of the substrate. Similar

MICROARRAY TECHNOLOGY

201

to microspotting, ink-jet ejection allows the spotting of virtually any biological molecule of interest, including cDNA, genomic DNA, antibodies, and small molecules. In contrast to microspotting, ink-jets have the advantage of avoiding direct surface contact but cannot be used to manufacture microarrays as dense as those prepared by photolithography or microspotting approaches. Currently two types of non-contact ink-jet print technologies, piezoelectric pumps and syringe-solenoid, are used for printing microarrays. Piezoelectric Pumps. This printing technology utilizes a piezoelectric crystal, which contacts a glass capillary containing the sample liquid (Englert, 2000; Mace et al., 2000; Rose, 2000) and is still in the early stages of development. When the crystal is biased with a voltage and subsequently deformed, the capillary is squeezed and a small volume (0.05 –10 nl) of ﬂuid is ejected through the tip from the reservoir. Piezoelectric printing has the advantages of an extremely fast dispensing rate (on the order of several thousand drops per second), very small print volumes, and consistency of droplet size. The main problem with this technology is clogging by air bubbles and particulates, which makes the system less reliable compared to other printing methods. In addition, the void volume of sample solution contained in the capillary is very large (100 – 500 ml) and not recoverable. It is also difﬁcult to change samples using piezoelectric printing. Syringe-solenoid Printing Technology. This technology uses a syringe pump and a microsolenoid valve for dispensing samples (Rose, 2000). The sample is taken by a syringe, and sample droplets, ranging from 4 to 100 nl in volume, are ejected by pressure onto the surface through the microsolenoid valve. The main advantages of this technology are reliability and low cost. However, it is not suitable for fabricating high-density microarrays because of the large printing volume and spot size.

D. CRITICAL ISSUES

FOR

MICROARRAY FABRICATION

This section highlights some practical issues that are important for microarray fabrication: microarray density, reproducibility, storage time, contamination and printing quality.

1.

Microarray Density

Microarray density is one of the most important parameters for microarray fabrication. The number of DNA elements that can be deposited on a slide will depend on spot size and pin conﬁguration.

202

J. ZHOU AND D. K. THOMPSON

Spot Size. Microarray density is determined by the size of the DNA spot and depends directly on the volume of sample deposited on the substrate surface. The volume of sample deposited per position on the array generally ranges from 50 to 500 pl, with high and low extremes of 10 pl and 10 nl possible, respectively (Mace et al., 2000). Several factors affect the volume of sample that can be applied to an array surface, including the surface properties (e.g., surface energy) of the slides and pins, and the sample solution characteristics (e.g., viscosity). For printing high-density microarrays, a hydrophobic glass surface (e.g., aldehyde-modiﬁed slide) is preferred, because spotted hydrophilic samples will spread less on a hydrophobic surface than on a hydrophilic one. Because the pin contact surface area determines the initial contact between the sample and slide, the spot size increases as the pin contact surface area increases. In addition, pin velocity has a great effect on the spot size. The loading sample volume for a split pin (e.g., ChipMaker and Stealth pins from TeleChem) typically ranges from 0.2 to 0.6 ml. Thus, if the pins tap the surface at high speed (. 20 mm/s), a large sample volume may be forced out of the pin and large spots will be produced (Rose, 2000).

Pin Conﬁguration. Printing pins are mounted in a print head, which can hold up to 64 pins. The distance between pins on the print head is 4.5 mm and precisely matches the well spacing of a 384-well microtitre plate. DNA samples are ﬁrst taken from 96-well or 384-well source plates by dipping the pins into the sample wells with either a single pin or multiple pins, and then depositing the sample on the slide surface by gently touching the pins to the surface. Fabrication of arrays using a single pin is the most straightforward approach, but it is also timeconsuming. The main advantage of single pin printing is that the DNA samples in the source plates can be positioned on the array in the same order as they occur in the source plate, thus making sample tracking and post-hybridization analysis easier. Another advantage is that pin-to-pin variations are not a problem when using a single pin for microarray printing. Using a single pin and 250-mm spot-tospot spacing, for example, more than 20,000 spots can be deposited on 22 £ 72 mm2 printing area. Multiple pins are generally used for printing high-density microarrays because of the increased printing speed, even though different pins can cause variations in array quality. To print with multiple pins, the pins are dipped into sample wells of a 384-well plate and then touched to the slide surface simultaneously to create separate spots at a 4.5-mm spacing in the ﬁrst round. Later rounds of printing are achieved by spotting with a predeﬁned spot-to-spot offset distance from the previous location. Each pin deposits samples in a sub-grid. Since some areas within each sub-grid might not be completely ﬁlled with spots due to the restriction of the layout, the density of microarrays will generally decrease as the number of pins used increases. Printing microarrays with multiple pins requires

MICROARRAY TECHNOLOGY

203

more time in designing the array layout, as well as sophisticated sample tracking and deconvolution in the data analysis phase. 2.

Reproducibility

Fabricating microarrays of reproducible quality is of the utmost importance in microarray-based experimentation. For reliable and reproducible data, the uniformity of individual spots across the entire array is very important for simplifying image analysis and enhancing the accuracy of signal detection. Several factors can affect the uniformity of spots, including array substrate, pins, printing buffer, and environmental controls. As mentioned previously, non-homogeneous surface treatment will cause variations in the amount of attached DNA. Variations in array quality can be caused by differences in pin geometry, pin age and sample solutions. Movement of the pin across the surface in the XY direction may cause the tip to bend (Rose, 2000). Tapping the pins on the surface may result in deformation of the pin tips. In addition, dragging the pin tip across the surface may cause clogging of the pin sample channel. Therefore, great care is needed in handling pins, even though they are robust. Pins should be cleaned with an ultrasonic bath after each printing (Rose, 2000). Environmental conditions have signiﬁcant effects on spot uniformity and size (Hegde et al., 2000). Humidity control is absolutely necessary for preventing sample evaporation from source plates and the pin channel during the printing process. Sample evaporation can cause changes in DNA concentration and viscosity. Reducing the extent of evaporation can help the small spotted volume of DNA have more time to bind at equal rates across the entire spot. As a result, DNA spots of high homogeneity will be obtained (Diehl et al., 2001). Generally, the relative humidity is controlled between 65 and 75% (Rose, 2000). Condensation could occur if the relative humidity is greater than 75%. Producing homogeneous spots on arrays also depends on the printing or deposition buffer. Saline sodium citrate is commonly used as a printing buffer in microarray construction; however, spot homogeneity and binding efﬁciency with this buffer can be poor. The addition of 1.5 M betaine to the printing buffer can signiﬁcantly improve spot homogeneity and binding efﬁciency (Diehl et al., 2001), because betaine increases the viscosity of a solution and reduces the rate of evaporation. More uniform spots can also be achieved with a deposition buffer that contains 50% dimethyl sulfoxide (DMSO) (Hegde et al., 2000; Wu et al., 2001). 3.

Storage Time

Another important practical issue concerns the shelf life of unused microarrays. The maximum time that microarrays can be stored prior to use is

204

J. ZHOU AND D. K. THOMPSON

currently unknown. The shelf time could depend on the coating chemistry of the slide and the storage conditions. Unprocessed microarrays can be stored in a dessicator for many months without deterioration of performance (Worley et al., 2000).

4. Contamination To produce high-quality microarrays, the collection of airborne dust and impurities on the slide surface must be eliminated or at least minimized during array fabrication. Dust and particulate matter can settle on the slide surface and cause printing inaccuracies as well as poor quality scanned image displays. Enclosing the array device in a humidity chamber can minimize dust contamination. Because pins are generally reused for depositing different biological samples, sample carryover during the printing process is a practical concern and can complicate interpretation of hybridization results. Efﬁcient cleaning of the pins is therefore required for the printing process. Generally, the pins are cleaned by dipping them into distilled water or detergent and then using a vacuum to remove the wash solution from the pin channel. Repeating this process three times is generally sufﬁcient to eliminate sample carryover problems. Cross-contamination during sample preparation and handling is another important concern in the microarray printing process. For making microarrays, plasmids containing the desired cDNA clones are generally extracted from bacterial cultures and the desired genes are ampliﬁed from the plasmid DNA. Recent studies showed that up to 30% of clones contained the wrong cDNA (Knight, 2001). This is most likely due to bacterial contamination and handling errors during sample preparation. Therefore, great care must be taken to eliminate or minimize such errors. Errors in public sequence databases are also possible and can lead to failures in microarray-based detection. For instance, some mouse sequences in the public databases correspond to the wrong strand of the DNA double helix. As a result, the designed oligonucleotide probes were not able to detect their target mRNAs (Knight, 2001).

5.

Evaluation of Printing Quality

After printing, it is important to assess the quality of the arrayed slides prior to hybridization in terms of surface quality, integrity and homogeneity of each DNA spot, the amount of deposited DNA, and consistency among replicated spots. Staining prior to hybridization will identify any problems introduced during the fabrication process. Microarrays can be stained with various ﬂuorescent dyes,

MICROARRAY TECHNOLOGY

205

such as PicoGreen, SYBR Green II and Topo Green (Battaglia et al., 2000), followed by ﬂuorescent scanning. Such information is critical in assessing the quality of the printed arrays and the data output from those arrays.

IV. MICROARRAY HYBRIDIZATION AND DETECTION A. PROBE DESIGN

AND

SYNTHESIS

Probe design and synthesis are critical steps in generating high-quality microarrays for gene expression analysis. Three types of probes are used for microarray fabrication: PCR products, cDNA clones, and oligonucleotides. For the construction of cDNA microarrays, individual open reading frames (ORFs) can be ampliﬁed using gene-speciﬁc primers. Because cross-hybridization among homologous genes is a potential problem, full-length genes cannot be used for microarray construction. Several computer programs are available that can identify DNA fragments speciﬁc (, 75% sequence identity) to each ORF by comparing the target gene with all other genes in a genome (Xu et al., 2002). Once the speciﬁc fragments are identiﬁed, more than one set of primers can be obtained based on the identiﬁed unique fragments using the PCR primer design program Primer 3 (Whitehead Institute). The designed primers are generally synthesized commercially. Optimal forward and reverse primers are generally selected based on the following considerations. First, for genes shorter than 1000 bp, the PCRampliﬁed unique fragments should be as long as possible. For genes longer than 1000 bp, the optimal ampliﬁed fragments should be within 500 –1200 bp. Second, each oligonucleotide primer should be 20 –28 bp in length and the set of primer pairs (typically stored in 96-well plates) should have an annealing temperature of approximately 65 8C to simplify PCR ampliﬁcation. If the desired target annealing temperature cannot be obtained, a lower annealing temperature can be used. In the case where speciﬁc fragments cannot be identiﬁed for some homologous genes, fragments with higher than 75% sequence identity will be selected and appropriate primers can be designed based on these fragments. However, hybridization signals for these genes should be carefully interpreted during microarray data analysis. One of the great practical problems with probe ampliﬁcation is that PCR product yields vary considerably among different genes and some primers may fail to yield PCR products. This may cause signiﬁcant variation in the DNA concentration present on the slide surface. The cDNA clone-based probes are generally derived from whole genes or fragments of genes that are ampliﬁed from clone libraries using vector-speciﬁc primers. The size of clone probes generally ranges from a few hundred to a few

206

J. ZHOU AND D. K. THOMPSON

thousand base-pairs. Generally, since vector-speciﬁc primers are used for amplifying the cloned inserts, clone-based probes cannot be speciﬁcally designed for regions of low homology to other genes. As mentioned above, a substantial portion of clones may contain the wrong cDNA due to bacterial contamination and mishandling (Knight, 2001). Oligonucleotide probes are different from other probes in that they can be deposited by printing or synthesized in situ on a solid surface. Speciﬁc oligonucleotide probes can be designed based on gene sequences. Generally, the sizes of the oligonucleotide probes are shorter than 25 bp, and several different oligonucleotide probes are used per gene for high-density oligonucleotide microarrays. To discriminate mispriming, a probe is designed deliberately to have a single mismatch at the central position (Lockhart et al., 1996; Warrington et al., 2000). Recently, the utility and performance of oligonucleotide microarrays containing 50- to 70-mer oligonucleotide probes were evaluated (Kane et al., 2000). The results indicate that such oligonucleotide microarrays can be used as a speciﬁc and sensitive tool for monitoring gene expression.

B. TARGET LABELING

AND

QUALITY

Target labeling is another critical step in successful microarray-based experimentation. The methods available for labeling nucleic acids for microarray hybridization can be classiﬁed into two categories: direct and indirect labeling. 1. Direct Labeling In direct labeling, ﬂuorescent tags are directly incorporated into the nucleic acid target mixture before hybridization by enzymatic synthesis in the presence of either labeled nucleotides (e.g., Cy3- or Cy5-dCTP) or PCR primers (Fig. 4A). The most commonly used method is to label the target mRNA or total cellular RNA using reverse transcriptase. In a ﬁrst-strand reverse transcription reaction, ﬂuorescently labeled cDNA copies of RNA are synthesized by incorporation of a ﬂuorescein-labeled nucleotide analog. Random hexamers, oligo(dT) or gene-speciﬁc primers can serve as primers for the initiation of reverse transcription. Since prokaryotic mRNA has no poly-(A) tail, random hexamers are generally used for reverse transcription. In this case, total cellular RNA is used as the template for cDNA synthesis, and hence a greater degree of background ﬂuorescence intensity can occur. Although genespeciﬁc primers can reduce such background levels by copying gene-speciﬁc

MICROARRAY TECHNOLOGY

207

Figure 4 Labeling strategies. (A) Direct incorporation of ﬂuorescent dyes into target sample through reverse transcription. (B) Incorporation of ﬂuorescent dyes into target samples through reverse transcription in the presence of amino-allyl-dUTP, followed by chemical coupling with ﬂuorescent dyes. (C) Dendrimer-based indirect labeling. (Courtesy of Molecular Probes).

fragments, it requires reverse transcription with hundreds or thousands of primers. A variation of the direct labeling approach is that mRNA is ampliﬁed up to 1,000 –10,000-fold by T7 polymerase to obtain antisense mRNA (aRNA), and then the aRNA is reverse transcribed to obtain labeled cDNA (Salunga et al., 1999). One of the advantages of the T7 polymerase-based ampliﬁcation method over other ampliﬁcation methods is that all mRNAs are almost equally ampliﬁed, because ampliﬁcation with T7 polymerase is a linear process. Another advantage is that mRNA can be labeled easily with reverse transcriptase, which incorporates ﬂuorescent tags much more readily than DNA polymerase.

208

J. ZHOU AND D. K. THOMPSON

Figure 4 (continued )

One of the problems with the reverse transcriptase-based labeling approach is that nucleotides tagged with structurally different ﬂuorescent dyes are differentially and non-uniformly incorporated into cDNA. To resolve this problem, a two-step approach was proposed. The FairPlaye system developed by Stratagene Corporation uses a two-step chemical coupling method to ﬂuorescently label cDNA. First, an amino allyl-dNTP is uniformly and efﬁciently incorporated into cDNA by reverse transcriptase (Fig. 4B), because the amino allyl-dNTP does not exhibit steric hindrance. Then, an amine-reactive cyanine is chemically coupled to the amino-modiﬁed cDNA. The main advantage of this approach is that this system efﬁciently produces uniformly labeled cDNA without any dye bias. As a result, this system is highly sensitive (5-fold increase in sensitivity), requires less RNA, and allows detection of low abundance genes. Any labeling bias resulting from ﬂuorescent dye incorporation also appears to be negligible and thus the dual labeling experimental approach is not needed.

MICROARRAY TECHNOLOGY

2.

209

Indirect Labeling

In the indirect labeling approach, ﬂuorescence is introduced into the detection procedure following hybridization. Brieﬂy, epitopes are incorporated into the target samples during cDNA synthesis. After hybridization with the epitope-tagged target samples, the microarray is incubated with a ﬂuorescently tagged protein that binds to the epitopes. The most common indirect labeling method uses a biotin epitope and a ﬂuorescent streptavidin – phycoerythrin conjugate (Warrington et al., 2000). The biotinylated nucleotides are incorporated into cDNA by reverse transcription and hybridized with the microarrays. After hybridization, the array is stained with a streptavidin – phycoerythrin conjugate, which binds to biotin tags and emits ﬂuorescent light when excited with a laser. Another indirect labeling approach is known as Tyramide Signal Ampliﬁcation (TSA) (Adler et al., 2000). This approach uses biotin and dinitrophenol (DNP) epitopes as well as streptavidin and antibody conjugates linked to horseradish peroxidase (HRP). In the presence of hydrogen peroxide, HRP catalyzes the deposition of Cy3- or Cy5-tyramide compounds on the microarray surface. By this method, a DNP- or biotin-dCTP analog is ﬁrst incorporated into cDNA, and then the epitope-tagged cDNA is hybridized with the microarray. Following hybridization, the microarray is incubated with anti-DNP-HRP, and Cy3-tyramide is deposited on the microarray surface, followed by incubation with streptavidin – HRP and deposition of Cy5-tyramide (Adler et al., 2000). The main advantage of this approach is that it can provide10- to 100-fold signal ampliﬁcation over the direct labeling approach. Thus, this approach can be used effectively to monitor the expression or abundance level of rare transcripts or to analyze samples prepared from small numbers of cells. The main disadvantage of this method is that it is generally less precise for comparative analysis due to variations arising from differences in labeling efﬁciencies and protein-binding afﬁnities (Schena and Davis, 2000). In addition, the signal intensity is only semiquantitative because of the involvement of enzymatic signal ampliﬁcation (Alder et al., 2000). The third indirect approach is to use DNA dendrimer technology (Stears et al., 2000) (Fig. 4C). Dendrimers are stable, spherical complexes of partially doublestranded oligonucleotides with a determined number of free ends, which are tagged with ﬂuorescent dyes, Cy3 or Cy5. In this technology, the cDNA is ﬁrst synthesized by reverse transcriptase with primers containing speciﬁc capture sequences that can bind the Cy3- or Cy5-tagged dendrimers through sequence complementarity. The synthesized cDNAs are then hybridized to microarrays, and the bound cDNAs on the microarrays are detected by incubating the arrays with the ﬂuorescent dye-tagged dendrimers. The dendrimer detection approach is highly sensitive, requiring up to 16-fold less RNA for probe synthesis. Since the ﬂuorescent dye is attached to the free end of the dendrimers, signal intensity is

210

J. ZHOU AND D. K. THOMPSON

independent of probe size and composition. In addition, this detection system has a high signal-to-background ratio and can be used for multiple channel detection on a single microarray.

C. HYBRIDIZATION After microarray fabrication, the most important issue in microarray-based analysis is probe –target hybridization. Conceptually, microarray hybridization and detection are quite similar to the traditional membrane-based hybridization (Eisen and Brown, 1999). Before hybridization, the free functional groups (e.g., amine) on the slide should be blocked or inactivated to eliminate nonspeciﬁc binding, which causes high background and depletion of probes. Any unbound DNA on the slides can be washed away during the pre-hybridization process. Removal of unbound DNA in pre-hybridization is important, because any DNA that washes from the surface during hybridization competes with DNA bound to the slide. Since the rate of hybridization in solution is much faster than that on surfaces, the presence of unbound probe DNA can lead to a dramatic decrease in the measured signals obtained from microarrays. After pre-hybridization, the microarray is hybridized with ﬂuorescently labeled target DNA or RNA for a certain period of time. Post-hybridization washing then removes unbound labeled material. Regardless of the hybridization format, the hybridization solution should be mixed well so that the labeled targets are evenly distributed across the array surface to obtain the maximum number of optimal target – probe interactions. In addition, the wash solutions should be uniformly distributed to eliminate unbound probes, remove non-speciﬁc hybridization, and minimize background signal.

D. DETECTION Both the confocal scanning microscope and coupled charge device (CCD) camera have been successfully used for the detection of microarray hybridization signals, and many such devices are commercially available (Hegde et al., 2000). Although the confocal scanning microscope and CCD camera systems both have advantages and disadvantages (as described below), the former is more commonly used. Generally, a confocal scanner uses laser excitation of a small region of the glass slide (~100 mm2), and the entire array image is acquired by moving the glass slide, the confocal lens, or both across the slide in two directions (Schermer, 1999). The ﬂuorescence emitted from the hybridized target molecule is gathered with an objective lens and converted to an electrical signal with a photomultiplier (PMT) or an equivalent detector. The main drawbacks of using

MICROARRAY TECHNOLOGY

211

a confocal scanning microscope for signal detection is that each excitation wavelength must have its own laser, which can be expensive, and the device is very sensitive to any non-uniformity of the glass slide surface. The CCD camera exploits many of the same principles as a confocal scanner, but the CCD camera utilizes substantially different excitation and detection technologies (Schermer, 1999). CCD systems typically use broad-band xenon bulb technology and spectral ﬁltration (Basarsky et al., 2000). The key advantage of the CCD camera-based imaging systems is that they allow simultaneous acquisition of relatively large images of a slide (1 cm2) and hence do not require moving stages and optics, which reduces cost and simpliﬁes instrument design. However, since the CCD camera does not move the optics or stages, several images need to be captured from different ﬁelds of the microarray and then stitched together to represent the entire information on the slide. Because most commonly used ﬂuoresceins have a small difference between excitation and emission maxima, it is difﬁcult to effectively separate excitation and emission light in the spectral ﬁltration process.

E. CRITICAL ISSUES

IN

HYBRIDIZATION

AND

DETECTION

This section highlights some important practical issues related to microarray hybridization and detection, namely, probe retention and quantitative hybridization, target labeling and availability, spatial resolution and cross-talk, and photobleaching and scanning parameters.

1. Probe DNA Retention and Quantitative Hybridization In solution-based hybridization, signal intensity depends on both target and probe DNA concentrations. In gene expression proﬁling studies, it is assumed that the concentrations of all probe DNAs deposited on the microarrays are much higher than the mRNA concentrations in the ﬂuorescently labeled target samples, so that signal intensity is dependent exclusively on the mRNA concentration in the target samples. Therefore, many factors causing probe deposition variations will have negligible effects on hybridization signal intensity. For the accurate quantitation of gene expression, it is essential to ensure that the arrayed DNA probes are in excess relative to the labeled target cDNAs. Generally, a DNA concentration of 100 – 200 ng/ml is used for spotting, which corresponds to 100 –200 pg/spot for a 1-nl deposition. The retention is about 20– 30% on silanized glass surfaces (Worley et al., 2000). Thus, after boiling and hybridization, this corresponds to approximately 20– 60 pg of doublestranded DNA present in each spot for binding. Studies indicate that the arrayed

212

J. ZHOU AND D. K. THOMPSON

DNA appears to be in excess for all the protein-coding genes in Escherichia coli (Worley et al., 2000). However, probe DNA retention depends on slide surface, coating chemistry, post-processing, hybridization, and washing conditions. Therefore, to ensure accurate quantitative results for highly expressed genes, it is important to understand how much spotted DNA can actually be retained after hybridization. Whether the probe DNA concentration represented on the array substrate is in excess obviously depends on the amount of target sample used. Typically, 10 – 20 mg of total cellular RNA is used for monitoring gene expression in prokaryotes. However, for monitoring rare transcripts, higher RNA concentrations (e.g., 50 mg) are generally used. In this case, probes corresponding to abundant transcripts may not be in excess relative to the target samples, resulting in hybridization that is not quantitative. Hence, it is important to select the appropriate amount of RNA to ensure that the microarray signal is within the range of linear response for the system being used.

2.

Target Labeling and Availability

The integrity and purity of the RNA are crucial for obtaining high-quality microarray hybridization results. Impurities in RNA preparations could have an adverse effect on both labeling efﬁciency and the stability of the ﬂuorescent dyes. Thus, the RNA must be free of contaminants such as polysaccharides, proteins and DNA. Many commercial RNA puriﬁcation kits are available for producing RNA of sufﬁcient purity for microarray studies. In addition, unincorporated nucleotides present in the labeling reaction must be removed to reduce background noise. Finally, both Cy3 and Cy5 are sensitive to light, and thus great caution must be taken to minimize exposure to light during labeling, hybridization, washing and scanning. The most frequently encountered experimental problem is the variation in hybridization signal between labeling reactions. In many cases, poor hybridization signals result from poor dye incorporation. Very low dye incorporation (, 1 dye molecule/100 nucleotides) gives unacceptably low hybridization signal intensities. However, studies showed that very high dye incorporation (e.g., . 1 dye molecules/20 nucleotides) is also not desirable, because high dye incorporation signiﬁcantly destabilizes the hybridization duplex (Worley et al., 2000). Thus, it is important to measure the dye incorporation efﬁciency prior to hybridization. The speciﬁc activity of dye incorporation can be determined by measuring the absorbance at wavelengths of 260 and 550 nm for Cy3 or 650 nm for Cy5. A suitable labeling reaction should have 8– 15 A260/A550 ratio for Cy3 and 10 –20 A260/A650 for Cy5.

MICROARRAY TECHNOLOGY

213

Another problem encountered routinely is the quality of ﬂuorescent dyes. The labeling efﬁciency and hybridization vary signiﬁcantly sometimes from batch to batch, especially for Cy5. Fresh reagents are very important in achieving a high degree of detection sensitivity (Wu et al., 2001). Microarray hybridization is generally performed in the absence of mixing. Since the diffusion coefﬁcient is very small for large labeled target DNA molecules, the probe at each arrayed spot is in effect hybridizing with its labeled counterpart from its immediate or nearly immediate local environment (Worley et al., 2000). Thus, the target solution should be mixed well and uniformly distributed over the microarray surface area. Otherwise, the availability of the labeled target molecules to the arrayed spots could be signiﬁcantly different across the microarray surface. As a result, signiﬁcant differences in signal intensity can be observed.

3.

Spatial Resolution and Cross-talk

The spatial resolution of microarray detection systems is usually expressed as a pixel size, the physical “bin” in which a single datum is acquired. The spatial resolution for many commercial systems usually ranges from 5 to 20 mm. The selection of spatial resolution depends on spot size, and in general, the pixel dimension should be less than 1/10 of the diameter of the smallest spot on the array. For example, microarrays containing 100-mm spots require ﬂuorescent detectors with a spatial resolution of 10 mm pixel size. Cross-talk refers to the phenomenon in which an emission signal from one channel is detected in another channel, resulting in an elevated, erroneous ﬂuorescence reading. Cross-talk is most likely from the shorter wavelength channel into the longer wavelength channel. For example, the ﬂuorescence intensity from the Cy3 channel can contaminate the Cy5 channel but not vice versa. Cross-talk is the most common potential problem in the simultaneous scanning approach, which acquires both images from two channels at the same time (Basarsky et al., 2000). For gene expression experiments, cross-talk should be kept to less than 0.1%. The most common and cost-effective way to minimize cross-talk is to use emission ﬁlters that reject light outside the desired wavelengths. Cross-talk can also be greatly minimized by selecting ﬂuorescent dyes and lasers with sufﬁcient differences in wavelength (Schermer, 1999).

4.

Photobleaching and Scanning Parameters

Light is emitted from a ﬂuorescent dye when it is illuminated by a light source. Generally, the emitted ﬂuorescence is directly proportional to the power of

214

J. ZHOU AND D. K. THOMPSON

the excitation light. Therefore, to increase detection sensitivity, higher power of excitation light is preferred. If the excitation light is excessive, however, the incoming photons can damage the dyes and reduce the ﬂuorescent signals during successive scans, leading to photobleaching of the signal intensity. Photobleaching depends on the duration of sample illumination. More powerful light sources and/or a longer laser exposure time can result in signiﬁcant photobleaching. When acquiring an array image, it is best to keep photobleaching to less than 1% per scan. Different dyes have considerable differences in their photostabilities. For example, Cy5 is more sensitive to photobleaching than Cy3. The differences in photostability among different dyes could be a signiﬁcant problem when multiple dyes are used in an experiment, because the ratios measured can lead to signiﬁcant quantitative errors. To minimize photobleaching, the Cy5 channel is always scanned ﬁrst, followed by the Cy3 channel. Although Cy3 (0.15, no unit) has a lower quantum yield than Cy5 (0.28), Cy3 is more efﬁciently incorporated into cDNA during reverse transcription. Such dye characteristics can cause variations in the signal intensity obtained in reverse labeling experiments. The signal should be balanced during scanning by using a higher PMT setting for the dye with the weaker signal to allow detection of more spots of low signal intensity.

V. MICROARRAY IMAGE PROCESSING A. DATA ACQUISITION The fundamental aim of image processing is to measure the signal intensity of arrayed spots and then quantify gene expression levels based on the signal intensities acquired for each spot. Therefore, spots on the array image must be correctly identiﬁed. The spots on microarrays are arranged in grids. An ideal microarray image for easy spot detection should have the following properties: (i) the location of spots should be centered on the intersections between the row and column lines; (ii) the spot size and shape should be circular and homogeneous; (iii) the location of the grids on the images should be ﬁxed; (iv) the slides should have no dust or other contaminants; and (v) the background intensity should be very low and uniform across the entire image. In practice however, it is difﬁcult to obtain such ideal images. First, the spot position variation occurs because of mechanical limitations in the arraying process, including inaccuracies in robotic systems, the printing apparatus and the platform for holding slides. Second, the shape and size of the spots may ﬂuctuate considerably across the array because of variations in the size of the droplets of DNA solution, DNA and salt concentration in

MICROARRAY TECHNOLOGY

215

the printing solution, and slide surface properties. In addition, contamination from airborne dust and impurities on the slide surface is a major problem for processing array images. To obtain accurate measurements of hybridization signals, all of these potential problems should be taken into consideration. Many methods are available for resolving spot location errors, spot size and shape irregularities and contamination problems (Zhou et al., 2000) in order to accurately estimate spot intensities. Commercial and free software, including ImaGenee from BioDiscovery (Los Angeles, CA), QuantArraye from GSI Lumonics, and the software on Axon GenePixe systems (Bassett Jr. et al., 1999), can be used for microarray image processing. Typically, a user-deﬁned gridding pattern is overlaid on the image, and the areas deﬁned by patterns of circles are used for spot intensity quantiﬁcation. The data are extracted and generally expressed as the total (the sum of the intensity values of all pixels in the signal region), mean (the average intensity of pixels), and median (the signal intensity of the median pixel). Microarray output corresponding to the total intensity is not the best measurement of hybridization signal, because it is particularly sensitive to variations in the amount of DNA deposited on the surface and the presence of contamination (Zhou et al., 2000). The mean is probably the best measurement when using an advanced image processing system that permits accurate segmentation of contaminated pixels, because the mean measurement reduces variations caused by differences in the amount of DNA deposited within a spot. However, the mean measurement is vulnerable to outliers (Petrov et al., 2002). The median is a better choice than the mean if the image processing software is not good enough for correctly identifying signal, background and contaminated pixels. The median intensity value is very stable and is close to the mean if the distribution proﬁle of pixels is uni-modal. The median is equal to the mean when the distribution is symmetric. An alternative to the median measurement is to use a trimmed mean (the mean of the pixel intensity after a certain percentage of the pixels are removed from both tails of the distribution). Some comparative studies indicate that the choice of measurements depends on the segmentation techniques used. The mean is the best measurement if the combined and trimmed segmentation techniques are used, whereas the median is the best without trimming (Petrov et al., 2002).

B. ASSESSMENT

OF

SPOT QUALITY SUBTRACTION

AND

BACKGROUND

For some spots, signal intensity data may not be reliable because of the inherently high variation associated with array fabrication, hybridization, and image processing. Thus, the ﬁrst step in data processing is to assess the quality of spots, with the removal or ﬁltering of unreliable poor spots or outlying spots

216

J. ZHOU AND D. K. THOMPSON

(outliers) prior to data analysis (Heyer et al., 1999; Tseng et al., 2001). It is critical to identify problematical slides, because without assessing the quality of the spot signals, conclusions drawn from the analysis of such data could be misleading.

1.

Identiﬁcation of Poor Slides

Due to the multiple steps involved in microarray experiments, it is important to evaluate array slides as a whole and to eliminate unreliable hybridization signals prior to rigorous data analysis. Two measures can be used to assess the overall slide quality if replicate spots are present on the arrays (Worley et al., 2000): one can calculate (i) the average coefﬁcient of variation (CV) of replicates in the spot pairs and (ii) the r 2 value of the regression line from a scatter plot of duplicate spots. Although there is no general consensus on the appropriate threshold value for rejecting slides, slides are generally accepted if the average CV is less than 50%. If there are no replicate spots on the microarrays, slide quality can be assessed by determining the number of spots that are of poor quality. Generally, microarray experiments should be repeated if more than 30% of the spots on the microarray are ﬂagged as poor spots.

2. Identiﬁcation of Poor Quality Spots There are no rigorously deﬁned rules for identifying poor spots from a biological or statistical perspective. The spot quality and integrity are generally assessed based on the following criteria: Spot size and shape. Spots with excessively large or small diameters compared to the majority of spots should be discarded. Discarding such low-quality spots signiﬁcantly improves the reliability of the data (Zhou et al., 2000). Spot homogeneity. The distribution of pixels within the spots can be used to assess spot homogeneity. Generally, spots with less than a certain percentage (e.g., 55 – 60%) of all pixels having intensities greater than the average background intensities (Khodursky et al., 2000) or one standard deviation (SD) above local background are ﬂagged as poor quality spots (Murray et al., 2001). Spot intensity. Spots with signals not signiﬁcantly above background should be identiﬁed using various standards. For example, spots with median or mean signals less than one to three SDs above background in both channels (Chen et al., 1997; Basarsky et al., 2000; Hegde et al., 2000) are ﬂagged as poor quality spots.

MICROARRAY TECHNOLOGY

217

In addition, spots whose signal is not at least 2.5 times higher than the background signal in both channels are excluded (Evertsz et al., 2000). Another way to deﬁne poor spots is based on signal-to-noise ratio (SNR), which is often deﬁned as the ratio of the difference between signal and background divided by SD of background intensity (Verdnik et al., 2002). This ratio indicates how well one can resolve a true signal from the instrumental noises. A commonly used criterion for the minimum signal that can be accurately determined is an SNR value equal to 3. Below that value, the signal cannot be accurately quantiﬁed, and such spots are treated as poor spots. The commercially available software, ImaGene from Biodiscovery, is able to automatically ﬂag poor spots. Spots identiﬁed as poor quality are not included in the data analysis. Although the criteria for deﬁning poor spots are based on subjective thresholds rather than statistically robust tests, they take into account the major factors affecting the quality of data and are likely to be very effective in reducing the amount of noise.

3.

Removal of Outlying Spots

Outliers represent extreme values in a distribution of replicates. Outlying spots can be caused by uncorrected image artifacts such as dust or by factors undetectable by image analysis such as cross-hybridization. Outliers signiﬁcantly affect the estimation of expression values and its associated random errors. Thus, removal of outlying spots is an important step in data ﬁltering. However, distinguishing outliers from differentially expressed genes is very challenging, because there is no general deﬁnition describing outliers. In this section, we brieﬂy describe several commonly used methods for identifying outliers. Simple threshold cutoff. A gene whose CV is greater than a certain threshold (Murray et al., 2001), e.g., 30 –50%, is excluded from the data analyses. Intensity-dependent threshold cutoff by windowing procedure. (Tseng et al., 2001). The CV values for individual genes are plotted against the average signal intensity of the two channels [(Cy3 þ Cy5)/2]. For each gene, a windowing subset is constructed by selecting a certain number of genes (e.g., 50) whose mean intensities are closest to this gene. If the CV of this gene is within a top certain percentage (e.g., 10%) among genes in its windowing subset, then data on this gene are regarded as unreliable, and hence all replicate data for this gene are discarded. However, in many cases, not all replicate spot data for this gene are unreliable. To salvage some information for this gene, the most outlying spots can be eliminated, and the CV of the intensity ratios of the remaining spots corresponding to this gene can be recalculated. If the CV is signiﬁcantly reduced

218

J. ZHOU AND D. K. THOMPSON

below the threshold level, the data for the remaining spots can be used in subsequent analyses. The CV can also be used for assessing the quality of different slides and different experiments (Tseng et al., 2001). Removal of outliers with jackknife procedure. The jackknife correlation can be used to remove outliers for expression data obtained from time-series microarray experiments (Heyer et al., 1999). In this statistical approach, the correlation coefﬁcient is calculated for each pair of genes using all of the timeseries data points. Then the data at one time point are deleted, and the correlation coefﬁcient is recalculated respectively for each pair of genes with all of the timeseries data points but one. The jackknife correlation is the minimum correlation coefﬁcients obtained above and can then be used for further cluster analysis. Jackknife correlation is robust and insensitive to single outliers. Applying jackknife correlation reduces false positives, while capturing the shape of an expression pattern. Hyer et al. (1999) showed that the genes showing similar expression patterns generally had a jackknife correlation of 0.7 or higher. Identiﬁcation of outliers based on pooled error methods. Several methods are used for statistical detection of outliers, but they are generally less adequate for typical microarray studies due to the small number of replicates (Nadon et al., 2001). The random error estimation for each gene based on a small number of replicates is imprecise, which makes statistical tests insensitive. As a result, many replicate spots may be falsely identiﬁed as outliers or many true outliers may not be identiﬁed (Nadon and Shoemaker, 2002). The potential violation of the normality assumption makes inferences of outliers and gene differential expression less reliable (Nadon et al., 2001). The pooled error method assumes that all probes or probes of similar intensities within a speciﬁc study have the same true random error. Variance estimates therefore can be pooled together across many genes and the precision of error estimation can be greatly improved. Furthermore, it is assumed that the standardized residuals have a normal distribution if the pooled error model is correct. Under these assumptions, the existence of outliers will cause the distribution of the entire data set to deviate from normal. Removal of spots with large residuals will improve the normality of the entire data set. Generally, outliers are identiﬁed in an iterative fashion: spots with large absolute residuals are removed from the data set; data are examined for normality and the residuals are calculated again. The process is iterated until the index asymptotes approach a stable value, which indicates that further removal of data values would not improve the normality of the distribution of the remaining data set. Software is available (ArrayState) to facilitate array-based statistical analysis (Nadon et al., 2001). In this software package, outliers are automatically detected. The pooled error method is a better, more sensitive method for outlier detection and can be used appropriately for microarray experiments having as few as two replicates.

MICROARRAY TECHNOLOGY

4.

219

Background Subtraction

Subtraction of background ﬂuorescence from hybridization signals is the second step in microarray data processing. Background subtraction is necessary to distinguish actual signals based on hybridization from noise and allows the comparison of speciﬁc spots. There are two approaches to background subtraction. The ﬁrst approach is to take signal intensity levels from blank areas on the array and use this for subtraction. The problem with this approach is that the background varies across the array and thus the background noise among spots can be signiﬁcantly different. The second approach is to use a local background for the area around each spot for background determination. Local sampling of the background is generally used to specify a threshold that the true signal must exceed. By doing this, it is possible to detect weak signals and extract an average density above the background for each array element (Chee et al., 1996). After removing poor slides, poor quality spots, outliers and background, the microarray data are ready for further normalization and data analysis.

VI. MICROARRAY DATA ANALYSIS A. DATA NORMALIZATION Microarray hybridization possesses intrinsic variation, which can potentially occur at every step in the microarray process. One key question prior to applying statistical analyses is whether such variations represent true random variations in expression values or are due to systematic variations arising from differences in the experimental conditions. Before pursuing further statistical analyses of microarray data, the systematic variations must be removed by normalization to allow statistical comparisons among different slides and different experiments.

1.

Sources of Systematic Variations

Systematic variation stems from a number of sources during microarray experiments. The major anticipated sources of variations include the following (Tseng et al., 2001). Variations within a Slide or Spatial Effect. Many studies show that substantial signal variation occurs for the same gene within a slide (Dudoit et al., 2001). Differences in pin geometry, slide homogeneity, hybridization and target ﬁxation could all contribute to variations observed among repeated spots within a slide.

220

J. ZHOU AND D. K. THOMPSON

Some systematic differences may occur between different pins due to differences in the length or in the opening of the tips, pin deformation following multiple rounds of printing, and slight differences in surface properties. All of these can lead to differences in target DNA transfer and hence may cause systematic variations in microarray signal intensity. The amount of deposited target DNA also ﬂuctuates for the same pins, while studies showed that the variation among different pins was signiﬁcantly higher (Dudoit et al., 2001). In addition, the fraction of target DNA that is chemically ﬁxed onto the slides is unknown and could vary considerably within slides. For various reasons, the labeled targets may also be distributed unevenly over the slide and/or the hybridization reaction may occur unequally in different parts of the slides. Finally, some areas of a slide may be contaminated and have a high background. The inﬂuence of these factors on signal intensity measurement within a slide is generally referred to as spatial effect. Variation among slides or slide effect. Differences in surface properties, microarray fabrication, hybridization and imaging could lead to systematic variations in hybridization signals among different slides. The amount of probe DNA immobilized on the slide during array printing and probe ﬁxation can be substantially different among different slides due to various factors such as differences in slide surface properties and sample evaporation during printing. Also, the amount of cDNA added to the slides, especially when different RNA preparations are used, and the local environment and hybridization conditions, such as temperature, buffer pH, target concentration, incubation and washing time in each hybridization chamber, could be considerably different. Background noise and the local curvature of the surface among different slides may have a large impact on scanning, especially for confocal scanners which are sensitive to focus. The inﬂuence of these factors on measuring signal intensity is deﬁned as slide effect. Tseng et al. (2001) showed that such effects are signiﬁcant, and normalization is slide-dependent. Variation from probe labeling or label effect. The most commonly used ﬂuorescent dyes, Cy3 and Cy5, are not equally incorporated into DNA molecules by reverse transcriptase and DNA polymerase. Cy3 is incorporated more efﬁciently than Cy5 with the same preparation and amount of RNA. While both Cy dyes are relatively unstable, Cy3 and Cy5 have different quantum efﬁciencies and are detected by the array scanner with different efﬁciencies. While the detection limit of Cy5 with the scanner is lower than that of Cy3, Cy5 is more sensitive to photobleaching. The inﬂuence of these factors on intensity measurements is referred to as label effect. The use of two ﬂuorescent dye labels may also introduce gene-label interactions. For instance, ﬂuorescent labeling may ﬂuctuate systematically, depending on the nucleotide composition of the target sequences, and one or the

MICROARRAY TECHNOLOGY

221

other dye may be preferentially incorporated into speciﬁc gene sequences. Also, the length of Cy3- and Cy5-labeled cDNA by reverse transcription with random priming could be signiﬁcantly different from sequence to sequence. This longer labeled cDNA could potentially lead to higher intensity levels for certain arrayed probes. If such interaction occurs, certain sequences will always show higher intensities in one channel than the other channel even under non-differential conditions and after normalization. Variation in growth conditions and mRNA preparation or sample effect. In a comparative microarray experiment, two RNA samples extracted from cells grown under different conditions are labeled with different ﬂuorescent dyes. Because of the differences in genetic identity (e.g., wild type versus mutant strains) and environmental growth conditions, cell biomass and mRNA abundance could ﬂuctuate signiﬁcantly among different cultures. The RNA purity also could be very different from sample to sample and this could lead to different labeling and hybridization efﬁciencies. Furthermore, sensitivity to mRNA degradation could be considerably different between preparations. All of these factors affecting signal intensity are referred to as sample effects. Due to experimental variations, hybridization signals from microarrays should be normalized prior to comparing data from a single array or multiple arrays.

2. Genes Used for Normalization Two critical issues in the analysis of microarray data, are how to eliminate systematic variations and which genes should be selected as references for normalization. Experimental design largely determines the strategy used for normalization, three of which are described in detail below. Using all genes on the array or global normalization. Under a certain condition, only a small portion of the genes is expected to be differentially expressed. Thus, the remaining genes should exhibit constant expression levels between two channels and can be used for normalization to calibrate spatial effects (Dudoit et al., 2001), slide effects and label effects (Tseng et al., 2001). The prerequisite for using almost all genes on the array for normalization is that only a small fraction of the genes are expressed, and the numbers of down- or up-regulated genes are approximately equal. Using constantly expressed housekeeping genes. The housekeeping genes that are constantly expressed across a variety of conditions can be used for normalization (Duggan et al., 1999). Although it can be difﬁcult to identify a set of housekeeping genes that do not change signiﬁcantly under any condition, it

222

J. ZHOU AND D. K. THOMPSON

may be possible to identify small sets of temporary housekeeping genes for particular experimental conditions. A limitation in using housekeeping genes for normalization is that housekeeping genes are generally highly expressed and thus may not be representative of other genes of interest. Using controls. A third normalization approach is to use spiked controls or a titration series of control sequences. In the spiked control method, DNA sequences from organisms different from the ones being studied are printed on the array and then the mRNAs of the control sequences are mixed with the two different mRNA samples in equal amounts. These spotted controls should have equal Cy5- and Cy3-derived intensities, and thus can be used for normalization. One limitation is that the composition of the control sequences could be considerably different from the target sequences, and as a result, they may not be representative of the genes of interest. Another limitation is that it may be difﬁcult to determine how much mRNA to spike, because there are always varying amounts of rRNA and tRNAs present, and the degree of RNA degradation varies from sample to sample. In the titration series method, a series of concentrations of the control sequences are printed on the arrays. These control spots are expected to have equal Cy5- and Cy3-derived intensities across a range of concentrations. Genomic DNA could be used in the titration method, because it should have a consistent expression level across various conditions.

3.

Experimental Design and Normalization Strategies

Since microarray experiments have inherently high variation, careful experimental design and execution are critical for accurately identifying differentially expressed genes under different conditions. Appropriate normalization is therefore necessary to eliminate different types of systematic variations. Minimizing spatial effects. To minimize spatial effects, multiple spots for a gene or control DNA should be fabricated on the microarrays. For control sequences, various concentrations of sequences should be spotted on arrays. Multiple spots of genes or control DNAs within the same slide are very useful for identifying contaminated spots, spots having high background noise, and problematical slides in each experiment (Tseng et al., 2001). To minimize spatial effects, normalization can be performed for each sector of the microarray-based on all the genes in that sector. Since DNAs in different sectors are deposited by different pins, normalization is an effective way to eliminate pin-to-pin variations (Dudoit et al., 2001). By comparing the normalization results for different genes and

MICROARRAY TECHNOLOGY

223

control sequences among different sectors, one should be able to assess and minimize the systematic variations associated with slide surface properties. Minimizing labeling effects and label –gene interaction. To eliminate systematic variations in probe labeling and gene-label interaction, a reverse labeling experimental design is recommended (Kerr and Churchill, 2001a; Tseng et al., 2001). For this, two aliquots of the two RNA samples (A and B) are labeled with Cy3 or Cy5 separately, and then hybridized with two microarrays. The hybridization solution for the ﬁrst microarray consists of Cy3-labeled sample A and Cy5-labeled sample B, whereas the labeling for the second microarray hybridization is reversed for the two target samples. Then the signal intensity for each microarray is normalized based on all genes on the microarray or on a set of housekeeping genes using different normalization approaches (see below). After normalization, the signal intensities from both channels for each sample are averaged, and the intensity ratios of the two samples are calculated based on the averaged signals. The reverse labeling experimental design is effective in eliminating labeling effects and gene-label interaction (Tseng et al., 2001). Minimizing slide and sample effects. In a typical comparative study, multiple replicated treatments (e.g., 3) under each condition are used and mRNAs from two different conditions are ﬂuorescently labeled and co-hybridized to the same single or multiple slides. Under such an experimental design, the signal intensity is impacted by both slide and sample effects, and it will be difﬁcult to eliminate the resulting systematic variation based on using all arrayed genes for normalization. In this situation, one should identify a sufﬁcient number of nondifferentially expressed genes on each slide and use them to construct a normalization curve, because the expression level of the non-differentially expressed genes are expected to remain constant under the experimental conditions tested. One challenge is to select the non-differentially expressed genes. Although predetermined housekeeping genes are good candidates, they may not provide a good ﬁt for normalization due to the high level of expression and natural variability of their expression level. A rank invariant selection approach (Schadt et al., 2000; Tseng et al., 2001) can be used for selecting non-differentially expressed genes. This method presumes that for an up-regulated gene, the signal intensity rank for a channel will be signiﬁcantly higher than the rank in the other channel, and vice versa. Brieﬂy, the signal intensities of individual genes from both channels are ranked. If the ranks of Cy3 and Cy5 intensities for a gene differ by less than a certain threshold value, and the rank of the averaged intensity is not within the known levels of the lowest and highest ranks, then this gene is classiﬁed as a non-differentially expressed gene (Tseng et al., 2001). This method works well if the majority of the genes are not differentially expressed. However,

224

J. ZHOU AND D. K. THOMPSON

this method may fail if majority of the genes are up- or down-regulated (Tseng et al., 2001).

4.

Normalization Approaches

Methods for the normalization of microarray hybridization data can generally be categorized as linear or nonlinear. The major difference between these two types is that linear methods multiply all values in one channel by a correction factor, whereas nonlinear methods, which are preferred by most researchers working with microarrays, take the channel intensity into account and therefore are thought be more accurate. Here, we brieﬂy describe the most commonly used normalization methods. Correction factor based on total intensity. This method calculates a correction factor based on the total measured ﬂuorescence intensity. The primary underlying assumption is that the total amount of RNA labeled with Cy3 and Cy5 is equal because the same amount of RNA from the same sample is used in separate labeling reactions. Although the spot for any one gene in one channel may be higher than that in the other, such variations should be averaged out over thousands of spots on the array. Therefore, the total integrated intensity of all spots should be equal in both channels, and a constant signal correction factor can be derived to rescale the signal intensity of the other channel. Linear regression method. For differential experiments, it is expected that many genes will be expressed at a nearly constant level under two different growth conditions or treatments. Thus, the slope of the intensity in a scatter plot of both channels should be 1. Based on this assumption, the slope can be calculated by linear regression to obtain a correction factor, and then all values in one channel are multiplied by the correction factor to adjust the slope to 1. Trimmed geometric mean (TGM). This nonlinear method was initially described by Morrison et al. (1999) and is generally recommended for most normalization needs. The method assumes that under a certain condition, only a small proportion of the genes will be differentially expressed. Thus, the remaining genes should display a constant level of expression and can be used for normalization (Beliaev et al., 2002; Thompson et al., 2002). The signals from each channel are log transformed and sorted based on the intensity, then 5% of the extreme values (minimum and maximum) are discarded. The log-TGM and the SD of the log-trimmed means are calculated. The normalized value for a gene is obtained by dividing the difference between log intensity and log-trimmed means by the SD of the log-trimmed means. The normalized values are then

MICROARRAY TECHNOLOGY

225

converted back from log to normal values, which are then used to calculate expression ratios. Intensity-dependent nonlinear normalization method. In many cases, the dye bias is dependent on spot density (Dudoit et al., 2001; Tseng et al., 2001). Thus, an intensity-dependent normalization method may be preferable. Yang et al. (2001) proposed an intensity-dependent nonlinear normalization method that utilizes most of the genes on an array. Since this method is complicated, the reader is referred to the original paper for details (Dudoit et al., 2001). Brieﬂy, the log intensity ratio and the mean log intensity of both channels are calculated. The normalized intensity ratio is the difference between the actual log intensity ratio and the intensity ratio estimated based on Lowess function. Theoretically, this normalization method should be the most robust.

B. DATA TRANSFORMATION Prior to statistically analyzing the microarray data, it is important to establish whether the data meet the underlying assumptions of the particular statistical model that will be used. The most common requirements for statistical techniques are that the data have a normal distribution and homogeneous variance. If the data do not meet these assumptions, they may be transformed and reevaluated to determine if they meet the underlying assumptions. If the data do not meet the assumptions, the statistical analyses will not be valid. Although there are many different approaches to data transformation, the most commonly used approach in microarray studies is taking the logarithm of the quantiﬁed expression values. The rationale for this is three-fold. First, the variation in logs of intensities and logs of ratios of intensities are less dependent on absolute magnitude. Log transformation can equalize variability in microarray data with high variability. Second, log transformation evens out highly skewed distributions and thus brings the data closer to a normal distribution. Third, normalization is additive for logs of intensities. Studies show that log transformation is very effective in bringing the microarray data approximately to a normal distribution and is the best approach for the analysis of microarraybased gene expression data (Kalosai and Shams, 2001).

C. METHODS

FOR IDENTIFYING DIFFERENTIALLY EXPRESSED GENES

Generally, normalized intensity ratios under two different experimental conditions are used to assess differentially expressed genes. Standard statistical techniques cannot be easily used to determine which level of difference in gene

226

J. ZHOU AND D. K. THOMPSON

expression reﬂects an actual biological difference. This is because of the inherently high variation associated with microarray experiments and low-level replications. Three basic approaches are currently used for identifying differentially expressed genes. The ﬁrst approach, which is commonly reported in the literature, is based on arbitrarily assigned fold differences (Schena et al., 1996; Heller et al., 1997). If the average expression level varies by more than a constant factor (e.g., 2) between the treatment and control conditions, then this gene is considered to have changed signiﬁcantly in its expression. However, such a ﬁxed fold rule is unlikely to identify real biologically differences, because a factor of two has a different signiﬁcance, depending on the levels of gene expression and variation. The fold rule method is applicable only when the variance among the replicates within a treatment is identical for every gene so that the sample variance can be ignored. However, in practice, the variance differs among genes, and it is critical to incorporate such information into a statistical test. The second approach is to use standard statistical t-test (Baldi and Long, 2001; Beliaev et al., 2001; Thompson et al., 2002) or paired t-tests (Rogge et al., 2000) using the intensity ratio or log of the intensity ratio to test whether the fold change is signiﬁcantly different from 1 or 0. When the t value exceeds a certain threshold, depending on the conﬁdence level selected (typically the 95% conﬁdence level or P , 0.05), the gene expression level is considered to be signiﬁcantly different between two conditions. The t-test incorporates variance information and could potentially overcome the drawbacks of the fold rule method. Application of the t-test requires that all microarray experiments be highly replicated to obtain accurate estimates of the variance within experimental treatments. However, the level of replication within experimental treatments is often too low to permit t-tests, because the microarray experiments are costly and time-consuming to repeat or the amount of biological samples is very limited. A small number of replicates could lead to inaccurate estimation of variance and a correspondingly poor performance of the t-test itself (Baldi and Long, 2001). The third approach is to apply Bayesian probabilistic model-based regularized t-test to improve the conﬁdence in interpreting DNA microarray data with a low number of replicates (Baldi and Long, 2001). This method assumes that genes of similar expression levels have similar measurement errors, and that data from all of the genes with similar expression can serve as pseudo-replication of the experiment. Thus, variance of any single gene can be estimated by the weighted average of the variances from a number of genes with similar expression levels. This method has been applied to identify global expression proﬁles in E. coli K12 (Long et al., 2001). The results showed that the Bayesian approach identiﬁed a stronger set of genes that were signiﬁcantly up- or down-regulated and required less replication to achieve the same level of reliability as the t-test method. Since this method is computationally demanding, a program for accommodating this

MICROARRAY TECHNOLOGY

227

approach, Cyber-T, is available at the Web interface, www.genomics.uci.edu/ software.html. Various statistical methods are also available in ArrayState for identifying differentially expressed genes.

D. MICROARRAY DATA ANALYSIS A massive amount of data is generated by microarray hybridization, and the great challenge is how to extract meaningful biological information. One of the key goals for microarray data visualization and analysis is to identify statistically signiﬁcant up- and down-regulated genes and co-regulated genes exhibiting similar expression patterns. Although many different statistical methods have been used for analyzing microarray data, they are still in the early stages of development. In this section, several current methods will be brieﬂy reviewed.

1.

Scatter Plot

Scatter plots are the simplest way to visualize microarray expression data. In a comparative experiment, microarray hybridization is generally performed with two samples from two different conditions. One can use a scatter plot to visualize up- and down-regulated genes by assigning x- and y-axis values to represent signal intensity under the two different conditions. In the scatter plot, genes with equal expression values for two conditions fall along the diagonal identity line, whereas genes that are differentially expressed fall-off the diagonal line; the greater the deviation from the diagonal line, the greater the difference in the expression of a given gene between two samples.

2.

Similarity Measurement

In a typical microarray experimental design, multiple experimental conditions at multiple time points are generally compared. In large experiments analyzing thousands of genes, the increased data volume makes it very difﬁcult to identify gene expression patterns using scatter plots. More sophisticated multivariate analysis techniques should therefore be used in such cases. To use different multivariate analysis methods, the relationships among different genes should ﬁrst be quantiﬁed based on signal intensity using appropriate metrics. Two approaches are generally used for quantifying the relationships among different genes. One approach is to use Euclidean distance, which is deﬁned as the square root of the summation of the squares of the differences between all pair-wise comparisons. This metric measures the absolute distance between two

228

J. ZHOU AND D. K. THOMPSON

points in space deﬁned by the two gene expression proﬁles. In general, such distance measures are suitable when the objective is to cluster genes with similar expression patterns. The other approach is to use Pearson correlation coefﬁcient. For understanding regulatory networks, it is biologically more interesting to search for genes expressed at different levels but with similar overall proﬁles. Pearson correlation coefﬁcient is ideal for identifying proﬁles of similar shape. The values of this correlation coefﬁcient range from 21 (negative correlation) to 1 (positive correlation), and the method can detect both negatively and positively correlated genes. Several variations of the correlation metric have been used such as the correlation coefﬁcient with an offset of zero for speciﬁcally taking into account the reference state (Eisen et al., 1998) and jackknife correlation to counter against outlier effects (Heyer et al., 1999).

3.

Principal Component Analysis

Principal component analysis (PCA) is an exploratory multivariate statistical method for simplifying data sets that reduces the dimensionality of the variables by ﬁnding new variables, which are independent of each other. A few of the new variables, typically 2– 3, are selected to explain the majority of variance in the original data. Since each principal component is a linear combination of the original variables, it is often possible to assign meaning to what the principal components represent. For microarray data analysis, genes or experiments can be considered as variables. PCA has been used in a variety of biochemical studies, including the analysis of microarray data in identifying outlier genes and/or experiments (Hilsenbeck et al., 1999). The main advantage of PCA is that it identiﬁes outliers in the data or genes that behave differently than most of the genes across a set of experiments. It can also be used to visualize clusters of genes that behave similarly across different experiments. However, the number of clusters in the data sets is arbitrary and dependent on the user’s intuition or experience.

4.

Cluster Analysis

One of the most commonly used methods is cluster analysis. Cluster analysis is used to identify groups of genes, or clusters that have similar expression proﬁles. Clusters and the genes within them can be subsequently examined for commonalities in functions as well as sequences in order to gain a better understanding of how and why they behave similarly. Cluster analysis can help

MICROARRAY TECHNOLOGY

229

establish functionally related groups of genes and can predict the biochemical and physiological roles of functionally unknown ORFs. Despite the emergence of many methods for microarray data analysis, the optimal way of classifying such data is still open to debate. Depending on the way in which the data are clustered, cluster analysis can be divided into hierarchical clustering and non-hierarchical clustering. (i)

Hierarchical clustering. This method attempts to group genes and/or experiments in small clusters and then group these clusters into higher-level clusters and so on. As a result of this grouping process, a tree structure called a dendrogram is generated for visualization of the relationships between genes and/or experiments. There are three common options for hierarchical analysis based on the deﬁnition of the distance between two clusters: single linkage, average linkage, and complete linkage (Heyer et al., 1999). Although there are numerous versions of the basic algorithm, the most common is known as average linkage. Applications of hierarchical clustering to gene expression data have been described in recent studies (Eisen et al., 1998). Hierarchical clustering methods are very popular due to their simplicity and analysis speed. However, there are several problems associated with these methods (Heyer et al., 1999). First, decisions to group two elements are based only on the distance between them and once elements are joined, it is impossible for them to be separated. In addition, it is a local decisionmaking method and does not consider the data as a whole. It suffers from a lack of robustness and solutions may not be unique and dependent on the data order, leading to incorrect clustering overall. Finally, the tree is extremely complex for large data sets, with the performance decreasing with the square of the number of genes requiring classiﬁcation. (ii) Non-hierarchical clustering. One of the typical non-hierarchical clustering methods is k-means clustering, which identiﬁes predetermined k points as cluster centers. Each data point is assigned to one of these centers in a way that minimizes the total of distance between all points and their centers. The subsequent centers are chosen by identifying the data points farthest from the centers already chosen, and this process is iterated until the cluster memberships do not change appreciably (Tavazoie and Church, 1998). The advantage of k-means clustering is that it provides sufﬁcient clustering without having to create the full distance and similarity matrix or scan the whole dataset excessively (Zhou et al., 2000). This is particularly useful for microarray data with large numbers of genes and many different experimental conditions. The algorithm converges quickly for good initial choices of the cluster centers. The main disadvantage of this method is that the number of clusters, k, must be speciﬁed prior to running the algorithm, and the ﬁnal clustering relies heavily on

230

J. ZHOU AND D. K. THOMPSON

the choice of k. Generally, the number of clusters is not known in advance. In addition, the quality of the clusters identiﬁed by k-means is not guaranteed (Heyer et al., 1999). Recently, a new version of k-means (progressive k-means) was proposed to analyze gene expression data. This new procedure identiﬁes the number of different clusters from the data itself and is independent of a priori speciﬁed number of clusters (Ben-Dor et al., 1999; Herwig et al., 1999). Despite such limitations, the k-means methods appear to perform quite well with a large number of genes (Dopazo et al., 2001). To avoid limitations of hierarchical clustering and k-means clustering, another non-hierarchical clustering procedure, quality cluster, was developed (Heyer et al., 1999) that focuses on identifying large clusters with a quality guarantee. The quality cluster allows each ORF to initiate a candidate cluster, which is formed by starting each ORF and grouping the ORF with the greatest jackknife correlation coefﬁcient. Other ORFs are iteratively added in a way to minimize the increase in cluster diameter without removing the ORFs, which previous clusters included (Heyer et al., 1999). One characteristic of this procedure is that the number of candidate clusters is equal to the ORF numbers and many candidate clusters overlap, with the largest candidate cluster being retained. The ORFs it contains are eliminated and the entire procedure is iterated on the smaller set of ORFs until the largest remaining cluster has fewer than some pre-speciﬁed number of elements. There are several advantages of the quality cluster over both hierarchical and k-means clustering. First, the total number of clusters is not required prior to running the algorithm, and the quality of all clusters are guaranteed. Second, although the quality cluster algorithm is similar to the complete linkage hierarchical procedure, the clusters identiﬁed at a speciﬁed threshold are much larger on average. Third, since each ORF is considered a potential cluster center, local decisions do not have a great impact on the ﬁnal clustering results. Thus, it is expected that this method is less sensitive than hierarchical approaches to small changes in the data such as removal of ORFs through ﬁltering. Finally, this method is not sensitive to the order in which the similarity or distance data appear. Since this is a new clustering method, its value as an analysis tool remains to be determined.

5. Neural Network Analysis Since clustering methods have some serious drawbacks in dealing with data with a signiﬁcant amount of noise, a fundamentally different neural networkbased approach has been proposed for microarray data analysis (Tamayo et al., 1999; Toronen et al., 1999; Herrero et al., 2001). Unsupervised neural networks, and in particular self-organizing maps (SOMs), are a more robust and accurate

MICROARRAY TECHNOLOGY

231

method for grouping large data sets. The algorithm for neural network analysis works in the following way. First, a two-dimensional grid typically of hexagonal or rectangular geometry is deﬁned. Then, similar to k-means clustering, the number of clusters (k) is speciﬁed to correspond to the representative points in the speciﬁed geometrical conﬁguration. Data points are mapped onto the grid and the positions of the representative points are iteratively relocated in a way that each center has one representative point. Clusters close to each other in the grid will be more similar to each other than those further apart. The main advantage of SOMs is that they are robust to noise. In other words, they are able to handle large data sets containing noisy, poorly deﬁned items with irrelevant variables and outliers. This is particularly useful for analyzing microarray data. SOMs are also reasonably fast and can be easily scaled up to large data sets. One disadvantage of SOMs is that they require pre-determined choices about geometry, like the k-means method. The number of clusters is arbitrarily ﬁxed from the beginning and consequently, it is difﬁcult to recover the natural cluster structure of the data set. SOMs also yield non-proportional classiﬁcation. If irrelevant data or some particular type of proﬁle is abundant, the most interesting proﬁle will be mapped in few clusters and hence their resolution could be low. In addition, it is very difﬁcult to detect higher-order relationships between clusters of proﬁles due to the lack of a tree structure (Herrero et al., 2001). To overcome some of the limitations of SOMs, an unsupervised neural network with a binary tree topology, termed the self-organizing tree algorithm (SOTA), was proposed (Dopazo and Carazo, 1997). This new algorithm combines the advantages of hierarchical clustering (tree topology) and neural network (accuracy and robustness) and was used to analyze gene expression data (Herrero et al., 2001).

VII. USING MICROARRAYS TO MONITOR GENOMIC EXPRESSION Microarrays have been used widely to quantify and compare global gene expression in a high-throughput fashion. This section brieﬂy reviews the fundamental basis, general approaches to experimental design, and hybridization performance of microarrays in monitoring gene expression levels.

A. GENERAL APPROACHES TO REVEALING DIFFERENCES GENE EXPRESSION

IN

Temporal and spatial information concerning gene expression, as well as changes in mRNA abundance levels in response to different environmental

232

J. ZHOU AND D. K. THOMPSON

conditions, are important for understanding gene function and regulation. Three comparative approaches have been used for the display of differential gene expression.

1.

Differential Display of mRNA Under Different Physiological Conditions

Cells of interest are cultured under different physiological conditions, and the differences in mRNA abundance between the test and reference samples are compared using high-density microarrays. This is the most straightforward and widely used approach for identifying gene expression patterns associated with various physiological states (DeRisi et al., 1997; Tao et al., 1999; Ye et al., 2000; Beliaev et al., 2002).

2.

Differential Display of Temporal Gene Expression

Cells of interest are grown under a speciﬁc physiological condition and then harvested at different time points during growth. Changes in mRNA levels are revealed using microarrays. Information on the temporal dynamics of gene expression is very useful in understanding when genes are turned on or off and how genes interact with each other (DeRisi et al., 1997; Liu et al., 2003).

3.

Comparison of Gene Expression Patterns Between Wild-type and Mutant Cells

Differences in gene expression in response to changing environmental conditions can be very complicated, and oftentimes the expression proﬁles of many genes are altered as a result. Changes in the expression proﬁles for many genes present a great challenge to understand the underlying molecular mechanisms controlling these genes. The most effective approach to deﬁne the contributions of individual regulatory genes in a complex metabolic process is to use DNA microarrays to identify genes whose expression is affected by mutations in putative regulatory genes (DeRisi et al., 1997; Beliaev et al., 2002; Thompson et al., 2002). The basic approach to microarray-based gene expression studies is outlined in Fig. 5. In a typical microarray experiment for monitoring gene expression, genespeciﬁc PCR primers are designed based on whole-genome sequence information and synthesized. Gene-speciﬁc fragments are then ampliﬁed with speciﬁc primers, puriﬁed, and arrayed on solid substrates. Once the microarrays are ready, total cellular RNA isolated from bacterial cells grown under two different

MICROARRAY TECHNOLOGY

233

Figure 5 General approach for using microarrays for monitoring gene expression.

conditions (a control and experimental condition) is ﬂuorescently labeled with different dyes (Cy3 or Cy5) via the enzyme, reverse transcriptase. The microarray is then simultaneously hybridized with ﬂuorescently tagged cDNA from the test and reference samples. The signal intensity of each ﬂuorescent dye on the array is

234

J. ZHOU AND D. K. THOMPSON

then measured with a confocal laser scanning microscope or CCD camera. The quantitative ratio of red (Cy5) to green (Cy3) signal for each spot reﬂects the relative abundance of that particular gene in the two experimental samples. With appropriate controls, the intensity can be converted into biologically relevant outputs (e.g., the number of transcripts per cell). A series of samples can be compared with each other through separate co-hybridizations with a common reference sample, and the data can be analyzed with various statistical methods. Eisen and Brown (1999) provide a detailed discussion of the technical aspects of microarray experiments for monitoring gene expression.

B. EXPERIMENTAL DESIGN FOR MICROARRAY-BASED MONITORING OF GENE EXPRESSION Microarray experiments generate massive data sets, which must be analyzed and interpreted in a rapid and meaningful way. To improve the efﬁciency and reliability of experimental data, careful experimental design is needed. Without this, the collected data may fail to answer the research question of interest or lead to a biased, inadequate interpretation of the experimental results (Yang and Speed, 2002). The main objective of experimental design is to make the data analysis and interpretation as simple and powerful as possible. For a competitive microarray hybridization experiment in which two ﬂuorescent dyes are used, the most important experimental design issue is how the mRNAs are labeled and which mRNAs are hybridized together on the same slide (Yang and Speed, 2002). In most experiments, several designs can be devised. The selection of the most appropriate design will depend on the particular research questions being asked, the number of comparisons, the number of slides available for hybridization, the amount of mRNAs available, and cost. Various design schemes have been described in great detail by Yang and Speed (2002) and several designs could be devised for a particular microarray experiment. The microarray experiment design scheme can be classiﬁed into the three categories (Fig. 6): reference design, all-pairs design, and loop design. In reference design, all treatment samples are labeled with one dye and are hybridized, respectively, with a common reference sample labeled with another dye (Fig. 6A). This indirect design is used widely in gene expression studies. This design is especially suitable when the amount of mRNA from treatment samples is limited and when many treatment samples are compared. Another advantage of this design is that data analyses and interpretation are easy and do not require sophisticated statistical tools. However, the average variance for this indirect reference design is considerably higher than that for the other designs. Since it is straightforward, the reference design is used much more often than the other designs.

MICROARRAY TECHNOLOGY

235

Figure 6 Illustrations of basic types of microarray experimental design schemes with ﬁve treatment samples. By convention, the green-labeled sample (Cy3) is placed at the tail while the redlabeled sample (Cy5) is placed at the head of the arrow. (A) Reference design. The ﬁve treatment samples (A –E) are labeled with one dye and hybridized, respectively, with the common reference sample R, which is labeled with the other dye. Altogether ﬁve hybridizations are needed. (B) All-pair design. Each sample is labeled twice with red and twice with green. Ten pair-wise hybridizations are needed. (C) Loop design. Each sample is labeled once with red and once with green. Five successive pair hybridizations are needed.

In the all-pairs design scheme, all of the treatment samples are labeled with different ﬂuorescent dyes and directly hybridized together in pair-wise fashion (Fig. 6B). The main advantage of this design is that more precise comparisons among different treatment samples can be obtained. However, this design is unlikely to be feasible and desirable when a large number of comparisons are made, due to the constraints on mRNA quantity and cost. Finally, in the loop design, all of the treatments are successively connected as a loop (Fig. 6C) (Kerr and Churchill, 2001a, b). Using the same number of microarrays as the reference design, the loop design obtains twice as much data on the treatments of interest.

236

J. ZHOU AND D. K. THOMPSON

The loop design requires far fewer slides than the all-pairs design. However, long paths between some pairs of treatment samples are needed in larger loops, and thus some comparisons are much less precise than others (Yang and Speed, 2002). Another practical problem is that each sample should be labeled with both Cy dyes, which doubles the number of labeling reactions. In addition, the failure of microarray hybridization in one sample will affect the analysis of other samples in the loop.

C. MICROARRAY-BASED FUNCTIONAL ANALYSIS ENVIRONMENTAL MICROORGANISMS

OF

Genome sequence information for a number of bacteria and archaea of potential environmental or biotechnological relevance is accumulating rapidly and includes representatives from such genera as dissimilatory metal-reducing bacteria (Shewanella oneidensis [Heidelberg et al., 2002]), extreme radiationresistant bacteria (Deinococcus radiodurans [White et al., 1999]), photosynthetic cyanobacteria (Anabaena sp. strain PCC 7120 [Kaneko et al., 2001], Synechocystis sp. strain PCC6803 [Kaneko et al., 1996]), thermophilic and hyperthermophilic archaea (Pyrococcus horikoshii [Kawarabayasi et al., 1998], Aeropyrum pernix [Kawarabayasi et al., 1999], Thermotoga maritima [Nelson et al., 1999], Thermoplasma volcanium [Kawashima et al., 2000], Pyrococcus furiosus [Robb et al., 2001], Pyrobaculum aerophilum [Fitz-Gibbon et al., 2002]), thermoacidophilic archaea (Sulfolobus tokodaii [Kawarabayasi et al., 2001], Sulfolobus solfataricus [She et al., 2001]), methanogens (Methanococcus jannaschii [Bult et al., 1996], Methanobacterium thermoautotrophicum [Smith et al., 1997], Methanopyrus kandleri [Slesarev et al., 2002], Methanosarcina acetivorans [Galagan et al., 2002]), sulfate-reducing archaea (Archaeoglobus fulgidus [Klenk et al., 1997]), and halophilic archaea (Halobacterium species NRC-1 [Ng et al., 2000]). However, to-date, very few studies have explored the transcriptomes of these organisms using microarray technology. The large majority of microarray-based genomic expression analyses have focused on bacterial pathogens and such model organisms as E. coli, B. subtilis, and yeast. In this section, we will brieﬂy discuss microarray proﬁling of gene expression in three organisms of environmental signiﬁcance, namely, S. oneidensis, D. radiodurans, and P. furiosus, as examples of the application of microarrays to environmental microbiology.

1.

Shewanella oneidensis, a Dissimilatory Metal-reducing Bacterium

S. oneidensis MR-1 (formerly Shewanella putrefaciens strain MR-1 [Venkateswaran et al., 1999]) is a facultatively anaerobic g-proteobacterium

MICROARRAY TECHNOLOGY

237

that is noted for its remarkably diverse respiratory capacities. In addition to utilizing oxygen as a terminal electron acceptor during aerobic respiration, S. oneidensis can anaerobically respire various organic and inorganic substrates, including oxidized metals (e.g., Mn(III) and (IV), Fe(III), Cr(VI), U(VI)), fumarate, nitrate, nitrite, thiosulfate, elemental sulfur, trimethylamine N-oxide, DMSO, and anthraquinone-2,6-disulphonate (Lovley, 1991; Nealson and Saffarini, 1994; Moser and Nealson, 1996). This unusual versatility in the use of alternative electron acceptors for anaerobic respiration is conferred in part by complex electron transport networks, the components of which remain to be elucidated (Richardson, 2000). The metal ion-reducing capabilities of this bacterium, in particular, have important implications with regard to the potential for in situ bioremediation of metal contaminants in the environment. However, the effective prediction and assessment of bioremediation performance or activity is complicated due to insufﬁcient knowledge concerning the gene networks and regulatory mechanisms enabling microbial metal reduction. To expedite understanding of metal reduction by S. oneidensis MR-1, its ~5-Mb genome was determined recently by The Institute for Genomic Research (TIGR) under the support of the U.S. Department of Energy (DOE) (Heidelberg et al., 2002), making it feasible to apply microarray technology to the study of energy metabolism in this bacterium. The transcriptional response of S. oneidensis to different respiratory growth conditions (Beliaev et al., 2002b) and to the disruption (inactivation) of genes encoding putative transcriptional regulators (Beliaev et al., 2002a; Thompson et al., 2002) were examined using DNA microarrays containing 691 arrayed genes. These partial genome microarrays consisted of PCR-ampliﬁed MR-1 ORFs putatively involved in energy metabolism, transcriptional regulation, adaptive responses to environmental stress, iron acquisition, and transport systems according to the sequence annotation. These arrays were constructed prior to the closure and publication of the S. oneidensis genome sequence. To identify genes speciﬁcally involved in anaerobic respiration, differential mRNA expression proﬁles of S. oneidensis were monitored under aerobic and fumarate-, Fe(III)-, or nitrate-reducing conditions using partial genome microarrays (Beliaev et al., 2002b). Gene expression proﬁling indicated that 121 of the 691 arrayed ORFs showed at least a 2-fold difference in mRNA abundance in response to changes in growth conditions (Beliaev et al., 2002b), with a number of genes required for aerobic growth being repressed in the transition from aerobic to anaerobic respiration. Genes induced in a general response to anaerobic respiration, irrespective of the terminal electron acceptor, belonged to several different categories of cellular function: cofactor biosynthesis and assembly, substrate transport, and anaerobic energy metabolism. Of particular importance was the observation that certain genes preferentially displayed increased transcript levels in response to speciﬁc electron acceptors. For example, the expression of genes encoding a periplasmic nitrate reductase

238

J. ZHOU AND D. K. THOMPSON

(napBHGA operon), cytochrome c552, and prismane was elevated 8- to 56-fold speciﬁcally in response to the presence of nitrate, while genes encoding a tetraheme cytochrome c (cymA), a ﬂavocytochrome c (ifcA), and a fumarate reductase ( frdA) were preferentially induced 3- to 8-fold under conditions of fumarate reduction. In addition, the mRNA abundance levels for two oxidoreductaserelated genes of unknown function and several cell envelope genes involved in multidrug resistance increased speciﬁcally under Fe(III)-reducing conditions. This work represented the ﬁrst attempt to characterize a complex system in S. oneidensis on a genome scale. Other microarray-based transcriptomic studies have focused on deﬁning the functions of putative S. oneidensis regulatory genes encoding a ferric uptake regulator ( fur; Thompson et al., 2002) and an electron transport regulator (etrA; Beliaev et al., 2002a).

2.

Deinococcus radiodurans, an Extreme Radiation-resistant Bacterium

D. radiodurans strain R1 is the most characterized member of the DNAdamage resistant bacterial family Deinococcaceae, which is comprised of at least seven different species that form a distinct eubacterial phylogenetic lineage (Makarova et al., 2001). D. radiodurans is a Gram-positive, non-sporulating bacterium that was originally isolated in 1956 from canned meat that had spoiled following exposure to X-ray sterilization (Anderson et al., 1956). Species in the genus Deinococcus, particularly D. radiodurans, are extremely resistant to a number of physicochemical agents and environmental conditions that damage DNA, including ionizing and ultraviolet radiation, desiccation, heavy metals, and oxidative stress (reviewed in Minton, 1996; Battista, 1997; Battista et al., 1999). Studies have demonstrated that D. radiodurans can survive acute exposures to gamma radiation that exceed 15,000 Gy without lethality or induced mutation (Daly et al., 1994; Daly and Minton, 1995) and ﬂourish in the presence of highlevel chronic irradiation (60 Gy/h) (Lange et al., 1998; Venkateswaran et al., 2000). D. radiodurans also expresses an intrinsic ability to reduce metals and radionuclides (Fredrickson et al., 2000) and thus has potential applications for the bioremediation of metal- and radionuclide-contaminated sites where the presence of radioactivity prohibitively restricts the activity of more sensitive dissimilatory metal-reducing bacteria such as Shewanella. To enhance the understanding of the molecular basis of extreme DNA damage resistance, the complete genome of D. radiodurans R1 was sequenced by TIGR (White et al., 1999) under DOE support. Sequence analysis of this organism’s multigenomic content indicates that essentially the entire repertoire of recombinational DNA repair genes identiﬁed in D. radiodurans has functional homologs in other prokaryotes (White et al., 1999; Makarova et al., 2001), suggesting that the extreme radioresistance of R1 may be attributable to novel genes, repair pathways, and mechanisms yet to be described.

MICROARRAY TECHNOLOGY

239

Detailed computational genomic analyses alone, therefore, are unlikely to uncover the fundamental answers underlying the remarkable ability of D. radiodurans to withstand DNA-damaging conditions. The transcriptome dynamics of D. radiodurans in response to cellular recovery from acute ionizing radiation was examined using DNA microarrays covering ~94% of the organism’s predicted protein-encoding genes (Liu et al., 2003). In this time-series study, D. radiodurans cells exposed to acute irradiation (15 kGy) were allowed to recover at 378C for time intervals ranging from 0 to 24 h. Deinococcus transcriptome dynamics were monitored in cells representing early (0 – 3 h), middle (3 –9 h), and late (9 –24 h) phases of recovery from ionizing radiation and compared to non-irradiated control cells. Microarray analysis of genomic expression patterns revealed a large number of D. radiodurans genes responding to acute irradiation: 832 genes (28% of the genome) were induced and 451 genes (15% of the genome) were repressed two-fold or greater at one point during D. radiodurans recovery (Liu et al., 2003). Genes exhibiting increased transcription in the early phase of cell recovery belonged to a number of broad functional groups, including DNA replication, DNA repair, recombination, cell wall metabolism, cellular transport, and uncharacterized proteins. Hierarchical clustering of genes showing differential expression revealed similar expression patterns for groups of genes and clusters of presumably co-regulated genes (Fig. 7). Genes responding to recovery from irradiation clustered into three distinct groups: (1) recA-like activation pattern (based on the expression proﬁle of recA, which is critical for D. radiodurans recovery and is substantially upregulated during early-phase recovery and down-regulated before the onset of late phase), (2) growth-related activation pattern, and (3) repressed patterns. Unexpectedly, genes encoding tricarboxylic acid (TCA) cycle components were repressed in the early and middle phases of recovery, whereas genes encoding the glyoxylate shunt pathway were induced during this interval (Liu et al., 2003). In addition, a number of poorly characterized genes showed high induction folds in expression during at least one phase of recovery, thus implicating their encoded proteins in the functional role of cell recovery. The response of metabolic gene systems, however, is not immediately clear and will require further, more focused experimentation. The study by Liu et al. (2003) represents the ﬁrst published description of the application of DNA microarrays to the functional analysis of D. radiodurans and suggests that the recovery process for this organism involves the complicated coordination of DNA repair and metabolic functions as well as other cellular functions.

3.

Pyrococcus furiosus, a Hyperthermophilic Archaeon

P. furiosus is a member of a phylogenetically distinct group of prokaryotes called the Archaea, which constitutes a primary, separate domain in the universal

240

J. ZHOU AND D. K. THOMPSON

MICROARRAY TECHNOLOGY

241

tree of life (Woese et al., 1990; Olsen and Woese, 1997). The Archaea domain is composed of organisms with diverse phenotypes, such as methane-producing methanogens, extreme halophiles, and extremely thermophilic sulfur-metabolizing species (Woese, 1987). Typically, archaeal genes involved in energy production, cell division, cell wall biosynthesis, and metabolism have homologs in bacteria, whereas genes encoding proteins that function in the informational processes of DNA replication, transcription, and translation are more similar to their eucaryal counterparts (Bult et al., 1996). Archaea also share certain RNA processing components with Eucarya, such as ﬁbrillarin (a pre-rRNA processing protein) (Bult et al., 1996; Belfort and Weiner, 1997) and tRNA splicing endonucleases (Belfort and Weiner, 1997; Kleman-Leyer et al., 1997). The mosaic nature of archaea makes this group of organisms extremely interesting from an evolutionary perspective. The sequencing and analysis of archaeal genomes should provide valuable insights into the origin or evolution of eukaryotes, as well as the molecular mechanisms enabling their adaptation to extreme environments. The hyperthermophilic archaeon P. furiosus is able to grow optimally at a temperature of 1008C (Fiala and Stetter, 1986). Studies support a highly regulated fermentative-based metabolism in P. furiosus (Adams et al., 2001), which can utilize the disaccharide maltose in the presence or absence of elemental sulfur (S0). In addition, P. furiosus can couple the reduction of S0 to the oxidation of catabolism-generated, reduced ferredoxin, but the molecular mechanism of this metabolic coupling is not presently known (Schut et al., 2001). The availability of the complete genome sequence of P. furiosus (Robb et al., 2001) permits the global analysis of gene function and expression using high-density DNA microarray technology. To investigate the molecular basis of S8 metabolism, Schut et al. (2001) used DNA microarrays containing 271 ORFs (of the ca. 2200 total ORFs predicted) from the P. furiosus genome (1.9 Mb) to measure

Figure 7 Hierarchical clustering analyses of expression proﬁle patterns. Gene expression patterns are displayed graphically. Three distinct patterns are sorted according to the hierarchical clustering analyses, i.e., (A) recA-like activation pattern, (B) growth-related activation pattern, and (C) repressed patterns. The top row represents the general pattern of the selected group where a Pearson correlation coefﬁcient (r) is shown on the left side. All displayed graphs are organized in a row/column format. Each row of colored strips represents a single gene whose expression levels are color-recorded sequentially in every column of the same row that represents recovery time intervals. Red color denotes up-regulation, whereas green indicates down-regulation. Black indicates the control level. The variation in transcript abundance is positively correlated with the color darkness. (a) Gene numbers are offered for tracking the primary information of the gene of interest. (b) The maximum (for recA-like and growth-related activation pattern) or minimum (for the repressed pattern) expression level for each of the exhibited genes over the 24-h recovery period is presented as the dye intensity ratio of the irradiated sample to the non-irradiated control at (c) the indicated time interval. Values in parentheses show the SD for each mean expression ratio (Courtesy of PNAS).

242

J. ZHOU AND D. K. THOMPSON

differential gene expression in cells grown at 958C on maltose in the presence or absence of S8. The arrayed PCR products represented ORFs with proposed functions in sugar and peptide catabolism, metal utilization, and the biosynthesis of cofactors, amino acids, and nucleotides. This study by Schut et al. (2001) represents the ﬁrst and, to-date, only account describing the application of DNA microarray analysis to a member of the Archaea. Currently published genomic analyses of archaea have been almost exclusively limited to the sequencing, annotation, and in silico comparative analysis of archaeal genomes. DNA microarray analysis revealed a number of ORFs whose expression was dramatically down-regulated (. 5-fold decrease) by S8, including 18 genes encoding various subunits associated with three different hydrogenase systems (Schut et al., 2001). Other genes displaying decreased transcription, when P. furiosus cells were grown with S8, encoded a hypothetical protein and two homologs (ornithine carbamoyltransferase and HypF) involved in hydrogenase biosynthesis. In the presence of S8, the expression of two previously uncharacterized ORFs (encoding products designated SipA and SipB for “sulfur-induced proteins”) increased by a striking . 25-fold. The encoded proteins of these ORFs were proposed by the authors to be part of a novel S8-reducing, membrane-associated, iron-sulfur cluster-containing complex in P. furiosus (Schut et al., 2001). The research reported by Schut et al. (2001) clearly illustrates the power of DNA microarray analysis in generating new lines of experimentation and in implicating previously uncharacterized ORFs identiﬁed by genome sequencing in biological processes. There is little doubt that the continuing determination of archaeal genomes will spawn more microarray-based functional studies of extremophiles.

VIII. APPLICATION OF MICROARRAYS TO ENVIRONMENTAL STUDIES In addition to monitoring transcription patterns on a genomic scale, microarray-based technology is well suited for detecting microorganisms in natural environments. Many target functional genes involved in biogeochemical cycling in environments are highly diverse, and it is difﬁcult or impossible to identify conserved regions for designing PCR primers or oligonucleotide probes. The microarray-based approach does not require such sequence conservation, because all of the diverse gene sequences from different populations of the same functional group can be fabricated on arrays and used as probes to monitor their corresponding populations. In contrast to studies using pure cultures, microarray analysis of environmental nucleic acids presents a number of technical challenges that must be overcome. First, target and probe sequences in environmental samples can be

MICROARRAY TECHNOLOGY

243

very diverse, and it is not clear whether the performance of microarrays with diverse environmental samples is similar to that with pure culture samples and how sequence divergence is reﬂected in hybridization signal intensity. Second, environmental samples are generally contaminated with substances such as humic matter, organic contaminants, and metals, which may interfere with the hybridization reaction on microarrays. Third, in contrast to pure cultures, the recoverable biomass in environmental samples is generally low; consequently, it is not clear whether microarray hybridization is sensitive enough to detect microorganisms in all types of environmental samples. Finally, it is uncertain whether microarray-based detection can be quantitative. Environmental and ecological studies require experimental tools that not only detect the presence or absence of particular groups of microorganisms but also provide quantitative data on their in situ biological activities. In the following sections, we discuss three different types of microarray formats that have been developed for use in environmental studies: functional gene arrays (FGAs), phylogenetic oligonucleotide arrays (POAs), and community genome arrays (CGAs).

A. FUNCTIONAL GENE ARRAYS Genes encoding functional enzymes involved in various biogeochemical cycling processes (e.g., carbon, nitrogen, sulfate and metals) are very useful as molecular signatures for assessing the physiological status and functional activities of microbial populations and communities in natural environments. Microarrays containing functional gene sequence information are referred to as FGAs, because they are primarily used for the functional analysis of microbial community activities in environments (Wu et al., 2001). Similar to gene expression proﬁling arrays, both oligonucleotides and PCR-ampliﬁed DNA fragments corresponding to functional genes can be used for fabricating FGAs.

1.

Selection of Gene Probes

FGAs are designed for studying functional gene diversity in natural environments. To construct FGAs, the gene probes should be carefully deﬁned and selected based on the speciﬁc research questions to be addressed. As an example, microarrays can consist of gene probes that are involved in such biogeochemical processes as nitriﬁcation (ammonia monooxygenase, amoA), denitriﬁcation (nitrite reductases, nirS and nirK), nitrogen ﬁxation (nitrogenases, nifH), sulﬁte reduction (sulﬁte reductase, dsvA/B), methanogenesis

244

J. ZHOU AND D. K. THOMPSON

(methyl coenzyme M reductase genes, mcrA), methane oxidation (methane mono-oxygenases, mmo), and plant and fungal polymer degradation (cellulases, xylanases, ligin peroxidases, chitinases). Probes for the construction of FGAs can be generated in three ways. The desired gene fragment can be ampliﬁed from genomic DNA extracted from pure bacterial cultures using speciﬁc primers or from cloned plasmids containing the desired gene insert using vector-speciﬁc primers. However, the availability of pure cultures and plasmid clones can be limited. In the second approach, desired gene fragments are recovered from natural environments using PCR-based cloning methods (Zhou et al., 1997). Generally, sequences that show . 85% identity can be used as speciﬁc probes for FGAs. These two approaches were used to construct FGAs containing nitrite reductase genes and ammonia monooxygenase genes for monitoring bacteria involved in nitriﬁcation and denitriﬁcation, respectively (Wu et al., 2001). Finally, in the third strategy, oligonucleotides, usually 50– 70-mers, are designed based on the functional sequences available in public databases and synthesized for microarray fabrication (Tiquia et al., unpublished).

2.

Speciﬁcity

Hybridization speciﬁcity is an important parameter that impacts any detection method. It is inﬂuenced by many factors, such as G þ C content, degree of sequence divergence, sequence length, secondary structure of the probe, temperature, and salt concentrations. To determine the speciﬁcity of DNA microarray hybridization, we have constructed and used FGAs consisting of heme- and copper-containing nitrite reductase genes, ammonia monooxygenase (amoA), and methane monooxygenase genes [ pmoA] (Wu et al., 2001). Small subunit (SSU) rRNA genes and yeast genes were used as positive and negative controls, respectively, on the FGAs. Cross-hybridization among different gene groups was not observed at either low (458C) or high (658C) stringency. Furthermore, no hybridization was observed with any of the ﬁve yeast genes, which served as negative controls for hybridization on the array (Fig. 8A). Based on the sequence similarities, it was estimated that microarray hybridization can differentiate between sequences exhibiting a dissimilarity of approximately 15% at 658C and 10% at 758C (Wu et al., 2001). In addition, at low stringency, most nirS, nirK or amoA genes hybridized well with their respective homologous target DNA, suggesting that a broad range of detection can be achieved by adjusting the conditions for microarray hybridization. These results indicate that speciﬁc hybridization can be achieved using the glass slide-based microarray format with bulk community DNA extracted from environmental samples.

MICROARRAY TECHNOLOGY

245

Figure 8 Speciﬁcity and sensitivity of DNA fragments-based FGAs. (A) Fluorescence images showing the speciﬁcity of nirS in DNA microarray hybridization. Target DNA was labeled with either Cy3 from a pure culture using the method of PCR ampliﬁcation and hybridized separately at high stringency (658C) to FGAs containing nirS, nir K, and amoA gene probes from both pure bacterial cultures and environmental clones. The 16S rRNA and yeast genes served as positive and negative controls, respectively. (B) Array hybridization images showing the detection sensitivity with labeled pure genomic DNA Genomic DNA from a pure culture of nirS-containing P. stutzeri E4-2 was labeled with Cy5 using the random primer labeling method. The target DNA was hybridized to the microarrays at total concentrations of 0.5, 1, and 5 ng.

To determine the potential performance of oligonucleotide microarrays for environmental studies, an FGA consisting of 50-mer oligonucleotide probes was constructed and evaluated using 1033 genes involved in nitrogen cycling (nirS, nir K, nif H, amoA, and pmoA) and sulﬁte reduction (dsrA/B) from public databases and our own sequence collections (Tiquia et al, unpublished). Under hybridization conditions of 508C and 50% formamide, genes having , 86 –90% sequence identity were clearly distinguished. As expected, the oligonucleotide-based FGAs showed a higher degree of hybridization speciﬁcity than the DNA-based FGAs. Comparison of probe sequences from pure cultures of bacteria involved in nitriﬁcation, denitriﬁcation, nitrogen ﬁxation, methane oxidation and sulfate reduction indicated that the average similarity of these functional genes at the species level ranged from 74 to 84%. These results suggest that the 50-mer FGAs could provide species-level resolution for analyzing microorganisms involved in these biogeochemical processes.

246

J. ZHOU AND D. K. THOMPSON

Compared to DNA-based FGAs, the 50-mer oligonucleotide arrays offer the following main advantages. First, higher hybridization speciﬁcity can be achieved because the probe sizes in oligo arrays are much smaller than those used in the DNA-based FGAs. Thus, this type of environmental array could provide a higher level of resolution in differentiating microbial populations. Since the probes can be directly designed and synthesized based on sequence information from public databases, construction of oligonucleotide arrays is much easier than that of the DNA-based FGAs. To construct microarrays containing large DNA fragments, the probes used for array fabrication are generally ampliﬁed by PCR from environmental clones or from pure genomic DNA. However, obtaining all the diverse environmental clones and bacterial strains from various sources as templates for ampliﬁcation can be a considerable challenge. As a result, the construction of comprehensive microarrays representing all functional genes of interest is not practically feasible. With oligo arrays, a great number of genes can easily be arrayed for comprehensive survey of the populations and activities of diverse microbial communities in the environment. In addition, since no PCR ampliﬁcation is involved in oligonucleotide microarray fabrication, potential cross-contamination due to PCR ampliﬁcation is minimized. 3.

Sensitivity

Sensitivity is another critical parameter that impacts the effectiveness of microarray-based detection of microorganisms. The detection sensitivity of hybridization with a prototype DNA-based FGA was determined using genomic DNA from both pure cultures and soil community samples. At high stringency, strong hybridization signals were observed with 5 ng of DNA for both nirS and SSU rRNA genes, whereas hybridization signals were weaker but detectable with 1 ng of DNA (Fig. 8B). The hybridization signals at low DNA concentrations were stronger for SSU rRNA genes than for nirS genes. Hybridization signals derived from 0.5 ng of genomic DNA were measurable, but the ﬂuorescence intensity was poor. As a result, the detection limit was estimated to be approximately 1 ng with randomly labeled pure genomic DNA under the tested hybridization conditions. The detection sensitivity of FGA hybridization was also evaluated using community genomic DNA isolated from surface soil that contained a high level of chromium and organic matter. All of the arrayed genes, with the exception of the ﬁve yeast genes, showed hybridization with 50 and 25 ng of labeled community DNA. Only the SSU rRNA genes could be detected when as little as 10 ng of the soil community DNA was used in the hybridization reaction. Thus, the detection sensitivity of nirS and SSU rRNA genes in this soil sample was considered to be approximately 25 and 10 ng of the total environmental DNA, respectively. These approximate levels of detection sensitivity should be

MICROARRAY TECHNOLOGY

247

sufﬁcient for many studies in microbial ecology and suggest that microarray hybridization can be used as a sensitive tool for analyzing microbial community composition in environmental samples. The detection limit with 50-mer FGAs was approximately 8 ng of pure genomic DNA. As expected, the sensitivity of the 50-mer FGAs is 10 times lower than the DNA-based FGAs and 100 times lower than CGAs (Tiquia et al., unpublished; Zhou, 2003), which are discussed below. One of the main reasons for the lower sensitivity of the 50-mer FGAs is that the oligonucleotide probes are much shorter than the probes used in DNA-based FGAs and CGAs, which have more binding sites available for capturing the labeled target DNAs. In addition, good hybridizations were obtained with the 50-mer FGAs using 2 mg of bulk community DNA from marine sediments. These results suggest that the amount of DNA sample should not be a major limiting factor in using this type of microarray for environmental studies, because the average DNA yields generally range from 10 to 400 mg of DNA per g (dry weight) for many surface soil and sediment samples. Although sensitive detection can be obtained with microarray hybridization, the detection sensitivity is dependent on reagents, especially the ﬂuorescent dyes. We found that the sensitivity varies greatly with different batches of ﬂuorescent dyes. In addition, the sensitivity with direct microarray hybridization may still be 100 to 10,000-fold less than with PCR ampliﬁcation. Microarray hybridization is still not sensitive enough for some environmental studies where the amount of recoverable biomass is very low, thus requiring the development of more sensitive methods.

4.

Quantitation

Many environmental and ecological studies require quantitative data on the in situ abundance and biological activities of microbial communities. The accuracy of microarray-based quantitative assessment is still uncertain because of the inherently high variation associated with array fabrication, probe labeling, hybridization, and image processing. Comparison of microarray hybridization results with previously known results suggested that microarray hybridization appears to be quantitative enough for detecting differences in gene expression patterns under various conditions (DeRisi et al., 1997; Lockhart et al., 1996; Taniguchi et al., 2001). DNA microarrays have also been used to measure differences in DNA copy number in breast tumors (Pinkel et al., 1998; Pollack et al., 1999) and to detect single-copy deletions or additions (Pollack et al., 1999), suggesting that microarray-based detection is potentially quantitative. A recent study in which lambda (l) DNA was co-spotted with DNA from reference

248

J. ZHOU AND D. K. THOMPSON

bacterial strains also indicated that microarrays could accurately quantify genes in DNA samples (Cho and Tiedje, 2001). To evaluate whether microarray hybridization can be used as a quantitative tool to analyze environmental samples, the relationship between target DNA concentration and hybridization signal was examined with DNA-based FGAs (Wu et al., 2001). A strong linear relationship (r 2 ¼ 0.96) was observed between signal intensity and target DNA concentration with DNA from a pure bacterial culture within 1 to 100 ng. Similar to the DNA-based FGAs, a strong linear relationship was observed using the 50-mer oligonucleotide FGAs between signal intensity and target DNA concentrations from 8 to 1000 ng for all six different functional gene groups (r 2 ¼ 0.96 – 0.98) (Tiquia et al., unpublished). These results suggest that microarray hybridization is quantitative for pure bacterial cultures within a limited range of DNA concentration. With our optimized protocol, experimental variation between array slides can be reduced to below 15% with environmental samples (Wu et al., 2001). This is consistent with the ﬁndings of microarray studies on gene expression (Bartosiewicz et al., 2000). Since environmental samples contain a mixture of target and non-target templates, the presence of other non-target templates could affect microarraybased quantiﬁcation. To determine whether microarray hybridization is quantitative for targeted templates within the context of environmental samples, 11 different genes, exhibiting less than 80% sequence identity, were labeled and hybridized with the DNA-based FGAs. For this mixed DNA population, a linear relationship (r 2 ¼ 0.94) was observed between signal intensity and target DNA concentration (Fig. 9), further suggesting that microarray hybridization holds promise as a quantitative tool for studies in environmental microbiology. The target genes within functional groups present in environmental samples may have different degrees of sequence divergence. Such sequence differences will affect microarray hybridization signal intensities and hence its quantitative power. Although it was shown that microarray hybridization could be used to quantify mixed DNA templates, the difﬁcult challenge in quantifying the abundance of microbial populations in natural environments, based on hybridization signal intensity, is how to distinguish differences in hybridization intensity due to population abundance from those due to sequence divergence. One possible solution is to carry out microarray hybridization under conditions of varying stringency. Based on the relationships among signal intensity, sequence divergence, hybridization temperature, and washing conditions, it should be possible to distinguish, to some extent, the contributions of population abundance and sequence divergence to hybridization intensity (Wu et al., 2001). For instance, Wu et al. (2001) showed that at about 55 – 608C, sequence divergence had little or no effect on signal intensity for amoA genes with greater than 80% identity to the labeled target DNA. This suggests that under such hybridization conditions the effect of sequence divergence on signal intensity is negligible for genes with . 80% sequence identity; therefore, any signiﬁcant differences in

MICROARRAY TECHNOLOGY

249

Figure 9 Relationship of hybridization signal intensity to DNA target concentration using a mixture of target DNAs. The PCR products from the following nine strains were mixed together in different quantities (pg): E4-2 (nirS), 1000; G179 (nirK), 500; wc301-37 (amoA), 250; ps-47 (amoA), 125; pB49 (nirS), 62.5; Y32K (nirK), 31.3; wA15 (nirS), 15.6; ps-80 (amoA), 7.8; wB54 (nirK), 3.9. All of these genes are less than 80% identical. The mixed templates were labeled with Cy5. The plot shows the log-transformed average hybridization intensity versus the log-transformed target DNA concentration for each strain. The target DNA was prepared by labeling MR-1 genomic DNA with Cy5 using Klenow fragment with random hexamer primers. The data points are mean values derived from three independent microarray slides, with three replicates on each slide (nine data points). Error bars showing the SD are presented.

signal intensity are most likely due to differences in population abundance. Another possible solution to this problem is to use microarrays containing probes that are extremely speciﬁc to the target population of interest, such as those used in oligonucleotide microarrays.

5.

Applications

FGAs for microbial detection are still in the developmental stages, and thus their applications are still being explored. To demonstrate the applicability of DNA microarrays for microbial community analysis, Wu et al. (2001) used FGAs to analyze the distribution of denitrifying and nitrifying microbial populations in marine sediment and soil samples. The prototype FGA revealed differences in

250

J. ZHOU AND D. K. THOMPSON

the apparent distribution of nirS, nir K and amoA/pmoA gene families in sediment and soil samples. Recently, a 70-mer oligonucleotide microarray containing 64 nirS genes (14 from cultured microorganisms and 50 from environmental clones) was evaluated for studying functional gene diversity in the Choptank River-Chesapeake Bay system (Taroncher-Oldenburg et al., 2003). Signiﬁcant differences in the hybridization patterns were observed between the sediment samples from two stations in the Choptank River. The changes in the nirS-containing denitriﬁer population could have been caused by differences in salinity, inorganic nitrogen and dissolved organic carbon between these two stations. So far, very limited studies have been carried out to evaluate speciﬁcity, sensitivity, sequence divergence and quantitation of DNA microarrays for environmental applications. While this tool is potentially valuable for environmental studies, more development is needed, especially for improved sensitivity, quantitation, and the biological meaning of a detectable speciﬁcity before it can be used broadly and interpreted meaningfully within the context of microbial ecology.

B. PHYLOGENETIC OLIGONUCLEOTIDE ARRAYS Ribosomal RNA genes are powerful molecules for studying phylogenetic relationships among different organisms and for analyzing microbial community structure in natural environments, because these genes exist in all organisms and contain both highly conserved and highly variable regions, which are useful for differentiating microorganisms at different taxonomic levels (e.g., kingdom, phyla, family, genus, species, and strain). A very large database of ribosomal RNA genes exists (http://www.cme.msu.edu), making them ideal molecules for developing microarray-based detection tools. In addition, cells generally have multiple copies of rRNA genes, and the majority (. 95%) of total RNA isolated from samples is rRNA. Consequently, the detection sensitivity will be higher for rRNA genes than for functional genes. Therefore, rRNA genes are very useful targets for developing microarray-based detection approaches. Oligonucleotide microarrays containing information from rRNA genes are referred to as phylogenetic oligonucleotide microarrays (POAs), because such microarrays are used primarily for phylogenetic analysis of microbial communities. The POAs can be constructed for different phylogenetic taxa and used in microbial community analysis studies. The oligonucleotide probes can be designed in a phylogenetic framework to survey different levels of sequence conservation, from highly conserved sequences giving broad taxonomic groupings to hypervariable sequences giving genus- and potentially species- level groupings. Because highly conserved universal primers for amplifying rRNA

MICROARRAY TECHNOLOGY

251

genes are available, POA-based hybridization can be easily coupled with PCR ampliﬁcation, thus enabling the implementation of highly sensitive assays.

1.

Challenges of Plylogenetic Oligonucleotide Arrays

Non-rRNA gene-based oligonucleotide microarrays have been used successfully for monitoring genome-wide gene expression (e.g., Lockhart et al., 1996; de Saizieu et al., 1998) and detecting genetic polymorphisms (e.g., Wang et al., 1998). In contrast, rRNA gene-based oligonucleotide arrays present some unique technical challenges concerning hybridization speciﬁcity and sensitivity (Zhou and Thompson, 2002; Zhou, 2003). Speciﬁcity. Since rRNA genes are highly conserved and present in all microorganisms, speciﬁc detection with rRNA-targeted oligonucleotide microarrays can be difﬁcult. First, the probe length and G þ C content can signiﬁcantly impact microarray hybridization (Guschin et al., 1997a). Second, probe selection is limited by the sequence differences among target genes, and crosshybridization can be a problem. Oligonucleotide microarrays typically contain many probes. Ideally, all of the oligonucleotides should have similar or identical melting kinetics, so that all of the probes on an array element can be subjected to the same hybridization conditions at once. This can be difﬁcult to achieve, because the melting temperature depends on the length and composition of the oligonucleotide probe as well as the target 16S rRNA molecules in the samples. Secondary structure. The hybridization of oligonucleotide probes to target nucleic acids possessing stable secondary structures can be particularly challenging, since low stringency conditions (i.e., hybridization temperatures between 0 –308C) are required for stable association of a long target nucleic acid with a short immobilized oligonucleotide probe (Drobyshev et al., 1997; Guschin et al., 1997a, b; Southern et al., 1999). Any stable secondary structure of the target DNA or RNA must be overcome in order to make complementary sequence regions available for duplex formation. The stable secondary structure of SSU rRNA will have serious effects on hybridization speciﬁcity and detection sensitivity.

2. Speciﬁcity and Sensitivity In a study by Guschin et al. (1997a), gel-pad oligonucleotide microarrays were constructed using oligonucleotides complementary to SSU rRNA sequences from key genera of nitrifying bacteria. The results showed that speciﬁc detection could be achieved with this type of microarray. However, the probe speciﬁcity depends on various factors, such as probe length. Guschin et al. (1997a) showed that, as

252

J. ZHOU AND D. K. THOMPSON

the length of the oligonucleotide probe increases, mismatch discrimination is lost; conversely, as the length of the probe decreases, hybridization signal intensity (i.e., sensitivity) is sacriﬁced. A recent study showed that gel-pad-based oligonucleotide microarrays could also be used to distinguish between B. thuringiensis and B. subtilis (Bavykin et al., 2001). Using glass-based twodimensional microarrays, Small et al. (2001) detected such metal-reducing bacteria as Geobacter chapellei and Desulfovibrio desulfuricans. The potential advantage of oligonucleotide probes is that target sequences containing single-base mismatches can be differentiated by microarray hybridization. However, this has not been fully demonstrated with SSU rRNA gene-based probes. To systematically determine whether single mismatch discrimination can be achieved for SSU rRNA genes using microarray hybridization, we constructed a model oligonucleotide microarray consisting of probes derived from three different regions of the SSU rRNA molecule corresponding to different bacterial taxa (X. Zhou and J. Zhou, unpublished data). The probes had 1 – 5 mismatches in different combinations along the length of the oligonucleotide probe with at least one mismatch at the central position. Hybridization signal intensity with a single-base mismatch was decreased by 10 to 30%, depending on the type of mismatched nucleotide base. The signal intensity of probes with two base mismatches was 5 to 25% of that of the perfect match probes. Probes with three or four base-pair mismatches yielded signal intensities that were 5% of that of the perfect match probes. Maximum discrimination and signal intensity was achieved with 19-base probes. These results indicated that single base discrimination for SSU rRNA genes can be achieved with glass slide-based array hybridization, but complete discrimination appears to be problematic with SSU rRNA genes (Bavykin et al., 2001; Small et al., 2001; Urakawa et al., 2002). Urakawa et al. (2002) demonstrated that the single-base-pair near-terminal and terminal mismatches have a signiﬁcant effect on hybridization signal intensity. With SSU rRNA gene-based oligonucleotide microarrays, the level of detection sensitivity obtained using the G. chapellei 16S rRNA gene is about 0.5 mg of total RNA extracted from soils (Small et al., 2001).

3.

Applications

As with all the arrays developed for environmental applications, SSU rRNA gene-based oligonucleotide arrays are still in the early stages of development, and therefore, only a few studies have applied POAs to the analysis of microbial structure within the context of environmental samples. Using photolithographybased Affymetrix technology, Wilson et al. (2002) designed a gene chip (microarray) containing 31,179 and 20-mer oligonucleotide probes speciﬁc for SSU rRNA genes. All of the probes are derived from a small SSU rRNA gene

MICROARRAY TECHNOLOGY

253

region (i.e., E. coli positions 1409 to 1491), which is bound on both ends by universally conserved segments. The gene chip also contained control sequences, which were paired with the probe sequences. A control sequence was identical to the paired probe sequence except that there was a mismatch nucleotide at the 11th position. Thus, the gene chip contained a total of 62,358 features. The number of probes for individual sequences contained in the Ribosomal Database Project (RDP version 5.0, with about 3200 sequences) ranges from 0 to 70. A total of 17 pure bacterial cultures were used to assess the performance of this gene chip, and 15 bacterial species were identiﬁed correctly. However, it failed to resolve the individual sequences comprising complex mixed samples (Wilson et al., 2002). Rudi et al. (2000) constructed a small microarray containing 10 SSU rRNA probes derived from cyanobacteria, and used it to analyze the presence and abundance of these organisms in lakes with both low and high biomass. The probes were speciﬁc to the cultures analyzed, and reproducible abundance proﬁles were obtained with these lake samples. Relatively good qualitative correlations were observed between the community diversity and standard hydrochemical data, but the levels of correlation were lower for the quantitative data. Loy et al. (2002) developed a microarray containing 132 SSU rRNA-targeted oligonucleotide probes, which represented all recognized groups of sulfatereducing prokaryotes. Microarray hybridizations with 41 reference strains showed that, under the hybridization conditions used, clear discrimination between perfectly matched and mismatched probes were obtained for most, but not all of the 132 probes. This microarray was used to determine the diversity of sulfate-reducing prokaryotes in periodontal tooth pockets and a hypersaline cyanobacterial mat. The microarray hybridization results were consistent with those obtained using well-established conventional molecular methods. These results suggest that microarray hybridization is a powerful tool in analyzing community structure but great caution is needed in data interpretation because of the potential for cross-hybridization.

C. COMMUNITY GENOME ARRAYS Decades of scientiﬁc investigations have led to the isolation of many microorganisms from a variety of natural habitats. However, little or nothing is known about the genomic sequences for the majority of these microorganisms. Such a large collection of pure cultures should be very useful for monitoring microbial community structure and composition in natural environments. To exploit such a resource, a novel prototype microarray containing whole genomic DNA, termed community genome array (CGA), was developed and evaluated in my (Zhou’s) laboratory.

254

J. ZHOU AND D. K. THOMPSON

The CGA is conceptually analogous to membrane-based reverse sample genome probing (RSGP) (Voordouw, 1998), but CGA hybridization is distinctly different from RSGP in terms of the arraying substrate and signal detection strategies. In contrast to RSGP, the CGA uses a non-porous (i.e., glass) surface for fabrication and ﬂuorescence-based detection. The capability of accurate and precise miniaturization with robots on non-porous substrates is one of the two key advances of microarray-based genomic technologies. The miniaturized microarray format coupled with ﬂuorescent detection represents a fundamental revolution in biological analysis. Like RSGP, the main disadvantage of the CGA is that only the cultured components of a community can be monitored, because the construction requires the availability of individual pure isolates, even though CGA-based hybridization itself does not require culturing (Voordouw, 1998). With the recent advances in environmental genomics, high-molecular-weight DNA from uncultivated microorganisms could be accessed through bacterial artiﬁcial chromosomes (BACs). BAC clones could also be used to fabricate CGAs, thus allowing the investigation of uncultivated components of a complex microbial community. In the following sections, we will brieﬂy describe the performance of CGA-based hybridization in terms of speciﬁcity, sensitivity and quantitation.

1.

Speciﬁcity

To examine hybridization speciﬁcity under varying experimental conditions and to determine the threshold levels of genomic differentiation, a prototype microarray was fabricated that contained genomic DNA isolated from 67 different representative environmental microorganisms classiﬁed as a-, b-, and g-proteobacteria and Gram-positive bacteria. Many of the selected species are closely related to each other based on SSU rRNA and gyr B gene phylogenies and belong primarily to three major bacterial genera (Pseudomonas, Shewanella, and Azoarcus). The G þ C content of the genomes varies from 37 to 69.3%. By adjusting hybridization temperature and the concentration of additives such as formamide (which increases hybridization stringency), different threshold levels of phylogenetic differentiation could be achieved using the CGAs. For instance, under hybridization conditions of 558C and 50% formamide, strong signals were obtained for genomic DNAs of corresponding species to the labeled target. Little or no cross-hybridization (~0 –4%) was observed for non-target species as well as for negative controls (yeast genes), thus indicating that species-speciﬁc differentiation can be achieved with CGAs under the hybridization conditions used. However, different strains of Pseuodomonas stutzeri, Azoarcus tolulyticus, Bacillus methanolicus, and Shewanella algae could not be clearly distinguished under these conditions (Wu et al., unpublished). By further increasing

MICROARRAY TECHNOLOGY

255

hybridization temperature (65 and 758C), strain-level differentiation was obtained for closely related Azoarcus strains (Wu et al., unpublished). Due to the complicated nature of microarray hybridization, it is unlikely that such assays will completely eliminate some degree of hybridization to non-target strains. The central question is how to distinguish true hybridization signals from non-speciﬁc background noise. One common approach is to determine SNRs and discard values below certain threshold value. Our studies showed that the average SNR for hybridizations with different species within a genus is about 3.35 ^ 0.32, which is substantially lower than hybridizations with different strains from the same species. This value is very close to the commonly used threshold value (SNR ¼ 3.0). CGAs could be used to determine the genetic distance between different bacteria at the taxonomic levels of species and strain. Signiﬁcant linear relationships were observed between CGA hybridization ratios and sequence similarity values derived from SSU rRNA and gyr B genes, DNA-DNA reassociation, or REP- and BOX-PCR ﬁngerprinting proﬁles (r 2 ¼ 0.80 2 0.95) (Wu et al., unpublished), suggesting that CGAs could provide meaningful insights into relationships between closely related strains. Because of its high capacity, one can construct CGAs containing bacterial type strains plus appropriately related strains. By hybridizing genomic DNA from unknown strains with this type of microarray, one should be able to quickly and reliably identify unknown strains provided a suitably related probe is on the array. When using CGAs for strain identiﬁcation, less stringent hybridization conditions (e.g., 458C and 50% formamide) should be used ﬁrst to ensure that good hybridization signals can be obtained for distantly related target species. If multiple probes have signiﬁcant hybridization with the unknown target strains, highly stringent hybridization conditions should then be used. Compared to the traditional DNA –DNA reassociation approach, CGAs have several advantages for determining species relatedness. Since many bacterial genomes can be deposited on microarray slides, the tedious and laborious pairwise hybridizations associated with the traditional DNA – DNA reassociation approach among different species are not needed with CGAs. In contrast to the traditional DNA – DNA reassociation approach, which generally requires about 100 mg DNA, CGA-based hybridization requires only about 2 mg of genomic DNA. This is important for determining the relationships between bacterial species that are recalcitrant to cultivation or grow very slowly.

2. Sensitivity and Quantitative Potential To determine the detection sensitivity of CGAs, genomic DNA from a pure bacterial culture was ﬂuorescently labeled and hybridized with the CGA at

256

J. ZHOU AND D. K. THOMPSON

different concentrations. Under stringent hybridization conditions (i.e., 658C), the detection limit with randomly labeled pure genomic DNA was estimated to be approximately 0.2 ng, whereas genomic DNA concentrations of 0.1 ng were barely detectable above background levels (Wul et al., unpublished). The level of CGA detection sensitivity should be sufﬁcient for many studies in microbial ecology. The detection sensitivity was approximately 10-fold higher than that of DNA-based FGAs and about 100 times higher than that of the 50-mer FGAs. These results were expected, because the CGA probes represent entire genomes rather than a single gene. The capacity of CGA hybridization to serve as a quantitative tool was explored by examining the relationship between the concentration of labeled target DNA and hybridization signal intensity. Quantitative potential was determined using labeled genomic DNA from a single pure culture and from 16 targeted bacteria representing different genera and species. In both cases, strong linear relationships between ﬂuorescence intensity and DNA concentration were observed within a certain range of concentrations (r 2 ¼ 0.92 2 0.95) (Wu et al., unpublished). The results indicate that CGAs can be used for quantitative analysis of microorganisms in environmental samples. The quantitative feature of CGA is similar to those of the DNA- and oligonucleotides-based FGAs (Wu et al., 2001; Tiquia et al., unpublished).

D. WHOLE-GENOME OPEN READING FRAME ARRAYS FOR REVEALING GENOME DIFFERENCES AND RELATEDNESS Many microorganisms that are closely related based on SSU rRNA gene sequences show dramatic differences in phenotypic characteristics. One way to understand the genetic basis for such phenotypic differences is to obtain wholegenome sequence information for all closely related species of interest. Patterns of sequence similarity and variability will provide insights on the conservation of gene functions, physiological plasticity and evolutionary processes. However, sequencing the entire genomes of all closely related species is expensive and time-consuming. In addition, it may not be necessary to sequence all closely related genomes once the complete genome sequence for one representative microorganism is available, because substantial portions of the genomic sequence will be common among closely related species. One way to circumvent the need for sequencing multiple genomes of closely related species is to use DNA microarrays containing individual ORFs of a sequenced microorganism to view genome diversity and relatedness of other closely related microorganisms. The whole-genome ORF array-based hybridization approach has been used to reveal genome diversity and relatedness among closely related organisms in

MICROARRAY TECHNOLOGY

257

several studies. Murray et al. (2001) used this approach to evaluate the genome diversity and relatedness of several related metal-reducing bacteria within the Shewanella genus using partial ORF microarrays for the sequenced metal-reducing bacterium, S. oneidensis MR-1. Both conserved and poorly conserved genes were identiﬁed among the nine species tested. Under the conditions used in this study, the hybridization results were most informative for the closely related organisms with SSU rRNA sequence similarities greater than 93% and gyrB sequence similarities greater than 80%. Above this level of homology, the similarities of microarray hybridization proﬁles were strongly correlated with gyrB sequence divergence. In addition, most genes in operons had high levels of DNA relatedness, suggesting that this approach can be used to identify genes or operons that were horizontally transferred (Murray et al., 2001). Using the ORF arrays for E. coli K-12, Dong et al. (2001) identiﬁed the genes in a common endophyte of maize, Klebsiella pneumoniae 342, which is closely related to E. coli. About 3000 (70%) of E. coli genes were found in strain 342 with greater than 55% identity, whereas about 24% of the E. coli genes were absent in strain 342. The genes with high sequence identity were those involved in cell division, DNA replication, transcription, translation, transport, regulatory proteins, energy, amino acid and fatty acid metabolism, and cofactor synthesis, whereas the genes that are less conserved were involved in carbon compound metabolism, membrane proteins, structural proteins, central intermediary metabolism, and proteins involved in adaptation and protection. Genes that were not identiﬁed in strain 342 included putative regulatory proteins, putative chaperones, surface structure proteins, mobility proteins, putative enzymes and hypothetical proteins. These results on genomic diversity are consistent with the physiological properties of these two strains, suggesting that the microarraybased whole-genome comparison is a powerful approach to revealing the genomic diversity and relatedness of closely related organisms. The whole-genome ORF array approach was also successfully used to identify genome differences among 15 Helicobacter pylori strains with more and less virulence (Salama et al., 2000) and to detect the deletions existing in other strains of Mycobacterium tuberculosis and M. bovis (Behr et al., 1999). All of these studies suggest that whole-genome ORF arrays will be useful for revealing genome difference and relatedness. Whole-genome ORF arrays are available from many microorganisms and they will be valuable for studying genome diversity and relatedness of closely related microorganisms. For example, the whole-genome arrays for six environmentally important microorganisms, including S. oneidensis MR-1, D. radiodurans R1, Rhodopseudomonas palustris, Nitrosomonas europaea, Desulfovibrio vulgaris, and Geobacter metallireducens are available at Oak Ridge National Laboratory, and we are also currently using these whole-genome ORF arrays to understand the genome diversity and relatedness of some important environmental isolates.

258

J. ZHOU AND D. K. THOMPSON

E. OTHER TYPES OF MICROARRAYS FOR MICROBIAL DETECTION AND CHARACTERIZATION DNA microarrays containing random genomic fragments have been used to determine species relatedness in instances where genome sequence information is not available. In this approach, 60– 96 genomic fragments of about 1 kb were randomly selected from four ﬂuorescent Pseudomonas species as reference genomes for microarray fabrication (Cho and Tiedje 2001). Cluster analysis of hybridization proﬁles from 12 well-characterized ﬂuorescent Pseudomonas species indicated that such types of microarray hybridization could provide species to strain level resolution. This approach could have higher resolution than CGA because extensive component information is obtained rather than an average for the whole-genome. However, this approach is more time-consuming and costly to develop than CGA and such an array would be more limited in scope since many of the array positions would be used for each reference microorganism (L. Wu, personal communication). Recently, a random nonamer oligonucleotide microarray was developed and evaluated for obtaining ﬁngerprinting proﬁles among closely related strains instead of using a gel electrophoresis-based method (Kingsley et al., 2002). A prototype array containing 47 randomly selected nonamer oligonucleotides was constructed and used to differentiate 14 closely related Xanthomonas strains. The REP-PCR was ﬁrst carried out to obtain the ﬁngerprints from different strains, then the ampliﬁed REP-PCR products were hybridized with the nonamer array, and ﬁngerprinting proﬁles for each strain were obtained based on microarray hybridization. The results showed that the microarray-based ﬁngerprinting methods provide clear resolution among all strains examined, including two strains (X. oryzae 43836 and 49072) which could not be resolved using traditional gel electrophoresis of REP-PCR ampliﬁcation methods. This suggests that the microarray hybridization-based approach could provide higher resolution in strain differentiation than the conventional gel electrophoresis-based ﬁngerprinting approach. This approach is attractive because a universal nonamer array can be developed to generate ﬁngerprints from any microorganisms.

IX. CONCLUDING REMARKS Microarray is a recently developed functional genomics technology that has powerful applications in a wide array of biological research areas, including the medical sciences, agriculture, biotechnology and environmental studies. Since many universities, research institutions and industries have established microarray-based core facilities and services, microarrays have become a readily accessible, widely used technology for investigating biological systems. As the

MICROARRAY TECHNOLOGY

259

technologies for array instrumentation are relatively mature, major trends are emerging in such issues as novel array platforms, attachment strategies and substrates, miniaturization with higher density, novel labeling strategies, scanning technologies and automation (Constans, 2003; Stears et al., 2003). Besides the DNA-based array assay, the microarray platform is also being rapidly expanded to include the analysis of other biomolecules such as proteins and carbohydrates (Stears et al., 2003). Along with exploration in microarray technology applications, novel strategies and approaches for experimental controls and design are needed to ensure that microarray hybridization data from different samples are comparable, interpretable and biologically signiﬁcant because of the inherent variability in microarray hybridization signals. Finally, more advanced automatic mathematical and computational tools, such as multivariate analysis, time-series analysis, neural network, artiﬁcial intelligence, and differential equation-based modeling approaches, should be extremely useful for rapid pattern recognition, visualization, data mining, cellular modeling, simulation and prediction. The development and application of microarray-based genomic technology for environmental studies has received a great deal of attention. Because of its highdensity and high-throughput capacity, it is expected that microarray-based genomic technologies will revolutionize the analyses of microbial community structure, function and dynamics. Microarray-based assays have great potential as speciﬁc, sensitive, quantitative, parallel, and high-throughput tools for microbial detection, identiﬁcation and characterization in natural environments. However, more rigorous and systematic assessment and development are needed to realize the full potential of microarrays for microbial ecology studies. Several key issues need to be addressed, including novel experimental designs and strategies for minimizing inherent high hybridization variations to improve microarray-based quantitative accuracy, novel approaches for increasing hybridization sensitivity to detect extremely low biomass in natural environments, novel computational tools for microarray data extraction and interpretation, and broad integration and application of microarray technologies with environmental studies to address ecological and environmental questions and hypotheses.

REFERENCES Adams, M. W. W., Holden, J. F., Menon, A. L., Schut, G. J., Grunden, A. M., Hou, C., Hutchins, A. M., Jenney, F. E. Jr., Kim, C., Ma, K., Pan, G., Roy, R., Sapra, R., Story, S. V., and Verhagen, M. F. (2001). Key role for sulfur in peptide metabolism and in regulation of three hydrogenases in the hyperthermophilic archaeon Pyrococcus furiosus. J. Bacteriol. 183, 716–724. Adler, K., Broadbent, J., Garlick, R., Joseph, R., Khimani, A., Mikulskis, A., Rapiejko, P., and Killian, J. (2000). MICROMAXe: A highly sensitive system for differential gene expression on microarrays. “In Microarray Biochip Technology” (M. Schena, Ed.), pp. 221–230. Eaton Publishing, Natick, MA.

260

J. ZHOU AND D. K. THOMPSON

Afanassiev, V., Hanemann, V., and Wolﬂ, S. (2000). Preparation of DNA and protein micro arrays on glass slides coated with an agarose ﬁlm. Nucl. Acids Res. 28, E66. Anderson, A., Nordan, H., Cain, R., Parrish, G., and Duggan, D. (1956). Studies on a radioresistant micrococcus, I. Isolation, morphology, cultural characteristics, and resistance to gamma radiation. Food Technol. 10, 575–578. Augenlicht, L., Taylor, J., Anderson, L., and Lipkin, M. (1991). Patterns of gene expression that characterize the colonic mucosa in patients at genetic risk for colonic cancer. Proc. Nat. Acad. Sci. 88, 3286–3289. Augenlicht, L., Wahrman, M., Halsey, H., Anderson, L., Taylor, J., and Lipkin, M. (1987). Expression of cloned sequences in biopsies of human colonic tissue and in colonic–carcinoma cells induced to differentiate invitro. Cancer Res. 47, 6017–6021. Bains, W., and Smith, G. (1988). A novel method for nucleic acid sequence determination. J. Theoret. Biol. 135, 303–307. Baldi, P., and Long, A. D. (2001). A Bayesian framework for the analysis of microarray expression data: regularized t-test and statistical inferences of gene changes. Bioinformatics 17, 509–519. Bartosiewicz, M., Trounstine, M., Barker, D., Johnston, R., and Buckpitt, A. (2000). Development of a toxicological gene array and quantitative assessment of this technology. Arch. Biochem. Biophys. 376, 66 –73. Basarsky, T., Verdnik, D., Zhai, J. Y., and Wellis, D. (2000). Overview of a microarray scanner: Design essentials for an integrated acquisition and analysis platform. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 265 –284. Eaton Publishing, Natick, MA. Bassett, D. Jr., Eisen, M., and Boguski, M. (1999). Gene expression informatics — it’s all in your mine. Nat. Genet. Sup. 21, 51–55. Battaglia, C., Salani, G., Consolandi, C., Bernardi, L. R., and De, Bellis, G. (2000). Analysis of DNA microarrays by non-destructive ﬂuorescent staining using SYBR green II. Biotechniques 29, 78–81. Battista, J. R. (1997). Against all odds: the survival strategies of Deinococcus radiodurans. Annu. Rev. Microbiol. 51, 203–224. Battista, J. R., Earl, A. M., and Park, M.-J. (1999). Why is Deinococcus radiodurans so resistant to ionizing radiation? Trends Microbiol. 7, 362–365. Bavykin, S. G., Akowski, J. P., Zakhariev, V. M., Barsky, V. E., Perov, A. N., and Mirzabekov, A. D. (2001). Portable system for microbial sample preparation and oligonucleotide microarray analysis. Appl. Environ. Microbiol. 67, 922 –928. Beattie, K. L., Eggers, M. D., Shumaker, J. M., Hogan, M. E., Varma, R. S., Lamture, J. B., Hollis, M. A., Ehrlich, D. J., and Rathman, D. (1992). Genosensor technology. Clin. Chem. 39, 719– 722. Beattie, W. G., Meng, L., Turner, S., Varma, R. S., Dao, D. D., and Beattie, K. L. (1995). Hybridization of DNA targets to glass-tethered oligonucleotide probes. Mol. Biotechnol. 4, 213 –225. Behr, M. A., Wilson, M. A., Gill, W. P., Salamon, H., Schoolnik, G. K., Rane, S., and Small, P. M. (1999). Comparative genomics of BCG vaccines by whole-genome DNA microarray. Science 284, 1520–1523. Beier, M., and Hoheisel, J. (1999). Versatile derivatisation of solid support media for covalent bonding on DNA-microchips. Nucl. Acids Res. 27, 1970–1977. Belfort, M., and Weiner, A. (1997). Another bridge between kingdoms: tRNA splicing in Archaea and Eukaryotes. Cell 89, 1003– 1006. Beliaev, A. S., Thompson, D. K., Fields, M. W., Wu, L., Lies, D. P., Nealson, K. H., and Zhou, J. (2002a). Microarray transcription proﬁling of a Shewanella oneidensis etrA mutant. J. Bacteriol. 184, 4612–4616. Beliaev, A. S., Thompson, D. K., Khare, T., Lim, H., Brandt, C. C., Li, G., Murray, A. E., Heidelberg, J. F., Giometti, C. S., Yates, III, J., Nealson, K. H., Tiedje, T. M., and Zhou, J. (2002b). Gene and protein expression proﬁles of Shewanella oneidensis during anaerobic growth with different electron acceptors. OMICS 6, 39– 60.

MICROARRAY TECHNOLOGY

261

Ben-Dor, A., Shamir, R., and Yakhini, Z. (1999). Clustering gene expression patterns. J. Comput. Biol. 6, 281–297. Broude, N. E., Woodward, K., Cavallo, R., Cantor, C. R., and Englert, D. (2001). DNA microarrays with stem-loop DNA probes: preparation and applications. Nucl. Acids Res. 29, E92. Bult, C. J., White, O., Olsen, G. J., Zhou, L. X., Fleischmann, R. D., Sutton, G. G., Blake, J. A., FitzGerald, M., Clayton, R. A., Gocayne, J. D., Kerlavage, A. R., Dougherty, B. A., Tomb, J. F., Adams, M. D., Reich, C. I., Overbeek, R., Kirkness, E. F., Weinstock, K. G., Merrick, J. M., Glodek, A., Scott, J. L., Geoghagen, N. S. M., Weidman, J. F., Fuhrmann, J. L., Nguyen, D., Utterback, T. R., Kelley, J. M., Peterson, J. D., Sadow, P. W., Hanna, M. C., Cotton, M. D., Roberts, K. M., Hurst, M. A., Kaine, B. P., Borodovsky, M., Klenk, H. P., Fraser, C. M., Smith, H. O., Woese, C. R., and Venter, J. C. (1996). Complete genome sequence of the methanogenic archaeon, Methanococcus jannaschii. Science 273, 1058–1073. Chee, M., Yang, R., Hubbell, E., Berno, A., Huange, X. C., Stern, D., Winkler, J., Lockhart, D. J., Morris, M. S., and Fodor, S. P. A. (1996). Accessing genetic information with high-density DNA arrays. Science 274, 610 –614. Chen, Y., Dougherty, E. R., and Bittner, M. L. (1997). Ratio-based decisions and the quantitative analysis of cDNA microarray images. J. Biomed. Optics. 2, 364–374. Cho, J. C., and Tiedje, J. M. (2001). Bacterial species determination from DNA –DNA hybridization by using genome fragments and DNA microarrays. Appl. Environ. Microbiol. 67, 3677–3682. Chrisey, L. A., Lee, G. U., and O’Ferrall, C. E. (1996). Covalent attachment of synthetic DNA to self-assembled monolayer ﬁlms. Nucl. Acids Res. 24, 3031–3039. Clewell, D. R. (1993). Sex pheromones and the plasmid encoded mating response in Enterococcus faecalis. In “Bacterial conjugation” (D. B. Clewell, Ed.), pp. 349–367. Plenum Press, New York, NY. Constans, A. (2003). Microarray instrumentation. The Scientist 17, 37–38. Daly, M. J., Ling, O., and Minton, K. W. (1994). Interplasmidic recombination following irradiation of the radioresistant bacterium Deinococcus radiodurans. J. Bacteriol. 176, 7506–7515. Daly, M. J., and Minton, K. W. (1995). Interchromosomal recombination in the extremely radioresistant bacterium Deinococcus radiodurans. J. Bacteriol. 177, 5495–5505. Dassy, B., and Fournier, J. M. (1996). Respiratory activity is essential for post-exponential-phase production of type 5 capsular polysaccharide by Staphylococcus aureus. Infect. Immun. 64, 2408–2414. de Saizieu, A., Certa, U., Warrington, J., Gray, C., Keck, W., and Mous, J. (1998). Bacterial transcript imaging by hybridization of total RNA to oligonucleotide arrays. Nat. Biotechnol. 16, 45–48. DeRisi, J. L., Iyer, V. R., and Brown, P. O. (1997). Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 278, 680 –686. Diehl, F., Grahlmann, S., Beier, M., and Hoheisel, J. D. (2001). Manufacturing DNA microarrays of high spot homogeneity and reduced background signal. Nucleic Acids Res. 29, E38. Dong, S., Wang, E., Hsie, L., Cao, Y., Chen, X., and Gingeras, T. R. (2001). Flexible use of highdensity oligonucleotide arrays for single-nucleotide polymorphism discovery and validation. Genome Res. 11, 1418–1424. Dopazo, J., and Carazo, J. M. (1997). Phylogenetic reconstruction using an unsupervised growing neural network that adopts the topology of a phylogenetic tree. J. Molecul. Evol. 44, 226 –233. Dopazo, J., Zanders, E., Dragoni, I., Amphlett, G., and Falciani, F. (2001). Methods and approaches in the analysis of gene expression data. J. Immunol. Meth. 250, 93– 112. Drmanac, R., Labat, I., Brukner, I., and Crkvenjakov, R. (1989). Sequencing of megabase-plus DNA by hybridization: Theory of the method. Genomics 4, 114–128. Drobyshev, A., Mologina, N., Shik, V., Pobedimskaya, D., Yershov, G., and Mirzabekov, A. (1997). Sequence analysis by hybridization with oligonucleotide microchip: identiﬁcation of betathalassemia mutations. Gene 188, 45–52.

262

J. ZHOU AND D. K. THOMPSON

Drobyshev, A. L., Zasedatelev, A. S., Yershov, G. M., and Mirzabekov, A. D. (1999). Massive parallel analysis of DNA-Hoechst 33258 binding speciﬁcity with a generic oligodeoxyribonucleotide microchip. Nucl. Acids Res. 27, 4100–4105. Dudoit, S. Y., Yang, H., Gallow, M. J., and Speed, T. P. (2001). Statistical methods for identifying genes with differential expression in replicated cDNA microarray experiments. Statist Sincia 12, 111–139. Duggan, D. J., Bittner, M., Chen, Y., Meltzer, P., and Trent, J. M. (1999). Expression proﬁling using cDNA microarrays. Nat. Genet. 21, 10–14. Dunman, P. M., Murphy, E., Haney, S., Palacios, D., Tucker-Kellogg, G., Wu, S., Brown, E. L., Zagursky, R. J., Shlaes, D., and Projan, S. J. (2001). Transcription proﬁling-based identiﬁcation of Staphylococcus aureus genes regulated by the agr and/or sarA loci. J. Bacteriol. 183, 7341– 7353. Eckmann, L., Smith, J. R., Housley, M. P., Dwinell, M. B., and Kagnoff, M. F. (2000). Analysis of high density cDNA arrays of altered gene expression in human intestinal epithelial cells in response to infection with the invasive enteric bacteria Salmonella. J. Biol. Chem. 275, 14084– 14094. Eggers, M. D., Hogan, M. E, Reich, R. K., Lamture, J., Ehrlich, D., Hollis, M., Kosicki, B., Powdrill, T., Beattie, K., Smith, S., Varma, R., Gangadharan, R., Mallik, A., Burke, B., and Wallace, D. (1994). A microchip for quantitative detection of molecular utilizing luminescent and radioisotope reporter groups. Biotechniques. 14, 516–525. Eisen, M., and Brown, P. (1999). DNA microarrays for analysis of gene expression. Meth. Enzymol. 303, 179 –205. Englert, D. (2000). Production of microarrays on porous substrates using noncontact piezoelectric dispensing. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 231–246. Eaton Publishing, Natick, MA. Evertsz, E., Starink, P., Gupta, R., and Watson, D. (2000). Technology and applications of gene expression microarrays. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 149– 166. Eaton Publishing, Natick, MA. Fiala, G., and Stetter, K. O. (1986). Pyrococcus furiosus sp. nov. represents a novel genus of marine heterotrophic archaebacteria growing optimally at 1008C. Arch. Microbiol. 145, 56–61. Fitz-Gibbon, S. T., Ladner, H., Kim, U.-J., Stetter, K. O., Simon, M. I., and Miller, J. H. (2002). Genome sequence of the hyperthermophilic crenarchaeon Pyrobaculum aerophilum. Proc. Natl. Acad. Sci. USA 99, 984 –989. Fodor, S. P. A., Read, J., Pirrung, M., Stryer, L., Lu, A., and Solas, D. (1991). Light-directed, spatially addressable parallel chemical synthesis. Science 251, 767–773. Fotin, A., Drobyshev, A., Proudnikov, D., Perov, A., and Mirzabekov, A. (1998). Parallel thermodynamic analysis of duplexes on oligodeoxyribonucleotide microchips. Nucl. Acids Res. 26, 1515–1521. Fredrickson, J. K., Kostandarithes, H. M., Li, S. W., Plymale, A. E., and Daly, M. J. (2000). Reduction of Fe(III), Cr(VI), U(VI), and Tc(VII) by Deinococcus radiodurans R1. Appl. Environ. Microbiol. 66, 2006–2011. Galagan, J. E., Nusbaum, C., Roy, A., Endrizzi, M. G., Macdonald, P., FitzHugh, W., Calvo, S., Engels, R., Smirnov, S., Atnoor, D., Brown, A., Allen, N., Naylor, J., Stange-Thomann, N., DeArellano, K., Johnson, R., Linton, L., McEwan, P., McKernan, K., Talamas, J., Tirrell, A., Ye, W. J., Zimmer, A., Barber, R. D, Cann, I., Graham, D. E., Grahame, D. A., Guss, A. M., Hedderich, R., Ingram-Smith, C., Kuettner, H. C., Krzycki, J. A., Leigh, J. A., Li, W. X., Liu, J. F., Mukhopadhyay, B., Reeve, J. N., Smith, K., Springer, T. A., Umayam, L. A., White, O., White, R. H., de, Macario, E. C., Ferry, J. G., Jarrell, K. F., Jing, H., Macario, A. J. L., Paulsen, I., Pritchett, M., Sowers, K. R., Swanson, R. V., Zinder, S. H., Lander, E., Metcalf, W. W., and Birren, B. (2002). The genome of M. acetivorans reveals extensive metabolic and physiological diversity. Genome Res. 12, 532–542.

MICROARRAY TECHNOLOGY

263

Guo, Z., Guilfoyle, R. A., Thiel, A. J., Wang, R., and Smith, L. M. (1994). Direct ﬂuorescence analysis of genetic polymorphisms by hybridization with oligonucleotide arrays on glass supports. Nucl. Acids Res. 22, 5456–5465. Guschin, D. Y., Mobarry, B. K., Proudnikov, D., Stahl, D. A., Rittmann, B. E., and Mirzabekov, A. (1997a). Oligonuclotide microchips as genosensors for determinative and environmental studies in microbiology. Appl. Environ. Microbiol. 63, 2397–2402. Guschin, D., Yershov, G., Zaslavsky, A., Gemmell, A., Shick, V., Proudnikov, D., Arenkov, P., and Mirzabekov, A. (1997b). Manual manufacturing of oligonucleotide, DNA, and protein microchips. Anal. Biochem. 250, 203 –11. Hacia, J. G. (1999). Resequencing and mutational analysis using oligonucleotide microarrays. Nat. Genet. 21, 42–47. Hegde, P., Qi, R., Abernathy, K., Gay, C., Dharap, S., Gaspard, R., Hughes, J. E., Snesrud, E., Lee, N., and Quackenbush, J. (2000). A concise guide to cDNA microarray analysis. Biotechniques 29, 548 –560. Heidelberg, J. F., Paulsen, I. T., Nelson, K. E., Gaidos, E. J., Nelson, W. C., Read, T. D., Eisen, J. A., Seshadri, R., Ward, N., Methe, B., Clayton, R. A., Meyer, T., Tsapin, A., Scott, J., Beanan, M., Brinkac, L., Daugherty, S., DeBoy, R. T., Dodson, R. J., Durkin, A. S., Haft, D. H., Kolonay, J. F., Madupu, R., Peterson, J. D., Umayam, L. A., White, O., Wolf, A. M., Vamathevan, J., Weidman, J., Impraim, M., Lee, K., Berry, K., Lee, C., Mueller, J., Khouri, H., Gill, J., Utterback, T. R., McDonald, L. A., Feldblyum, T. V., Smith, H. O., Venter, J. C., Nealson, K. H., and Fraser, C. M. (2002). Genome sequence of the dissimilatory metal ion-reducing bacterium Shewanella oneidensis. Nature Biotechnol. 20, 1118–1123. Heller, H. M., Schena, M., Chai, A., Shalon, D., Bedilion, T., and Gilmore, J. (1997). Discovery and analysis of inﬂammatory disease-related genes using cDNA microarrays. Proc. Natl. Acad. Sci. USA 94, 2150– 2155. Herrero, J., Valencia, A., and Dopazo, J. (2001). A hierarchical unsupervised growing neural network for clustering gene expression patterns. Bioinformatics 17, 126–136. Herwig, R., Poustka, A. J., Muller, C., Bull, C., Lehrach, H., and O’Brien, J. (1999). Large-scale clustering of cDNA-ﬁngerprinting data. Genome Res. 9, 1093–1105. Heyer, L. J., Kruglyak, S., and Yooseph, S. (1999). Exploring expression data: identiﬁcation and analysis of coexpressed genes. Genome Res. 9, 1106–1115. Hilsenbeck, S. G., Friedrichs, W. E., Schiff, R., O’Connell, P., Hansen, R. K., Osborne, C. K., and Fuqua, S. A. (1999). Statistical analysis of array expression data as applied to the problem of tamoxifen resistance. J. Natl. Cancer Inst. 91, 453–459. Hoheisel, J. (1997). Oligomer-chip technology. Trends Biotech. 15, 465–469. Hughes, T. R., Marton, M. J., Jones, A. R., Roberts, C. J., Stoughton, R., Armour, C. D., Bennett, H. A., Coffey, E., Dai, H., He, Y. D., Kidd, M. J., King, A. M., Meyer, M. R., Slade, D., Lum, P. Y., Stepaniants, S. B., Shoemaker, D. D., Gachotte, D., Chakraburtty, K., Simon, J., Bard, M., and Friend, S. H. (2000). Functional discovery via a compendium of expression proﬁles. Cell 102, 109 –126. Kalocsai, P., and Shams, S. (2001). Use of bioinformatics in arrays. Meth. Mol. Biol. 170, 223 –236. Kane, M. D., Jatkoe, T. A., Stumpf, C. R., Lu, J., Thomas, J. D., and Madore, S. J. (2000). Assessment of the sensitivity and speciﬁcity of oligonucleotide (50mer) microarrays. Nucl. Acids Res. 28, 4552–4557. Kaneko, T., Nakamura, Y., Wolk, C. P., Kuritz, T., Sasamoto, S., Watanabe, A., Iriguchi, M., Ishikawa, A., Kawashima, K., Kimura, T., Kishida, Y., Kohara, M., Matsumoto, M., Matsuno, A., Muraki, A., Nakazaki, N., Shimpo, S., Sugimoto, M., Takazawa, M., Yamada, M., Yasuda, M., and Tabata, S. (2001). Complete genomic sequence of the ﬁlamentous nitrogen-ﬁxing cyanobacterium Anabaena sp. strain PCC 7120. DNA Res. 8, 205–213.

264

J. ZHOU AND D. K. THOMPSON

Kaneko, T., Sato, S., Kotani, H., Tanaka, A., Asamizu, E., Nakamura, Y., Miyajima, N., Hirosawa, M., Sugiura, M., Sasamoto, S., Kimura, T., Hosouchi, T., Matsuno, A., Muraki, A., Nakazaki, N., Naruo, K., Okumura, S., Shimpo, S., Takeuchi, C., Wada, T., Watanabe, A., Yamada, M., Yasuda, M., and Tabata, S. (1996). Sequence analysis of the genome of the unicellular cyanobacterium Synechocystis sp. strain PCC6803. II. Sequence determination of the entire genome and assignment of potential protein-coding regions. DNA Res. 3, 109 –136. Kawarabayasi, Y., Hino, Y., Horikawa, H., Yamazaki, S., Haikawa, Y., Jin-no, K., Takahashi, M., Sekine, M., Baba, S., Ankai, A., Kosugi, H., Hosoyama, A., Fukui, S., Nagai, Y., Nishijima, K., Nakazawa, H., Takamiya, M., Masuda, S., Funahashi, T., Tanaka, T., Kudoh, Y., Yamazaki, J., Kushida, N., Oguchi, A., Aoki, K., Kubota, K., Nakamura, Y., Nomura, N., Sako, Y., and Kikuchi, H. (1999). Complete genome sequence of an aerobic hyper-thermophilic crenarchaeon, Aeropyrum pernix K1. DNA Res. 6, 83–101. Kawarabayasi, Y., Hino, Y., Horikawa, H., Jin-no, K., Takahashi, M., Sekine, M., Baba, S., Ankai, A., Kosugi, H., Hosoyama, A., Fukui, S., Nagai, Y., Nishijima, K., Otsuka, R., Nakazawa, H., Takamiya, M., Kato, Y., Yoshizawa, T., Tanaka, T., Kudoh, Y., Yamazaki, J., Kushida, N., Oguchi, A., Aoki, K., Masuda, S., Yanagii, M., Nishimura, M., Yamagishi, A., Oshima, T., and Kikuchi, H. (2001). Complete genome sequence of an aerobic thermoacidophilic crenarchaeon, Sulfolobus tokodaii strain7. DNA Res. 8, 123 –140. Kawarabayasi, Y., Sawada, M., Horikawa, H., Haikawa, Y., Hino, Y., Yamamoto, S., Sekine, M., Baba, S., Kosugi, H., Hosoyama, A., Nagai, Y., Sakai, M., Ogura, K., Otsuka, R., Nakazawa, H., Takamiya, M., Ohfuku, Y., Funahashi, T., Tanaka, T., Kudoh, Y., Yamazaki, J., Kushida, N., Oguchi, A., Aoki, K., and Kikuchi, H. (1998). Complete sequence and gene organization of the genome of a hyper-thermophilic archaebacterium, Pyrococcus horikoshii. DNA Res. 5, 55 –76. Kawashima, T., Amano, N., Koike, H., Makino, S., Higuchi, S., Kawashima-Ohya, Y., Watanabe, K., Yamazaki, M., Kanehori, K., Kawamoto, T., Nunoshiba, T., Yamamoto, Y., Aramaki, H., Makino, K., and Suzuki, M. (2000). Archaeal adaptation to higher temperatures revealed by genomic sequence of Thermoplasma volcanium. Proc. Natl. Acad. Sci. USA 97, 14257–14262. Kerr, M. K., and Churchill, G. A. (2001a). Statistical design and the analysis of gene expression microarray data. Genet. Res. 77, 123–128. Kerr, M. K., and Churchill, G. A. (2001b). Experimental design for gene expression microarrays. Biostatistics 2, 183– 201. Khodursky, A. B., Peter, B. J., Cozzarelli, N. R., Botstein, D., Brown, P. O., and Yanofsky, C. (2000). DNA microarray analysis of gene expression in response to physiological and genetic changes that affect tryptophan metabolism in Escherichia coli. Proc. Natl. Acad. Sci. USA 97, 12170– 12175. Khrapko, K. R., Lysov, Y. P., Khorlyn, A. A., Ivanov, I. B., Yershov, G. M., Vasilenko, S. K., Vlorentiev, V. L., and Mirzabekov, A. (1989). An oligonucleotide hybridization approach to DNA sequencing. FEBS Lett. 256, 118 –122. Kingsley, M. T., Straub, T. M., Call, D. R., Daly, D. S., Wunschel, S. C., and Chandler, D. P. (2002). Fingerprinting closely related Xanthomonas pathovars with random nanamer oligonucleotide microarrays. Appl. Environ. Microbiol. 68, 6361–6370. Kleman-Leyer, K., Armbruster, D. W., and Daniels, C. J. (1997). Properties of H. volcanii tRNA intron endonuclease reveal a relationship between the archaeal and eucaryal tRNA intron processing systems. Cell 89, 839–847. Klenk, H.-P., Clayton, R. A., Tomb, J.-F., White, O., Nelson, K. E., Ketchum, K. A., Dodson, R. J., Gwinn, M., Hickey, E. K., Peterson, J. D., Richardson, D. L., Kerlavage, A. R., Graham, D. E., Kyrpides, N. C., Fleischmann, R. D., Quackenbush, J., Lee, N. H., Sutton, G. G., Gill, S., Kirkness, E. F., Dougherty, B. A., McKenney, K., Adams, M. D., Loftus, B., Peterson, S., Reich, C. I., McNeil, L. K., Badger, J. H., Glodek, A., Zhou, L. X., Overbeek, R., Gocayne, J. D., Weidman, J. F., McDonald, L., Utterback, T., Cotton, M. D., Spriggs, T., Artiach, P., Kaine, B. P., Sykes, S. M., Sadow, P. W., Dandrea, K. P., Bowman, C., Fujii, C., Garland, S. A., Mason, T. M.,

MICROARRAY TECHNOLOGY

265

Olsen, G. J., Fraser, C. M., Smith, H. O., Woese, C. R., and Venter, J. C. (1997). The complete genome sequence of the hyperthermophilic, sulphate-reducing archaeon Archaeoglobus fulgidus. Nature 394, 364–370. Knight, J. (2001). When the chips are down. Nature 410, 860–861. Lange, C. C., Wackett, L. P., Minton, K. W., and Daly, M. J. (1998). Engineering a recombinant Deinococcus radiodurans for organopollutant degradation in radioactive mixed waste environments. Nat. Biotechnol. 16, 929–933. Lamture, J. B., Beattie, K. L., Burke, B. E., Eggers, M. D., Ehrlich, D. J., Fowler, R., Hollis, M. A., Kosicki, B. B., Reich, R. K., and Smith, S. R. (1994). Direct detection of nucleic acid hybridization on the surface of a charge coupled device. Nucl. Acids Res. 22, 2121–2125. Lipshultz, R. J., Fodor, S. P. A., Gingeras, T. R., and Lockhart, D. J. (1999). High density synthetic oligonucleotide arrays. Nat. Genet. 21, 20–24. Liu, Y., Zhou, J., Omelchenko, M., Beliaev, A., Venkateswaran, A., Stair, J., Wu, L., Thompson, D. K., Xu, D., Rogozin, I. B., Gaidamakova, E. K., Zhai, M., Makarova, K. S., Koonin, E. V., and Daly, M. J. (2003). Transcriptome dynamics of Deinococcus radiodurans recovering from ionizing radiation. Proc. Natl. Acad. Sci. USA 100, 4191–4196. Lockhart, D. J., Dong, H., Byrne, M. C., Follettie, M. T., Gallo, M. V., Chee, M. S., Mittmann, M., Wang, C., Kobayashi, M., Horton, H., and Brown, E. L. (1996). Expression monitoring by hybridization to high-density oligonucleotide arrays. Nat. Biotechnol. 14, 1675–1680. Long, A. D., Mangalam, H. J., Chan, B. Y., Tolleri, L., Hatﬁeld, G. W., and Baldi, P. (2001). Improved statistical inference from DNA microarray data using analysis of variance and a Bayesian statistical framework. Analysis of global gene expression in Escherichia coli K12. J. Biol. Chem. 276, 19937– 19944. Lovley, D. (1991). Dissimilatory Fe(III) and Mn(IV) reduction. Microbiol. Rev. 55, 259–287. Loy, A., Lahner, A., Lee, N., Adamczyk, J., Meier, H., Ernst, J., Schleifer, K. H., and Wagner, M. (2002). Oligonucleotide microarray for 16S rRNA gene-based detection of all recognized lineages of sulfate-reducing prokaryotes in the environment. Appl. Environ. Microbiol. 68, 5064–5081. Mace, M. L., Montagu, J., Rose, S. D., and McGuinnes, G. (2000). Novel microarray printing and detection technologies. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 39–64. Eaton Publishing, Natick, MA. Makarova, K. S., Aravind, L., Wolf, Y. I., Tatusov, R. L., Minton, K. W., Koonin, E. V., and Daly, M. J. (2001). Genome of the extremely radiation-resistant bacterium Deinococcus radiodurans viewed from the perspective of comparative genomics. Microbiol. Mol. Biol. Rev. 65, 44–79. Martinsky, T., and Haje, P. (2000). Microarray tools, kits, reagents, and services. In “Microarray Biochip Technology” (M . Schena, Ed.), pp. 201–220. Eaton Publishing, Natick, MA. McGall, G. H., and Fidanza, J. A. (2001). Photolithographic synthesis of high-density oligonucleotide arrays. In “DNA Arrays — Methods and Protocols” (J. B. Rampal, Ed.), pp. 71–101. Humana Press, Totowa, NJ. Minton, K. W. (1996). Repair of ionizing-radiation damage in the radiation resistant bacterium Deinococcus radiodurans. Mutat. Res. 363, 1 –7. Morrison, N., Rattray, M., Oliver, S. G., Hayes, A., Zhang, N., Brutsche, M., Penkett, C., Lockey, J., Rao, S., Hays, I., Jupp, R., and Brass, A. (1999). Robust normalization of microarray data over multiple experiments. Nature Genetics Microarray Meeting: Technology, Application and Analysis. Moser, D., and Nealson, K. H. (1996). Growth of the facultative anaerobe Shewanella putrefaciens by elemental sulfur reduction. Appl. Environ. Microbiol. 62, 2100–2105. Murray, A. E., Lies, D., Li, G., Nealson, K., Zhou, J., and Tiedje, J. M. (2001). DNA/DNA hybridization to microarrays reveals gene-speciﬁc differences between closely related microbial genomes. Proc. Natl. Acad. Sci. USA 98, 9853–9858.

266

J. ZHOU AND D. K. THOMPSON

Nadon, R., Shi, P., Skandalis, A., Woody, E., Hubschle, H., Susko, E., Rghei, N., and Ramm, P. (2001). Statistical inference methods for gene expression arrays. In “Proceedings of SPIE, Microarrays: Optical Technologies and Informatics” (Bittner, Ed.), V4266, pp. 46 – 55. International Society for Optical Engineering, Bellingham, WA. Nadon, R., and Shoemaker, J. (2002). Statistical issues with microarrays: processing and analysis. Trends Genet. 18, 265–271. Nelson, K. E., Clayton, R. A., Gill, S. R., Gwinn, M. L., Dodson, R. J., Haft, D. H., Hickey, E. K., Peterson, L. D., Nelson, W. C., Ketchum, K. A., McDonald, L., Utterback, T. R., Malek, J. A., Linher, K. D., Garrett, M. M., Stewart, A. M., Cotton, M. D., Pratt, M. S., Phillips, C. A., Richardson, D., Heidelberg, J., Sutton, G. G., Fleischmann, R. D., Eisen, J. A., White, O., Salzberg, S. L., Smith, H. O., Venter, J. C., and Fraser, C. M. (1999). Evidence for lateral gene transfer between Archaea and Bacteria from genome sequence of Thermotoga maritima. Nature 399, 323 –329. Nealson, K. H., and Saffarini, D. A. (1994). Iron and manganese in anaerobic respiration: environmental signiﬁcance, physiology, and regulation. Ann. Rev. Microbiol. 48, 311–343. Ng, W. V., Kennedy, S. P., Mahairas, G. G., Berquist, B., Pan, M., Shukla, H. D., Lasky, S. R., Baliga, N. S., Thorsson, V., Sbrogna, J., Swartzell, S., Weir, D., Hall, J., Dahl, T. A., Welti, R., Goo, Y. A., Leithauser, B., Keller, K., Cruz, R., Danson, M. J., Hough, D. W., Maddocks, D. G., Jablonski, P. E., Krebs, M. P., Angevine, C. M., Dale, H., Isenbarger, T. A., Peck, R. F., Pohlschroder, M., Spudich, J. L., Jung, K. H., Alam, M., Freitas, T., Hou, S. B., Daniels, C. J., Dennis, P. P., Omer, A. D., Ebhardt, H., Lowe, T. M., Liang, R., Riley, M., Hood, L., and DasSarma, S. (2000). Genome sequence of Halobacterium species NRC-1. Proc. Natl. Acad. Sci. USA 97, 12176–12181. Nuwaysir, E. F., Huang, W., Albert, T. J., Singh, J., Nuwaysir, K., Pitas, A., Richmond, T., Gorski, T., Berg, J. P., Ballin, J., McCormick, M., Norton, J., Pollock, T., Sumwalt, T., Butcher, L., Porter, D., Molla, M., Hall, C., Blattner, F., Sussman, M. R., Wallace, R. L., Cerrina, F., and Green, R. D. (2002). Gene expression analysis using oligonucleotide arrays produced by maskless photolithography. Genome Res. 12, 1749–1755. Olsen, G. J., and Woese, C. R. (1997). Archaeal genomics: an overview. Cell 89, 991–994. Pease, A. C., Solas, D., Sullivan, E. J., Cronin, M. T., Holmes, C. P., and Fodor, S. P. A. (1994). Light-generated oligonucleotide arrays for rapid DNA sequence analysis. Proc. Natl. Acad. Sci. USA 91, 5022–5026. Petrov, A., Shah, S., Draghici, S., and Shams, S. (2002). Microarray image processing and quality control. In “DNA Array Image Analysis — Nuts & Bolts” (S. Shah and G. Kamberova, Eds.), pp. 99– 130. DNA Press, LLC, Eagleville, PA. Pinkel, D., Segraves, R., Sudar, D., Clark, S., Poole, I., Kowbel, D., Collins, C., Kuo, W. L., Chen, C., Zhai, Y., Dairkee, S. H., Ljung, B. M., Gray, J. W., and Albertson, D. G. (1998). High resolution analysis of DNA copy number variation using comparative genomic hybridization to microarrays. Nat. Genet. 20, 207 –11. Pollack, R. J., Perou, C., Alizadeh, A., Eisen, M., Pergamenschikov, A., Williams, C., Jeffrey, S., Botstein, D., and Brown, P. (1999). Genome-wide analysis of DNA copy-number changes using cDNA microarrays. Nat. Am. 23, 41–46. Ramsay, G. (1998). DNA chips: state-of-the art. Nat. Biotechnol. 1, 640– 644. Rehman, F. N., Audeh, M., Abrams, E. S., Hammond, P. W., Kenney, M., and Boles, T. C. (1999). Immobilization of acrylamide-modiﬁed oligonucleotides by co-polymerization. Nucl. Acids Res 27, 649 –655. Richardson, D. J. (2000). Bacterial respiration: a ﬂexible process for a changing environment. Microbiol. 146, 551–571. Robb, F. T., Maeder, D. L., Brown, J. R., DiRuggiero, J., Stump, M. D., Yeh, R. K., Weiss, R. B., and Dunn, D. M. (2001). Genomic sequence of hyperthermophile Pyrococcus furiosus: implications for physiology and enzymology. Meth. Enzymol. 330, 134– 157.

MICROARRAY TECHNOLOGY

267

Rogers, Y., Jiang-Baucome, P., Huange, Z., Bogdanov, V., Anderson, S., and Boyce-Jacino, M. T. (1999). Immobilization of oligonucleotides onto a glass support via disulﬁde bonds: A method for preparation of DNA microarrays. Ann. Biochem. 266, 23 –30. Rose, D. (2000). Microﬂuidic technologies and instrumentation for printing DNA microarrays. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 19–38. Eaton Publishing, Natick, MA. Rudi, K., Skulberg, O. M., Skulberg, R., and Jakobsen, K. S. (2000). Application of sequence-speciﬁc labeled 16S rRNA gene oligonucleotide probes for genetic proﬁling of cyanbacterial abundance and diversity by array hybridization. Appl. Environ. Microbiol. 66, 4004–4011. Salama, N., Guillemin, K., McDaniel, T. K., Sherlock, G., Tompkins, L., and Falkow, S. (2000). A whole-genome microarray reveals genetic diversity among Helicobacter pylori strains. Proc. Natl. Acad. Sci. USA 97, 14668–14673. Salunga, R. C., Guo, H., Luo, L., Bittner, A., Joy, K. D., Chambers, J. R., Wan, J. S., Jackson, M. R., and Erlander, M. G. (1999). Gene expression analysis via cDNA microarrays of laser capture microdissected cells from ﬁxed tissue. In “DNA Microarrays: A Practical Approach” (M. Schena, Ed.), pp. 121 –136. Oxford University Press, New York, NY. Schadt, E. E., Li, C., Su, C., and Wong, W. H. (2000). Analyzing high-density oligonucleotide gene expression array data. J. Cell Biochem. 80, 192 –202. Schena, M. (2003). “Microarray analysis.” John Wiley & Sons, New York, NY. Schena, M., and Davis, R. W. (2000). Technology standards for microarray research. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 1 –18. Eaton Publishing, Natick, MA. Schena, M., Heller, R., Theriault, T., Konrad, K., Lachenmeier, E., and Davis, R. W. (1998). Microarrays: biotechnology’s disovery platform for functional genomics. Trends Biotech. 16, 301 –306. Schena, M., Shalon, D., Davis, R. W., and Brown, P. O. (1995). Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science. 270, 467 –470. Schena, M., Shalon, D., Heller, R., Chai, A., Brown, P., and Davis, R. W. (1996). Parallel human genome analysis: microarray-based expression monitoring of 1000 genes. Proc. Natl. Acad. Sci. USA 93, 10614– 10619. Schermer, M. J. (1999). Confocal scanning microscopy in microarray detection. In “DNA Microarrays: A Practical Approach” (M. Schena, Ed.), pp. 17–42. Oxford University Press, New York, NY. Schuchhardt, J., Beule, D., Malik, A., Wolski, E., Eickhoff, H., Lehrach, H., and Herzel, H. (2000). Normalization strategies for cDNA microarrays. Nucl. Acids Res. 28, E47. Schut, G. J., Zhou, J., and Adams, M. W. W. (2001). DNA microarray analysis of the hyperthermophilic archaeon Pyrococcus furiosus: Evidence for a new type of sulfur-reducing enzyme complex. J. Bacteriol. 183, 7027–7036. Shalon, D., Smith, S. J., and Brown, P. O. (1996). A DNA microarray system for analyzing complex DNA samples using two-color ﬂuorescent probe hybridization. Genome Res. 6, 639–645. Shchepinov, M. S., Case-Green, S. C., and Southern, E. M. (1997). Steric factors inﬂuencing hybridisation of nucleic acids to oligonucleotide arrays. Nucl. Acids Res. 25, 1155–1161. Shriver-Lake, L. C. (1998). Silane-modiﬁed surfaces for biomaterial immobilization. In “Immobilized Biomolecules in Analysis” (T. Cass and F. S. Ligler, Eds.), pp. 1– 14. Oxford University Press, New York, NY. Slesarev, A. I., Mezhevaya, K. V., Makarova, K. S., Polushin, N. N., Shcherbinina, O. V., Shakhova, V. V., Belova, G. I., Aravind, L., Natale, D. A., Rogozin, I. B., Tatusov, R. L., Wolf, Y. I., Stetter, K. O., Malykh, A. G., Koonin, E. V., and Kozyavkin, S. A. (2002). The complete genome of hyperthermophile Methanopyrus kandleri AV19 and monophyly of archaeal methanogens. Proc. Natl. Acad. Sci. USA 99, 4644–4649. Small, J., Call, D. R., Brockman, F. J., Straub, T. M., and Chandler, D. P. (2001). Direct detection of 16S rRNA in soil extracts by using oligonucleotide microarrays. Appl. Environ. Microbiol. 67, 4708–4716.

268

J. ZHOU AND D. K. THOMPSON

Smith, D. R., Doucette-Stamm, L. A., Deloughery, C., Lee, H., Dubois, J., Aldredge, T., Bashirzadeh, R., Blakely, D., Cook, R., Gilbert, K., Harrison, D., Hoang, L., Keagle, P., Lumm, W., Pothier, B., Qiu, D., Spadafora, R., Vicaire, R., Wang, Y., Wierzbowski, J., Gibson, R., Jiwani, N., Caruso, A., Bush, D., Reeve, J. N., et al., (1997). Complete genome sequence of Methanobacterium thermoautotrophicum KH: Functional analysis and comparative genomics. J. Bacteriol. 179, 7135–7155. Southern, E. M. (2001). DNA microarrays. History and overview. Meth. Mol. Biol. 170, 1–15. Southern, E. M., Case-Green, S. C., Elder, J. K., Johnsone, M., Mir, K. U., Wang, L., and Williams, J. C. (1994). Arrays of complementary oligonucleotides for analyzing the hybridizaiton behavior of nucleic acids. Nucl. Acids Res. 22, 1368– 1373. Southern, E. M., Maskos, U., and Elder, J. K. (1992). Analyzing and comparing nucleic acid sequences by hybridization to arrays of oligonucleotides: evaluation using experimental models. Genomics 13, 1108–1107. Southern, E., Mir, K., and Shchepinov, M. (1999). Molecular interactions on microarrays. Nat. Genet. 21(Suppl. S), 5 –9. Stears, R. L., Getts, R. C., and Gullans, S. R. (2000). A novel, sensitive detection system for highdensity microarrays using dendrimer technology. Physiol. Genomics 3, 93–99. Stears, B. L., Martinsky, T., and Schena, M. (2003). Trends in microarray analysis. Nat. Med. 9, 140–145. Steel, A. B., Levicky, R. L., Herne, T. M., and Tarlo, M. J. (2000). Immobilization of nucleic acids at solid surfaces: effect of oligonucleotide length on layer assembly. Biophys. J79, 975–981. Stillman, B. A., and Tonkinson, J. L. (2000). FAST slides: a novel surface for microarrays. Biotechniques 29, 630–635. Tamayo, P., Slonim, D., Mesirov, J., Zhu, Q., Kitareewan, S., Dmitrovsky, E., Lander, E. S., and Golub, T. R. (1999). Interpreting patterns of gene expression with self-organizing maps: methods and application to hematopoietic differentiation. Proc. Natl. Acad. Sci. USA 96, 2907–2912. Taniguchi, M., Miura, K., Iwao, H., and Yamanaka, S. (2001). Quantitative assessment of DNA microarrays — comparison with Northern blot analyses. Genomics 71, 34 –39. Tao, H., Bausch, C., Richmond, C., Blattner, F. R., and Conway, T. (1999). Functional genomics: expression analysis of Escherichia coli growing on minimal and rich media. J. Bacteriol. 181, 6425–6440. Taroncher-Oldenburg, G., Griner, E. M., Francis, C. A., and Ward, B. B. (2003). Oligonucleotide microarray for the study of functional gene diversity in the nitrogen cycle in the environment. Appl. Environ. Microbiol. 69, 1159–1171. Tavazoie, S., and Church, G. M. (1998). Quantitative whole-genome analysis of DNA-protein interactions by in vivo methylase protection in E-coli. Nat. Biotech. 16, 566 –571. Thompson, D. K., Beliaev, A. S., Giometti, C. S., Tollaksen, S. L., Khare, T., Lies, D. P., Nealson, K. H., Lim, H., Yates, III, J., Brandt, C. C., Tiedje, J. M., and Zhou, J. (2002). Transcriptional and proteomic analysis of a ferric uptake regulator (Fur) mutant of Shewanella oneidensis: Possible involvement of Fur in energy metabolism, transcriptional regulation, and oxidative stress. Appl. Environ. Microbiol. 68, 881–892. Toronen, P., Kolehmainen, M., Wong, G., and Castren, E. (1999). Analysis of gene expression data using self-organizing maps. FEBS Lett. 451, 142 –146. Tseng, G. C., Oh, M. K., Rohlin, L., Liao, J. C., and Wong, W. H. (2001). Issues in cDNA microarray analysis: quality ﬁltering, channel normalization, models of variations and assessment of gene effects. Nucl. Acids Res. 29, 2549–2557. Urakawa, H., Noble, P. A., El Fantroussi, S., Kelly, J. J., and Stahl, D. A. (2002). Single-base-pair discrimination of terminal mismatches by using oligonucleotide microarrays and neural network analyses. Appl. Environ. Microbiol. 68, 235–244.

MICROARRAY TECHNOLOGY

269

Vasiliskov, A., Timofeev, E., Surzhikov, S., Drobyshev, A., Shick, V., and Mirzabekov, A. (1999). Fabricaiton of microarray of gel-immobilized compounds on a chip by copolymerization. Biotechniques 27, 592 –606. Venkateswaran, A., McFarlan, S. C., Ghostal, D., Minton, K. W., Vasilenko, A., Makarova, K. S., Wackett, L. P., and Daly, M. J. (2000). Physiologic determinants of radiation resistance in Deinococcus radiodurans. Appl. Environ. Microbiol. 66, 2620–2626. Venkateswaran, K., Moser, D., Dollhopf, M., Lies, D. P., Saffarini, D. A., MacGregor, B. J., Ringelberg, D. B., White, D. C., Nishijima, M., and Sano, H. (1999). Polyphasic taxonomy of the genus Shewanella and description of Shewanella oneidensis sp. nov. Int. J. Syst. Bacteriol. 49, 705 –724. Verdnik, D., Handran, S., and Pickett, S. (2002). Key considerations for accurate microarray scanning and image analysis. In “DNA Array Image Analysis — Nuts & Bolts” (S. Shah and G. Kamberova, Eds.). DNA Press, LLC, Eagleville, PA. Voordouw, G. (1998). Reverse sample genome probing of microbial community dynamics. ASM News 64, 627–633. Wang, D. G., Fan, J. B., Siao, C. J., Berno, A., Young, P., Sapolsky, R., Ghandour, G., Perkins, N., Winchester, E., Spencer, J., Kruglyak, L., Stein, L., Hsie, L., Topaloglou, E., Hubbell, E., Robinson, M., Mittmann, M. S., Morris, N., Shen, D., Kilburn, T., Rioux, J., Nusbaum, C., Rozen, S., Hudson, T. J., Lander, E. S., and et al. (1998). Large-scale identiﬁcation, mapping, and genotyping of single-nucleotide polymorphisms in the human genome. Science 280, 1077– 1082. Warrington, J. A., Dee, S., and Trulson, M. (2000). Large-scale genomic analysis using Affymetrix GeneChipw probe arrays. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 119–148. Eaton Publishing, Natick, MA. White, O., Eisen, J. A., Heidelberg, J. F., Hickey, E. K., Peterson, J. D., Dodson, R. J., Haft, D. H., Gwinn, M. L., Nelson, W. C., Richardson, D. L., Moffat, K. S., Qin, H. Y., Jiang, L. X., Pamphile, W., Crosby, M., Shen, M., Vamathevan, J. J., Lam, P., McDonald, L., Utterback, T., Zalewski, C., Makarova, K. S., Aravind, L., Daly, M. J., Minton, K. W., Fleischmann, R. D., Ketchum, K. A., Nelson, K. E., Salzberg, S., Smith, H. O., Venter, J. C., and Fraser, C. M. (1999). Genome sequence of the radioresistant bacterium Deinococcus radiodurans R1. Science 286, 1571–1577. Wilson, K. H., Wilson, W. J., Radosevich, J. L., DeSantis, T. Z., Viswanathan, V. S., Kuczmarski, T. A., and Andersen, G. L. (2002). High-density microarray of small-subunit ribosomal DNA probes. Appl. Environ. Microbiol. 68, 2535– 2541. Wodicka, L., Dong, H., Mittmann, M., Ho, M.-H., and Lockhart, D. J. (1997). Genome-wide expression monitoring in Saccharomyces cerevisiae. Nat. Biotechnol. 15, 1359–1367. Woese, C. R. (1987). Bacterial evolution. Microbiol. Rev. 51, 221 –271. Woese, C. R., Kandler, O., and Wheelis, M. L. (1990). Towards a natural system of organisms: proposal for the domains Archaea, Bacteria, and Eucarya. Proc. Natl. Acad. Sci. USA 87, 4576–4579. Worley, J., Bechtol, K., Penn, S., Roach, D., Hanzel, D., Trounstine, M., and Barker, D. (2000). A systems approach to fabricating and analyzing DNA microarrays. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 65–85. Eaton Publishing, Natick, MA. Wu, L. Y., Thompson, D., Li, G.-S., Hurt, R., Huang, H., Tiedje, J. M., and Zhou, J.-Z. (2001). Development and evaluation of functional gene arrays for detection of selected genes in the environment. Appl. Environ. Microbiol. 67, 5780–5790. Xu, D., Li, G., Wu, L., Zhou, J.-Z., and Xu, Y. (2002). PRIMEGENS: A computer program for robust and efﬁcient design of gene-speciﬁc targets on microarrays. Bioinformatics 18, 1432– 1437. Yang, Y. H., and Speed, T. (2002). Design issues for cDNA microarray experiments. Nat. Rev. Genet. 3, 579–588. Ye, R., Tao, W., Bedzyk, L., Young, T., Chen, M., and Li, L. (2000). Global gene expression proﬁles of Bacillus subtilis grown under anaerobic conditions. J. Bacteriol. 182, 4458–4465.

270

J. ZHOU AND D. K. THOMPSON

Zammatteo, N., Jeanmart, L., Hamels, S., Courtois, S., Louette, P., Hevesi, L., and Remacle, J. (2000). Comparison between different strategies of covalent attachment of DNA to glass surfaces to build DNA microarrays. Anal. Biochem. 280, 143–150. Zhao, X., Nampalli, S., Serino, A. J., and Kumar, S. (2001). Immobilization of oligodeoxyribonucleotides with multiple anchors to microchips. Nucl. Acids Res. 29, 955–959. Zhou, J.-Z. (2003). Microarrays for bacterial detection and microbial community analysis. Curr. Opin. Microbiol. 6, 288– 294. Zhou, Y.-X., Kalocsai, P., Chen, J.-Y., and Shams, S. (2000). Information processing issues and solutions associated with microarray technology. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 167–200. Eaton Publishing, Natick, MA. Zhou, J., Palumbo, A. V., and Tiedje, J. M. (1997). Sensitive detection of a novel class of toluenedegrading denitriﬁers, Azoarcus tolulyticus, with small-subunit rRNA primers and probes. Appl. Environ. Microbiol. 63, 2384–2390. Zhou, J., and Thompson, D. K. (2002). Challenges in applying microarrays to environmental studies. Curr. Opin. Biotechnol. 13, 204 –207. Zlatanova, J., and Mirzabekov, A. (2001). Gel-immobilized microarrays of nucleic acids and proteins: production and application for macromolecular research. Meth. Mol. Biol. 170, 17–38.

THE AGRONOMY AND ECONOMY OF BLACK PEPPER (PIPER NIGRUM L. ) — THE “KING OF SPICES ” K. P. Prabhakaran Nair Distinguished Visiting Scientist, Indian Council of Agricultural Research, New Delhi, India

I. Introduction II. The Pepper Plant — Its Botany and Chemistry A. Pepper Botany B. Pepper Chemistry III. Pepper Agronomy A. The pepper Soils B. Nutrition of Black Pepper C. Evolution of Pepper Manuring D. Response of Pepper to Mineral Nutrients IV. The Role of “The Nutrient Buffer Power Concept” in Pepper Nutrition A. The Buffer Power and Effects on Nutrient Availability B. Basic Concepts C. Measuring the Nutrient Buffer Power and Its Importance in Affecting Nutrient Concentration on Root Surfaces D. Background Informaton on the Importance of Measuring Zn Buffer Power E. Quantifying Zn Buffer Power of Pepper Growing Soils V. Establishing a Pepper Plantation A. The Indian Experience B. The Indonesian Experience VI. Pepper Pests and Their Control VII. The Processing of Black Pepper on Farm A. Sun Drying of Pepper B. Solar Drying of Pepper C. Garbling, Cleaning, and Fractionation D. Packaging and Storing VIII. An Account of Indonesian Pepper Processing IX. Industrial Processing of Black Pepper A. White Pepper B. Cryoground Pepper C. Pepper Oil and Oleoresin

271 Advances in Agronomy, Volume 82 Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved. 0065-2113/03 $35.00

272

K. P. P. NAIR X. The Future of Global Pepper Economy A. The Supply Side of Pepper Economy B. The Demand Side of Pepper Economy C. Prices and World Market D. The Pepper Price Out Look by 2020 E. The Pepper Supply Out Look by 2020 F. The Pepper Demand Out Look by 2020 G. Country-wise Economic Growth Impacting Production and Consumption XI. Pepper Economy in India A. Pepper Production Scenario in India B. Bush Pepper C. Economics of Pepper Production in the State of Kerala D. Marketing E. Pepper Futures Market XII. Pepper Pharmacopoeia A. Antiinﬂammatory and Central Nervous System (CNS) Depressant Activity of Pepper B. Effect on Hepatic Enzymes C. Carcinogenic and Mutagenic Effects of Black Pepper D. Pepper as An Antioxidant E. Pepper as An Antimicrobial Agent F. The Pharmacological Effect of Pepper on Human Health G. Clinical Applications of Pepper H. Toxicological Effects I. The Insecticidal Activity of Pepper XIII. Consumer Products of Black Pepper XIV. Value Addition in Pepper XV. Conclusions and a Peep Into Pepper’s Future Acknowledgments References

Black pepper, popularly known as the “King of Spices”, has a very checkered history dating back to the times of Queen Sheeba and King Solomon (BC 1015– BC 66) and has inﬂuenced the destiny of nations and their people, spread across the world, both economically and culturally. Today pepper commands the leading position among the different spices as the spice of immense commercial importance in world trade and is ﬁnding its way into the dietary habits of millions around the world, even among people on the European and North American continents, hitherto unaccustomed to its use. Pepper use ranges from a simple dietary constituent to that of immense pharmacological beneﬁts. Though beset with many problems, both agronomic and economic, it is a safe bet that pepper will emerge as the world’s most sought after spice, its predicted global demand escalating colossally to about 280,000 metric tons by the year 2020 that will further climb to 360,000 metric tons by the year 2050. This very exhaustive review details the various constraints to enhancing productivity of pepper and charts

AGRONOMY AND ECONOMY OF BLACK PEPPER

273

contours of a new course of action. Among the primary constraints in pepper production is the absence of an ideotype that combines many positive traits to boost production potential, whilst at the same time resisting ravages of nature such as the onslaught of the dreaded disease “Foot Rot” caused by Phytophthora fungi, whch is the most devastating. Fertility management of pepper soils is still rooted in classical “text book” knowledge. The review also includes the relevance of a refreshing new concept, developed by the author, which is now universally known as “the nutrient buffer power concept”, in enhancing pepper productivity. q 2004 Academic Press.

I.

INTRODUCTION

Known as the “King” of spices, black pepper (Piper nigrum), a perennial crop of the tropics, is economically the most important and the most widely used spice crop of the world. The history of spices is very much entwined with the history of mankind. But, within the family of spices, black pepper predominates. In ancient Egypt, when the mummiﬁed body of the Pharaoh was laid to rest in the Pyramids, it was black pepper along with gold and silver that was kept adjacent to the body, in the belief of the ancient Egyptians that even in the after life this very important spice would be of use. The ancient scriptures, Bible, Koran, and the Vedas mention the use of spices. According to the Bible, it was during the royal visit of Queen Sheeba to King Solomon (BC 1015 –BC 66) that a caravan load of spices, primarily pepper, was presented by the former to the latter. Nearly 3000 years before the birth of Christ, both Babylonians and Assyrians were trading in spices, primarily black pepper, with the people of the Malabar coast in the state of Kerala on the Indian subcontinent. Also, the ancient Indian medical texts, such as Ashtangahridaya and Samhitas, mention the use of pepper in rare and unique medical formulations. That spices, and in particular, pepper, had such a lasting impact on the economic prosperity of places is revealed by the fact that cities like Alexandria, Genoa, and Venice can trace their economic prosperity back to the vigorous trade in spices (Rosengarten, 1973). Parry (1969) observed that due to the increased demand and consumption of pepper in England and Europe, a guild known as “Pepperers” — the wealthiest of the merchants — was established in London. The high price of pepper made it the exclusive commodity in use by the rich for culinary purposes. But, it is interesting to note that with the arrival of pepper, the western kitchen transformed when “dishes took on a fullness of ﬂavor previously unknown, beverages glowed with a redolent tang, and life experienced a new sense of warmth and satisfaction” (Parry, 1969). King Solomon of Israel and the Phoenician King, Hiram of Tyre, obtained their spices from the coast of Malabar (Rosengarten, 1973). Though the Jews and Arabs were well into the

274

K. P. P. NAIR

spice trade by this time, it was the latter who had a domineering role in the spice trade, which was later usurped by the Romans. However, the Arabs were privy to the knowledge of the country of origin of pepper and even secretly kept the sea route knowledge to themselves. But, the onslaught of the Roman empire changed all that during the 1st century A.D. when Rome captured Egypt. It was around A.D. 40, during the reign of the Roman Emperor Claudius that the enterprising Greek merchant mariner Hippalus, who after discovering the full power and velocity of wind movement of the Indian ocean, a secret guarded by the Arabs, that it became known that a round trip to Egypt via India could be completed in less than 1 year (Rosengarten, 1973). A consequence of the great economic scale of this discovery was that when the sea route came to be known, the dependence on the overland trade routes declined substantially and several wealthy cities on the overland trade route went into economic penury (Ummer, 1989) and Hippalus and many others reached the coast of Malabar and returned to Rome with loads of pepper and other spices. The Roman dominance on the spice trade completely eclipsed the Arabs. Interestingly, among all the spices, it was pepper that the Romans fancied most despite its “obnoxious pungency” as noted by Pliny the Elder in his Natural History compiled between A.D. 23 and A.D. 79. It is interesting to note that just about the time Arabs were getting involved in the spice trade with the Malabar coast, the Chinese also entered the fray. As early as the 1st century A.D., a royal messenger reached the Malabar coast in search of spices and the trade relationship between China and the Malabar coast began to ﬂourish as recorded by travelers, such as Sulaiman who reached the Malabar coast in A.D. 851 (Ummer, 1989). It was the Chinese who played an important part in spreading pepper to southeast Asia and far-east Asia and it was during the voyages of Zheng He (1405 –1433) that China imported as much pepper as the total quantity imported into Europe during the ﬁrst half of the 10th century (T’ien, 1981), and the Chinese trade was an imperial monopoly. It is interesting to note that in the 15th century, soldiers in China were partly paid in pepper, and even government ofﬁcials, as observed by T’ien (1981). The fabulous Chinese ships that carried pepper were greatly praised by the venerable explorer Marco Polo (Mahindru, 1982). The colonization of the Indian subcontinent had much to do with the pepper trade. It was the lure of the spice trade that led Vasco de Gama, the great Portuguese explorer to discover the sea route to India and he landed on the Malabar coast at the Kappad beach near the presently named city of Kozhikode (Calicut) on 20th May 1498. Vasco de Gama and his men returned to Portugal immensely rich, but in the process also put an end to the Arab trade. The Portuguese King, Dom Mannuel, was so much impressed with the pepper bounty that Gama brought back home, that he sent a naval contingent with a ﬂeet of 13 battle ships to India under the command of Pedro Alwarez Cabral in A.D. 1500 and went further to declare sovereignty over India, along with other countries such as Ethiopia and Arabia (Rosengarten, 1973). The Portuguese

AGRONOMY AND ECONOMY OF BLACK PEPPER

275

domination of Kerala through pepper production and trade was so complete that this tiny state of India turned out to be the cradle of world pepper. The Portuguese rulers were ruthless and their only aim was to make the most out of the pepper trade. However, the scene changed in the ﬁrst quarter of the 17th century with the arrival of the Dutch. The Dutch were temporarily successful in elbowing out the Portuguese until the British came on the scene in A.D. 1600. In hindsight, what is most astonishing is how a trade war in pepper between the Dutch and the British led to the establishment of the British empire on the Indian subcontinent. The historians Collins and Lapiere in their monumental work “Freedom At Midnight” (1976) make a remarkable observation as follows: “Sometimes history’s most grandiose accomplishments have the most banal of origins. Great Britain was set on the road to the great colonial adventure for ﬁve miserable shillings. They represented the increase in price of a pound of pepper proclaimed by the Dutch privateers who then controlled the spice trade. Incensed at what they considered a wholly unwarranted gesture, 24 merchants of the city of London gathered on the afternoon of 24th September 1599 A.D. in a decrepit building on Leaden Hall Street. Their purpose was to ﬁnd a modest trading ﬁrm with an initial capital of £72,000 subscribed by 125 shareholders. Only the simplest of concerns, i.e., proﬁt, inspired their enterprise, which expanded and transformed, would ultimately become the most noteworthy creation of the age of imperialism — the British Raj.” What started as the British East India Company on 31st December 1600, with the stamp of approval by Queen Elizabeth I, in just 36 years, in A.D. 1636 from the day of 24th August 1600, when the 500 ton ship Hector landed in the Surat port, north of Bombay, laid the long and tortuous road to the subjugation of the vast millions of Indians through the pepper trade. In more than one sense, pepper was the cause of India losing its sovereignty. It was only in the last quarter of the 18th century that the Americans entered the pepper trade. The ﬁrst sponsored trip to the East Indies was organized by Capt. Jonathan Carnes in 1795. Though the American pepper trade ﬂourished until 1810, it later declined coinciding with the American Civil War in 1861. Compared with the Portuguese, Dutch and British, the impact of Americans on the pepper trade was only marginal. The Americans traveled to Sumatra to fetch pepper. By 1933, pepper was introduced to Brazil, and in 1938, it reached the Republic of Malagasy. By 1954, pepper was introduced to the African continent. Within the Indian Republic, the tiny state of Kerala can pride itself as the home of pepper, in particular, the coastal region of Malabar, in the state of Kerala, which accounts for 95% of the country’s area and production (Anon, 1997). Besides Kerala, two other states in the southern region, namely, Tamil Nadu and Karnataka, put together, contribute the remainder. The ﬁrst research station of pepper in the world was established in Kerala in a small town named Panniyur on the Malabar coast during 1952 –53. In addition, the ﬁrst hybrid pepper — Panniyur 1 — in the world was released by this station in 1966 (Fig. 1).

276

K. P. P. NAIR

Figure 1 Panniyur-1, the ﬁrst Pepper hybrid developed in the world at the Panniyur Pepper Research Station, in Kerala State, India.

Following research at Panniyur, pepper research began in Sarawak (Malaysia) in 1955. Following the success of the “green revolution” in India, several “Co-ordinated Research Projects” catering to the speciﬁc needs of individual crops were established by the Indian Council of Agricultural Research, headquartered in New Delhi. Thus, the All India Co-ordinated Research Project on Spices, with an intense focus on pepper, was established in 1971. More than 30 years later the project has included other spice crops as well, yet pepper continues to receive most attention against the background of the current liberalization and globalization process and the World Trade Organization (WTO) mandated changes. The future of pepper is most crucial to the economy of India vis-a`-vis the economy of the state of Kerala, where it is the economic mainstay. Many other tropical countries have made concerted efforts to grow pepper in view of its global economic importance, the most important being Indonesia where the International Pepper Community (IPC) is headquartered in Jakarta. Indonesia is second to India as a pepper growing country. It was either through the Polynesian sea-farers or the Babylonian –Chinese sea route linking the Malabar coast and southeast and far-east Asia that pepper reached Indonesia. Indonesia, known as the Dutch East Indies during the pre-second world war period, was the largest pepper producer. It was during the Japanese occupation during the war that many plantations were abandoned and production declined sharply. During the pre-war period, Indonesia had in all close to 30,000 ha of pepper (Lawrence, 1981). Malaysia and Sri Lanka are the other two major pepper producers. It was the European settlers who introduced pepper to Malaysia, while in Sri Lanka, pepper was grown as a mixed crop with other crops like cocoa and the foreign occupation of Sri Lanka helped pepper

AGRONOMY AND ECONOMY OF BLACK PEPPER

277

cultivation expand in the country. In the southeast Asian countries such as, Thailand, Vietnam, Cambodia, South Korea, and parts of south China, pepper cultivation took hold in the post-war years. Within the southeast Asian countries, Vietnam is beginning to emerge as a major pepper grower. Within the South Paciﬁc islands, Fiji is the most important pepper growing country. In South America, Brazil is the leader followed by Mexico, Guatemala, Honduras, Saint Lucia, Costa Rica, and Puerto Rico. Within the African continent, Madagascar leads the pack followed by Malawi, Zimbabwe, Benin, Kenya, Nigeria, Cameroon, Congo, Ethiopia, etc. Within Asia, where pepper production is concentrated, Malaysia takes the third place next to Indonesia, with an Agricultural Research Station in Kutching, Sarawak. Indonesia, which takes the second place, has the Research Institute for Spice and Medicinal Plants in Bogor. Though India tops the list among the producers in acreage, with a total area of 191,426 ha and a production of 56,200 metric tons, its productivity is the lowest in the world with just 294 kg ha21, while Thailand with a total area of only 2808 ha tops the list with a productivity of 3594 kg ha21 (Table I). The International Trade Centre (ITC) in Geneva, estimates the current trade in spices at 400,000 – 450,000 metric tons with a total value of US $1.5 –2 billion annually. With an annual growth rate of 3.6% in quantity and 8.4% in value in spices, pepper contributes 34% of total trade in spices. Within the industrialized west, Denmark tops the list in pepper consumption, followed closely by Germany and Belgium. The USA is a sizable consumer, while Canada and Switzerland tail the list. Post-Doha negotiations, agriculture will increasingly play a crucial role in world economy. Among the spices, pepper will play a major role. By 2010,

Table I Acreage, Production and Productivity of Major Pepper Growing Countries in the Last Decade of the 20th Century Country

Area (ha)

Brazil China India Indonesia Madagascar Malaysia Mexico Sri Lanka Thailand Vietnam

26,500 13,170 191,426 110,580 4228 8960 1294 12,080 2808 15,700

Production (metric tons) 23,400 11,045 56,200 45,240 2160 16,920 1112 5058 10,091 17,266

Productivity (kg ha21) 883 839 294 409 511 1888 859 419 3594 1100

278

K. P. P. NAIR

projected world consumption will reach 230,000 metric tons, which scales up to 280,000 metric tons by 2020, which means an annual increase of 5000 metric tons. Present production is close to 200,000 metric tons. For the next two decades close to 100,000 metric tons will be needed to balance the projected demand and consumption. World wide, especially in the industrialized countries, there is a growing demand for premium organically grown pepper. The potential for the organic food market is close to US $8 billion now in USA, which is followed by Germany and Japan, each with a market share of close to US $2.5 billion. A substantial part of this market will be for pepper. This would imply that future production strategies would need to increasingly focus on clean pepper production, which has to withstand both biotic and abiotic stresses without recourse to “high input chemical technology” — the hall mark of the so-called “green revolution”. There are areas of pepper production that simultaneously pose great challenges, while opening up new avenues. One of the most daunting, in the former category, is the evolution of a totally resistant pepper variety to the dreaded disease “Foot Rot”, caused by the fungus Phytophthora, which has wiped out many a pepper plantation. Also, pepper nutrition is still far from being thoroughly understood. The fact that it is a perennial crop adds to the lack of thorough understanding. Despite the complexity of soil science and the emergent soil management practices, the basic concept of soil as a medium of plant growth can be expected to persist for an indeﬁnite length of time (Nair, 1996). But, it is becoming clearer that the earlier views on soil as merely a “supportive medium” for plant growth is giving place to new ones on “managerial concepts” of this supportive medium. This is amply illustrated by the shift in focus from the green revolution phase of the 1960s to mid-1970s where application of increasing quantities of soil inputs, such as, fertilizers and pesticides, was emphasized, to the “sustainable agriculture” phase from the early 1980s to the present (probably to continue?); sustainable agriculture places more reliance on biological processes by adopting genotypes to adverse soil conditions, enhancing soil biological activity and optimizing nutrient cycling to minimize external inputs, such as fertilizers, and maximize their efﬁciency of use. In fact, the paradigm of the earlier phase has given way to the emergent new paradigm (Sanchez, 1994), and this is clearly reﬂected in the dialogue of the world leaders during the Earth Summit in 1991 in Rio de Janeiro, Brazil, where Agenda 21 has incorporated six chapters on soil management (Keating, 1993). This review on pepper, while discussing its overall production proﬁle in the world, will lay a special emphasis on the second paradigm inasmuch as prescriptive soil management for pepper production is concerned with regard to understanding the soil nutrient bioavailability and its efﬁcient management in the pepper production. On account of the paucity of published literature on this aspect, the focus will

AGRONOMY AND ECONOMY OF BLACK PEPPER

279

only be with regard to speciﬁc nutrients, such as zinc, which is becoming increasingly important in pepper production.

II. THE PEPPER PLANT — ITS BOTANY AND CHEMISTRY A. PEPPER BOTANY Black pepper (P. nigrum) is a perennial plant and derives its name Piper, perhaps, from the Greek name for black pepper, Piperi (Rosengarten, 1973) and most of the European names for black pepper were derived from the ancient Indian language, Sanskrit, such as Pippali, the name for long pepper (P. longum). It was the great botanist Linnaeus (1753) who established the genus Piper in his Species Plantarum. In this monumental work, Linnaeus recognized 17 species in the Piper family, all of which were included in the same genus. Ruiz and Pavon in 1794 introduced the second genus in the family, namely, Peperomia (Trelease and Yuncker, 1950). The family name Piperaceae was used for the ﬁrst time in 1815 by L.C. Rich in Humboldt, Bonpland, and Kunth’s Nova Genera et Species Plantarum (Yuncker, 1958). All the species known in the family Piperaceae during the early years of systematic classiﬁcation were included in the classical monographic study Systema Piperacearum published in 1843 by F.A.W. Miquel. It was Rheede in the year 1678 who made the earliest record of the description of Piper in the Indian subcontinent. In his Hortus Indicus Malabaricus, the earliest printed document of plants on the Malabar coast, Rheede described ﬁve types of wild pepper that included both black and long pepper. Linnaeus (1753) included 17 species from India in his monumental work Species Plantarum. The ﬁrst major study of the Piper spp. from the Indian subcontinent was that of Hooker (1886) in his book Flora of British India. However, the most authoritative ﬂoristic study of the Western Ghats of southern India was that of Gamble in 1925 in his book Flora of Presidency of Madras in which the following 13 species with their taxonomic keys were described: P. argyrophyllum, P. attenuatum, P. barberi, P. brachystachyum, P. galeatum, P. hapnium, P. hookeri, P. hymenophyllum, P. longum, P. nigrum, P. schmidtii, P. trichostachyon, and P. wightii. Following the publication of Gamble, no new additions to the list were made until 1981, when in this year a new species P. bababudani, from the Bababudin hills of the state of Karnataka (the neighboring state of Kerala) was added (Rahiman, 1981a); however, this was not published. Ravindran et al. (1987) reported a new species, P. silentvalleyensis, the only bisexual wild species from the world-renowned “Silent Valley” in the Western Ghats. Other new species are P. pseudonigrum (Velayudhan and Amalraj, 1992), and P. sugandhi (Nirmal Babu et al., 1993a).

280

K. P. P. NAIR

Black pepper cultivars could possibly have originated from wild ones through the process of domestication and selection. More than a hundred cultivars are known, but most have vanished due to the onset of devastating diseases like Foot Rot and also replacement by hybrids. Human migration has contributed to the spread of these cultivars. Pepper distribution is extensive in moist evergreen forests and to a lesser extent in semi-evergreen and moist deciduous forests of the Western Ghats of southern India, growing from almost sea level to a height of 1500 m above mean sea level. Population structure of any species is determined mainly by the breeding system of the species, pollen, fruit, and seed dispersal mechanism and the presence or absence of isolation mechanisms. In Piper, male, female, and hermaphrodite forms exist. The cultivated P. nigrum is monoecious, having hermaphrodite ﬂowers, while the wild ones are mostly dioecious. Human selection has played a major role in the directional evolution of hermaphroditism in the cultivated pepper. Predominantly self-pollinated, pepper pollen dispersal is aided by rain or dew drops and also by geitonogamy — the gravitational descent of pollen. Though ﬂowers are protogynous, in the absence of an active pollen transfer mechanism, protogyny is ineffective in outbreeding. Active and efﬁcient pollen and seed dispersal mechanisms ensure gene ﬂow within and between population segments leading to the establishment of intergrading populations. Absence of such a mechanism in Piper ensures effective isolation barriers among individuals and population units. Segregation in the seedling progenies leads to variations in such units, and also, mutations and chance crossing followed by segregation. Any such variation stemming in population is ﬁxed immediately because of the vegetative mode of propagation and such a unit may over the course of time, diverge from other similar units. Quite often different types of P. nigrum or different species climb up a single tree, which enhances the chances of outcrossing, which results in hybrid seedlings. Progenies from such chance crosses grow, and later climb up the same or nearby trees and chance outcrossing with the parental vine or its clonal or other seedling progenies, which result in further back crossing or hybrid progenies, all of which lead to substantial variation within the population. These forces acting together would have most likely contributed to the evolution of many present day pepper cultivars (Ravindran et al., 1990). Though pepper has originated in the evergreen forests of the Western Ghats in southern India, with extensive occurrence of wild pepper in the less disturbed forest areas, no study has so far been carried out to ﬁnd out the origin of P. nigrum. The basic chromosome number of Piper is x ¼ 13 and that of P. nigrum 2n ¼ 52, a tetraploid (Mathew, 1958, 1972; Jose and Sharma, 1984). On the basis of morphological and biosystematic studies, Ravindran (1991) reported that three species, namely, P. wightii, P. galeatum, and P. trichostachyon, could be the putative parents of P. nigrum. These three are woody climbers with similar texture and leaf morphology. Their spikes and fruits are more similar to that of

AGRONOMY AND ECONOMY OF BLACK PEPPER

281

P. nigrum than those of other species. Of the three, P. wightii and P. galeatum are the most probable ancestors of P. nigrum. Pepper is a shade loving plant that requires a constant moisture supply during dry spells as it has a high evapotranspiration coefﬁcient (Raj, 1978). However, even under favorable soil moisture conditions, when exposed to direct sunlight the plant develops certain physiological disorders (Vijayakumar et al., 1984). Pepper is highly sensitive to the growth-light regime (GLR) and plant parts exposed to high GLR produce more fruits per unit surface area than those exposed to low GLR (Montaya et al., 1961). Where grown in permanently shaded situations productivity is poor. Under such conditions, plants respond positively to radiation and a positive correlation has been shown to exist between productivity and radiation. Except in the case of “bush pepper”, normal pepper vines are grown on artiﬁcial support. These supports are called “standards”. Ramadasan (1987) reports three types of pepper canopies. Canopy shaping is largely based on the shading provided by live support and that of the adjoining shading trees. When dead supports, such as brick pillars or reinforced concrete poles are used, as in Thailand, the canopy does not taper at the top. Trees that grow adjacent to dead supports lead to tapering of the canopy at the base due to partial shading. Such tapering at the base is not observed when there are no competing trees. In Sarawak, the wood of Bornean ironwood (Eusideroxylon zwageri) is used for support and generally, no shade tree is grown. In such dead supports, the top of the canopy unfolds like an umbrella. The pepper plant is endowed with two advantages that many others do not have. First, it can be vegetatively propagated and second, it is viable for sexual reproduction. Both offer much scope for the exploitation of hybrid vigor as well as selection breeding. Published literature show that clonal selection, hybridization, open pollinated progeny selection, mutation, and polyploidy have been successfully experimented with to improve pepper yield and impart desirable plant traits. In recent times, biotechnological methods are also being employed to impart pathogen resistance. Quality is the most important criterion in breeding as far as spices are concerned and pepper is no exception. The quality of pepper is decided mainly by the amounts of piperine, and oleoresin (essential oil). An alkaloid, piperine contributes to the pungency of the berry, and oleoresin content enhances the ﬂavor. The unique pepper ﬂavor is inﬂuenced by very many chemical compounds present in the essential oil and a number of genes control this unique character. Conservation of pepper genetic resources is the key to sound pepper breeding. Among the 18 biodiversity hot spots on this planet, the Western Ghats of southern India is one of the most unique, with their different forest types, such as tropical wet evergreen, tropical moist and dry deciduous, montane subtropical and temperate ones. It is on the western side of the Western Ghats that the population

282

K. P. P. NAIR

and diversity of Piper is in abundance. It is important to realize that during the past century the Western Ghats have been ravaged the most, leading to ecological breakdown, due to clandestine timber felling and uncontrolled human encroachment. The Western Ghats is home to spices that are endemic to the region. Of the 17 species of Piper found here, 11 are endemic to the region (Ravindran and Peter, 1995). More than a hundred cultivars were under cultivation in the region, but most of them have vanished from the pepper growers gardens, probably due to such dreaded diseases as Foot Rot (Phytophthora) and slow decline and also, replacement of old and unproductive cultivars by the few high yielding elite ones. It takes millennia for nature to nurture plant diversity, but human activity, productive or otherwise, can change all that in a very short span of time. As part of the conservation efforts, some pepper growing countries, notably India, has made a concerted effort to collect and catalogue pepper germplasm in a systematic manner. Mention should be made of the efforts in this direction undertaken at the Indian Institute of Spices Research at Kozhikode, Kerala State. The Institute has established a National Conservatory for in situ conservation. Conservation is carried out in four stages (Ravindran and Nirmal Babu, 1994a) as follows: conservation as a nursery gene bank by trailing each accession in split bamboo pieces of about 2– 3 feet length serially, which will be maintained and continuously multiplied. Conservation in the clonal repository where 10 rooted cuttings from each accession are maintained. Conservation in the ﬁeld, as ﬁeld gene bank, by planting accessions for pilot yield performance and characterization and conservation in vitro and in cryogen banks. With the establishment of the All India Co-ordinated Research Project on Spices in 1971, collection and conservation of pepper germplasm on a smaller scale has also been made by the Panniyur Pepper Research Station (the ﬁrst pepper research station established in the world) in the state of Kerala, Pepper Research Station at Sirsi, in the state of Karnataka, Horticultural Research Station, Chintapalli, in the state of Andhra Pradesh and Horticultural Research Station, Yercaud, in the state of Tamil Nadu. Among all of these centers, the last one has the largest collection, but it is the Indian Institute of Spices Research that can pride itself as the one holding the largest collection in the world. On a much smaller scale, collections and conservation have also been made, notably in Indonesia and Malaysia, with the former holding 40 cultivars and seven Piper species and Sarwak (Malaysia) holding 18 cultivars of P. nigrum, 18 identiﬁed Piper spp. and 98 unidentiﬁed accessions (De Waard, 1984). Despite the fact that they are not closely related to the cultivated black pepper, the vast number of Piper spp. found in central and southern America, north eastern India, Indonesia, and Malaysia have not been collected and systematically catalogued. The vast germplasms in these regions could form a very important reservoir of gene pool to develop varieties with in-built disease and pest resistance (Table II). A new avenue for pepper improvement is through biotechnological research, which of late has made great strides in many areas of biology. The

AGRONOMY AND ECONOMY OF BLACK PEPPER

283

Table II Some Improved Pepper Varieties from Major Pepper Growing Countries Country of origin and Average yield name of the variety (kg ha21)

Remarks

India Panniyur 1

1242

Panniyur 2 Panniyur 3 Panniyur 4 Panniyur 5

2750 1953 1277 1098

Subhakara Sreekara Panchami Pournami PLD-2

2352 2677 2828 2333 2475

Indonesia Natar1

–

Natar 2

–

Malaysia Semongok Perak Semongok Emas

– –

High yield Uniquely high yield, tolerant to Foot Rot, Black Berry disease and pepper weevil

Madagascar Sel. IV.1 Sel. IV.2

– –

High yield High yield

Sri Lanka PW 14

–

Claimed to be resistant to Radopholus similis

Well suited to most regions, performs rather poorly at higher elevations and under shade Well suited to most regions and shade tolerant Well suited to most regions, rather late in maturity Consistently good yielder suited to most regions Most important characteristic is tolerance to nursery diseases and shade Suited to most regions, berries of good quality Suited to most regions, berries of good quality Suited to most regions, late in maturity Good tolerance to root knot nematode Suited to most regions, berries of high quality High yield and particularly tolerant to the dreaded disease Foot Rot and nematodes High yield and particularly tolerant to the dreaded disease Foot Rot and nematodes

Note: Average yield unknown.

biotechnological approaches permit the researchers to manipulate plant tissues and cells in vitro. A better understanding of genetics is provided by tissue culture techniques and investigations at the molecular, cellular, and organismal levels. The possibilities of cell culture and plant tissue culture for breeding and plant propagation purposes has led to large-scale commercial exploitation. Technical perfection by manipulating plant cells at the molecular level through recombinant DNA technology has opened up enormous possibilities in the creation of transgenic plants and this is an area that is assuming industrial proportions. Inasmuch as pepper is concerned, one area where recombinant DNA technology could be successfully exploited is the evolution of pepper varieties tolerant to the

284

K. P. P. NAIR

dreaded disease Foot Rot caused by Phytophthora capsici. None of the varieties in cultivation is tolerant to this dreaded disease. Within the Piper spp., a distant relative of P. nigrum, namely P. colubrinum from South America, is resistant to Phytophthora and, hence, biotechnological approaches which will enable transfer of resistance governing genes from P. colubrinum to P. nigrum will be the only mode to contain the devastation from this dreaded disease. Another area of interest is micropropagation, where in vitro culture methods for pepper cloning is done. Micropropagation was ﬁrst introduced by Broome and Zimmerman (1978). For micropropagation, several plant parts, such as shoot tips, nodal segments, and apical meristems, from both juvenile and mature plants have been used (Mathews and Rao, 1984; Agarwal, 1988; Philip et al., 1992; Nazeem et al., 1993; Nirmal Babu et al., 1993a,b; Joseph et al., 1996; Lisamma et al., 1996). However, a serious limitation to micropropagation is the incidence of bacterial contamination (Kelkar and Krishnamoorthy, 1996), though contamination is minimal in seedling explants. The authors suggest the incorporation of antibiotics in the culture media to arrest the endogenous bacterial contamination in the in vitro pepper cultures. Field establishment of micropropagated pepper is feasible, though endogenous bacterial contamination and phenolic exudates from the cut surface could adversely affect ﬁeld establishment (Raj Mohan, 1985; Fitchet, 1988a,b). The authors suggest treating the explants with fungicides prior to routine sterilization followed by frequent transfer to fresh medium to keep in check the problem of contamination. Madhusudhanan and Rahiman (1966) suggest the use of activated charcoal, at the rate of 200 mg l21, which could reduce browning of the explants and culture medium. Micropropagation has been in use for quite some time (Chu, 1981; Fitchet, 1988a,b). Culture media commonly used in micropropagation are those of Murashige and Skoog (1962), Schenk and Hildebrandt (1972) and McCown and Amos (1979). Indole acetic acid (IAA), naphthalene acetic acid (NAA), 2,4,-dichlorophenoxy acetic acid (2,4,-D) are some the commonly used growth regulators and additives used in the culture formulation. One other area of pepper improvement is the Agrobacterium mediated gene transfer. Reference to the technique is made by Sasikumar and Veluthambi (1994) and Sim et al. (1995). The technique can also be helpful in transferring disease resistance genes from P. colubrinum to P. nigrum in the light of the dreaded disease Foot Rot. Work on transgenic pepper to evolve delayed ripening ensuring uniform maturity — which will ensure reduction in labor costs for harvest — in Sarawak (Malaysia) has led to promising results. Conservation of pepper germplasm assumes high practical importance in improving the crop for enhanced productivity and acquisition of many desirable traits, the most important being tolerance to dreaded diseases such as Phytophthora caused Foot Rot. Normal conservation of crop germplasm is through seed banks. Since pepper is propagated only vegetatively (seeds are heterozygous and unpredictable in behavior) in vitro storage of germplasm is

AGRONOMY AND ECONOMY OF BLACK PEPPER

285

a viable alternative. In vitro conservation of pepper germplasm and its related species, such as P. barberi, P. colubrinum, P. beetle, and P. longum by deﬁning protocols have been made by Geetha et al. (1995) and Nirmal Babu et al. (1996b) by maintaining cultures at reduced temperatures in the presence of osmotic inhibitors at low nutrient levels and also by reducing to the minimum evaporative losses using closed containers. When the conserved materials were transferred to multiplication medium after storage, all the species exhibited normal multiplication rates. When the normal-sized plantlets were transplanted into ﬁeld soil, an establishment rate exceeding 80% was obtained. Normal pepper plants, morphologically similar to mother plants, were observed to develop. There is a steep decline in viability in pepper seeds associated with moisture depletion. Seeds cryopreserved in liquid nitrogen at 2 1968C with an initial moisture percentage of 12 have a survival rate of only 45%, while at 6% moisture it reduced to just 10.5% (Chaudhary and Chandel, 1994) showing clearly how seed moisture is so crucial to seed survival. There is yet another manner in which pepper plants can be propagated and that is by way of producing synthetic seeds, which consist of somatic embryos or shoot buds encapsulated in a protective coating that is biodegradable. The system facilitates not only low-cost propagation but also germplasm conservation and exchange. Encapsulation of disease-free bud employing tissue culture techniques helps production of diseasefree plantlets. Pepper shoot buds 0.5 cm long were used for production of synthetic seeds, which could be stored up to nine months in sterile water. The method is described by Sajina et al. (1996).

B. PEPPER CHEMISTRY It is the unique pungency and aroma of pepper that has both intrigued as well as fascinated pepper chemists. As already mentioned in the review earlier, the essential oil in the berry contributes to the aroma while the alkaloid piperine imparts the unique pungency. Oleoresin, which has a very great commercial value, is extracted from the dry powdered berries by solvents, and is the product that imparts the unique aroma and pungency in pepper. Dramatically put, it is the chemistry of pepper emanating from oleoresin that imparts that rare and unique blend of both aroma and ﬂavor, which is behind wars and enslavement, as much as love and tragedy, as has been detailed in the introductory part of this review. The earliest investigations on pepper chemistry go back more than half a century (Guenther, 1950) and subsequent research has enriched it (Wealth of India, 1969; Govindarajan, 1977; Lawrence, 1981; Purseglove et al., 1981). The various compounds occurring in Piper spp. are listed by Parmar et al. (1977). It was in the early 19th century that the ﬁrst report on essential oil in

286

K. P. P. NAIR

pepper was made by Dumas and later by Subeiran and Capitaine (Guenther, 1950). These researchers concluded that pepper oil is almost free of oxygenated constituents. By treating a fraction of the oil boiling at 1768C with acid and alcohol, Eberhardt obtained terpin hydrate and the presence of 1-phellendrene, caryophyllene, and tentatively dipentene were reported by Schimmel and Company and by Schreiner and Kremers (Guenther, 1950). The steam distillation of essential oil obtained from dry powdered Malabar pepper berries showed the presence of a-pinene, b-pinene, 1-a-phellendrene, DL -limonene, piperonal, dihydrocarveol (a compound melting at 1618C), b-caryophyllene and a piperidine complex (Hasselstrom et al., 1957). Additionally, the presence of cryptone, epoxydihydrocaryophyllene, and possibly, citronellol and an azulene were also reported. The presence of a- and b-pinenes, limonene, and caryophyllene in the hydrocarbon part of black pepper oil was conﬁrmed by infrared spectroscopy (Jennings and Wrolstad, 1961). It is important to mention that a renewed thrust was given to the study of chemical compounds in pepper after the advent of gas chromatography. Modern researchers used thin layer chromatography, gas chromatography, column chromatography, and vacuum distillation, etc., to separate the constituents and employed ultra violet, infra red, nuclear magnetic resonance, and mass spectroscopy for identiﬁcation. As many as 135 compounds, consisting of monoterpenoids, sesquiterpenoids, aliphatic, aromatic, and those of miscellaneous nature, have been reported (Ikeda et al., 1962; Sharma et al., 1962; Nigam and Handa, 1964; Wrolstad and Jennings, 1965; Muller and Jennings, 1967; Muller et al., 1968; Richard and Jennings, 1971; Russel and Else, 1973; Debrauwere and Verzele, 1975a,b, 1976; Lawrence, 1981; Gopalakrishnan et al., 1993). Different researchers have reported wide variations in the chemical composition of essential oils in pepper and these variations originate on account of several reasons, such as varietal differences, their geographic origin, variations in the maturity of raw material, procedural differences in oil extraction and non-resolution of constituents in early gas chromatographic analysis employing packed columns. In general, composition of essential oils will depend to a certain extent on the method of preparation, as for example, steam distillation will give oils containing about 70– 80% monoterpene hydrocarbons, 20– 30% sesquiterpene hydrocarbons and less than 4% oxygenated constituents. Vacuum distilled oils will contain less monoterpene hydrocarbons and more sesquiterpene hydrocarbons and oxygenated constituents. It is the result of incomplete distillation and the poor recovery of the high boiling sesquiterpene hydrocarbons and oxygenated constituents which lead to their low contents in steam distilled oils compared with vacuum distilled oils. Seventeen pepper cultivars from the state of Kerala were analyzed for their essential oil contents by Lewis et al. (1969) and Richard et al. (1971) and comparable results were obtained. Gopalakrishnan et al. (1993) analyzed the four Panniyur genotypes (Panniyur-1, Panniyur-2,

AGRONOMY AND ECONOMY OF BLACK PEPPER

287

Panniyur-3, and Panniyur-4) developed at the Pepper Research Station in Panniyur using a combination of gas chromatography and mass spectroscopy and Kovats indices on a methyl silicone capillary column and found, on the whole, that Panniyur-4 contained the most chemical constituents as compared with the other three genotypes. Signiﬁcant differences could also be found in 12 samples from Lampong and 16 samples from Sarawak, both in Malaysia, using the procedure of Richard et al. (1971) by Rusel and Else (1973). An interesting aspect of research refers to sensory evaluation of essential oils in pepper. There are but scanty references with regard to the relationship between odor characteristics and oil composition. The characteristic odor of pepper oil has been attributed to the meager amounts of oxygenated constituents (Hasselstrom et al., 1957). Arctander (Purseglove et al., 1981) describes the odor of pepper oil as fresh, dry –woody, warm – spicy, and identical to that of black pepper corn. Pangburn et al. (1970) after a systematic and thorough sensory evaluation of pepper oil obtained from pepper along the Malabar coast, using column chromatographic fractions and mixtures of fractions of the oil, reported that the early fraction was considered pepper-like and ﬂoral, the late fraction pepper-like, fresh and woody and the middle fraction remaining in between. Direct snifﬁng at the eluting port of the gas chromatographic columns helped distinguish the distinct odor of black pepper and ascribe it to the three areas of the late fractions. Distinctive odor analysis has been developed by Harper et al. and Sydow et al. (Govindarajan, 1977). Using a descriptive odor proﬁle based on a four-point category scale and subjecting the oils to a ranking list, Gopalakrishnan et al. (1993) have described the odor evaluation of the four Panniyur genotypes described earlier in this review. Besides aroma (odor), the other aspect that has been of interest to the researchers is the pungency emanating from the alkaloid piperene. From the early 19th century onwards, the pungency of pepper has been investigated starting with the ﬁrst report of Oersted who isolated piperine in 1819 (Guenther, 1950). Piperene is a yellow crystalline substance that was subsequently identiﬁed as the trans form of piperoyl piperidine. Subsequent researchers showed that piperine was not the only substance imparting pungency to pepper and Bucheim (Govindarajan, 1977) obtained a dark oily resin, which he called “chavicine”, subsequent to the removal of piperine from the oleoresin. Chavicine was supposedly far more pungent than piperine leaving a much sharper bite on the tongue than the crystalline piperine. This was shown to be incorrect by subsequent researchers who demonstrated the extreme pungency of piperine in solution. The debate, whether it was chavicine or other possible isomers, such as isopiperine or isochavicine, which are more pungent than piperine, continued for almost a century. However, recent investigations reveal that it is piperine which is the most pungent and chavicine is a mixture of piperine and many other minor alkaloids. The question why pepper blackens has interested many researchers. Lewis et al. (1976) attributed blackening of pepper berries due to enzymatic oxidation

288

K. P. P. NAIR

of polyphenolic substrates present in the skin of green pepper. The blackening of pepper is a chemical reaction akin to browning in fruits and vegetables and can have an enzymatic or non-enzymatic origin. Formation of coloring pigments because of enzymatic browning is triggered by the enzyme polyphenol oxidase. It is due to the enzymatic oxidation of polyphenolic substrates in the skin of green pepper that leads to blackening (Lewis et al., 1976). The authors developed a process in which green pepper was blanched to arrest the enzyme reaction with subsequent drying in a cross ﬂow drier to obtain dehydrated green pepper. Pruthi et al. (1976) reported that green pepper could also be preserved in brine containing either acetic or citric acid. In addition to the above chemical constituents, starch is a predominant constituent of pepper and its content may vary between 35 and 40% in terms of weight. In addition to starch, pepper also contains protein and fat. Reports on both starch and protein are practically non-existent. As for fat, pepper contains 1.9 – 9% fat. Bedi et al. (1971) and Salzer (1975) have determined the fatty acid composition of pepper and reported that of the different fatty acids, such as palmitic, oleic, linoleic, and linolenic, it is the linoleic acid that is most abundant (25 –35%) and the least is linolenic (8 –19%). The survey of pepper chemistry shows that though some of the major constituents have been researched during the last few decades, in the future, one may still come across constituents hitherto unknown. The research so far has mostly been conﬁned to the varieties in cultivation in different regions, world wide, and yet, there are far more cultivars that remain totally out of the chemists’ laboratory. It is important to research these, because one may come across unique chemical compositions and ﬂavor proﬁles in the different cultivars. As with disease resistance, unique chemical constitution is also of immense commercial importance. Only when pepper breeders, biotechnologists, and pepper chemists join hands in an intense search for unique traits will the research on pepper chemistry have any meaningful outcome, because it is only through these intense efforts that unique pepper lines could be identiﬁed and isolated for future breeding and crop improvement programs.

III. PEPPER AGRONOMY Agriculture systems differ from natural systems in one fundamental aspect: while there is a net outﬂow of nutrients by crop harvests from soils in the ﬁrst, there is no such thing in the second (Sanchez, 1994). This is because nutrient losses due to physical effects of soil and water erosion are continually replenished by weathering of primary minerals or atmospheric deposition. Hence, the crucial element of sustainability is the nutrient factor. But, of all the factors, the nutrient factor is the least resilient (Fresco and Kroonenberg, 1992). The thrust of high

AGRONOMY AND ECONOMY OF BLACK PEPPER

289

input technology, the hallmark of the “green revolution”, in retrospect, or the moderation by low input technology, the foundation stone of sustainable agriculture, in prospect, both dwell on this least-resilient nutrient factor. If the pool of nutrients in the soil, both native and added, could be considered as the “capital”, efﬁcient nutrient management might be analogous to raising the “interest” accrued from this capital in such a way that there is no great danger of erosion of this capital. Hence, sound prescriptive soil management should aim at understanding the actual link between the capital and the interest so that meaningful practices can be prescribed (Nair, 1996). This is all the more important for a perennial crop like pepper whose life span can be upwards of 25 years. An effort in this direction was initiated by Nair recently (Nair, 2002). But, before we dwell on it, it is only fair that an objective review is made of what has been done so far on the agronomy of black pepper.

A. THE PEPPER SOILS Pepper easily grows on a variety of soils. De Waard (1969) reported that in Malaysia pepper mostly grows on soils that are developed from slate or sandstone and even on soils of alluvial origin low in fertility. Pepper is seen to grow in Indonesia on all types of soils ranging from fertile and friable volcanic soils to sticky clayey soils. In Sri Lanka, pepper growers prefer red clay loam or sandy loam. In India, pepper is seen to grow on a variety of soils, but, in its native Kerala state, it is grown mostly on red laterites. Pepper prefers a well-drained soil having adequate water holding capacity, rich in humus and essential plant nutrients. In India, the major pepper growing soils are Oxisols (6%), Alﬁsols (70%), Mollisols (10%), and Entisols (4%). The state-wise distribution is given below (Sadanandan, 1994) (Table III).

Table III Type, Order, and Sub Order of Major Pepper Growing Soils and Their State-Wise Distribution in India Soil type

Order

Forest loam

Mollisols

Laterite

Alﬁsols Oxisols Entisols Alﬁsols

Alluvium Red loam

Suborder Udolls Ustolls Ustalfs Ustorthcut Ustalfs Ustults

Indian state Kerala, Karnataka, and Tamil Nadu Kerala, Karnataka, and Tamil Nadu Kerala and Karnataka Kerala and Karnataka

290

K. P. P. NAIR

Forest loam is acidic in reaction, with a pH range of 5.0– 5.5 and is conﬁned mainly to the western Ghats Highlands. The soils are rich in organic carbon, well drained, brown to black in color and are very well suited for pepper. Soils, in general, are quite fertile, rich in N and K, medium in P. Laterites are, again, acidic, with a pH range of 5.0 –6.2, low in fertility and invariably run into problems with P fertilization because of high amounts of soluble Fe and Al, which render applied P immobile. Excessive soluble Al leads to Al toxicity. Of late micronutrient deﬁciencies, in particular, Zn, have been reported and recently Nair (2002) has demonstrated the lacunae in Zn nutrition of pepper based on classical textbook knowledge of Zn fertilization in these problematic soils. Alluvium is conﬁned to the river banks and their tributaries. The soils are only moderately fertile and respond well to management. The soils are acidic in nature, with the pH ranging from 5.0 to 6.5. Red loams are highly porous and friable, low in fertility and like laterites contain high amounts of soluble Al and Fe which lead to problems with P fertilization. Except in forest loam, where pepper is grown as a monocrop, as the soil is highly suitable for the crop, in all the others it is grown both singly or in association with other crops, such as coconut (Cocus nucifera), arecanut (Areca catachu), etc. Unlike the large exclusive pepper plantations, these soils support what is generally known as “homestead farming” in Kerala — small, less than 5 ha, self-supporting family farming, where even milch animals are reared in association.

B. NUTRITION

OF

BLACK PEPPER

The fundamental difference between agricultural systems and natural systems is that while there is a net outﬂow of plant nutrients by crop harvests from soils in the former, there is no such thing in the latter. This is because nutrient losses due to physical effects of soil and water erosion are continually replenished by weathering of primary minerals or atmospheric deposition in natural systems. Hence, the crucial element of sustainability of crop production is the nutrient factor. But, of all the factors, the nutrient factor is the least resilient to management. The thrust of the “high input technology”, the hallmark of the “green revolution”, in retrospect, or the moderation of the “low input technology”, the foundation stone of “sustainable agriculture”, in prospect, both dwell on this leastresilient nutrient factor. For production sustainability, the nutrient factor is the most crucial. It is becoming increasingly clear that it is not the quantum per se of a speciﬁc nutrient in the soil that is crucial as far as plant requirements are concerned, but its bioavailability. In the ﬁnal analysis, it is the plant and plant alone, which will decide whether a nutrient inherently present in the soil or externally added to it, will ultimately be “available” to the plant or not (Nair, 2002). Formidable effort has gone into the task of deﬁning this “availability”. It is the conviction of this author that if sound and accountable prescriptive fertilizer

AGRONOMY AND ECONOMY OF BLACK PEPPER

291

recommendations have to be made to sustain crop production in the decades to come, one must have a clear understanding of the dynamics of nutrient availability from which accurate predictions can be made and sustainable and accountable ﬁeld practices are devised. These considerations particularly are important to devise accurate fertilizer recommendations for pepper, because, unlike most other annual or biennial crops, pepper is a perennial one and the utilization pattern of applied or native nutrients over several years, often running into decades, could be uniquely different. This part of the review will, at ﬁrst, dwell on the available pool of information relating to the general and prevalent mode of pepper nutrition, and then will go on to the next stage, where it will encompass the newer vision, which is now known, the world over, as “the nutrient buffer power concept”.

C. EVOLUTION

OF

PEPPER MANURING

Pepper has been in cultivation for many decades now in India and elsewhere. In earlier days, farmers used only meager quantities of organic manures, such as leaf litter, animal manure (principally cow dung, which is a wide-spread practice of homestead farming in Kerala even now) or slashed stems and leaves of live support, such as Erythrina indica. Chemical manuring, through factorymanufactured fertilizers, is a relatively recent phenomenon. It was the “pepper boom” — the escalating prices of pepper in world market, especially in the post second world war period — that enthused the pepper farmers to go in for artiﬁcial manure. It was the spectacular initial yield increases in wheat and rice in the North Indian regions — the green revolution effect owing to liberal doses of chemical fertilizers — which prompted the afﬂuent pepper farmers to go in for factory-produced fertilizers. Though there is now a blacklash — in terms of degraded soils, drying aquifers and vanishing biodiversity, all of which are attributed to the indiscriminate use of chemical fertilizers — against the use for chemical fertilizers, chemical fertilizer use in a systematic manner is still conﬁned only to afﬂuent owners of pepper plantations. Organic manuring is still the mode of pepper nutrition in homestead farming. Signiﬁcantly, even pepper planters of fairly large farms are gradually switching to organic farming, because, “organic pepper” raised by the use of only organic manures fetches a much better price and a steady and premium market is developing, especially in European countries such as Germany, where the spices are much fancied lately.

D. RESPONSE

OF

PEPPER

TO

MINERAL NUTRIENTS

The major emphasis on pepper fertilization is still conﬁned to nitrogen, phosphorus, and potassium. There is an extensive body of literature on nitrogen

292

K. P. P. NAIR

fertilization in pepper globally (De Waard, 1964; Sim, 1971; Pillai and Sasikumaran, 1976; Pillai et al., 1979, 1987; Mohanakumaran and Cheeran, 1981; Sadanandan, 1994). These routine studies have dealt with partitioning of N in various plant parts (Adzemi et al., 1993), response functions ( Pillai et al., 1987; Sadanandan, 1990,), greenhouse cultivation (Murni and Faodji, 1990) and foliar N application (Anon, 1995). Adzemi et al. (1993) found the maximum N concentration in leaves (2.30%) while the branches transported the most (47.6 mg plant21 year21). In Kerala state, for Panniyur-1, the ﬁrst hybrid variety released in the world from the Pepper Research Station, Panniyur, 50 Kg N with 100 kg P2O5 and 150 kg K2O ha21 were found to be the optimum rates for laterite soil (Pillai et al., 1987). In greenhouse conditions 1.1 g N plant21 as urea and 2.0 g K2O plant21 produced the maximum of dry matter (Murni and Faodji, 1990). Between Panniyur-1 and Karimunda, another improved variety from the Panniyur Research Station, 292 kg ha21 year21 and 183 kg ha21 year21 of N, respectively, were found to be the optimum rates (Sadanandan, 1990). In Sarawak (Malaysia), foliar application of 0.7% urea at weekly intervals totaling nine sprays were found to increase pepper yield by 22% (Anon, 1995) (Table IV). Comparatively, research on P in pepper is less, viewed against that of N. Unlike N, pepper stems contain the most P (3.96 g vine21) because of large transport and a total of 22.8 kg ha21 of P is removed from the soil (Adzemi et al., 1993). Availability of P in laterite soil is a universal problem, because of the presence of excessive soluble Al and Fe, which render the applied P fertilizer immobile and some researchers have focused on alternate sources of P as a possible remedy. Phosphate rocks (PRs) are ideally suited for plantation crops and where indigenously available could be proﬁtably utilized. In northern India, a popular PR is “Mussorie Rock Phosphate” (MRP). Sadanandan (1994) reported that MRP is superior to ordinary super phosphate. Both cumulative yield increase and relative agronomic effectiveness were superior in MRP as compared to ordinary super phosphate. Pepper is a proliﬁc user of K and even with 2% content of K in the pepper plant, K deﬁciency will manifest (De Waard, 1969). Field trials conducted using Panniyur-1 as the test crop showed that as much as 200 g of K2O vine21 is required (Pillai et al., 1987). Response functions to K fertilizer application indicated that pepper needs a very high dose of K (270 kg K2O ha21) for high yield (Sadanandan, 1990). Pepper plants remove as much as 203.2 kg ha21 of K and white pepper stores most of it (42.0 g vine21) as reported by Adzemi et al. (1993). Most pepper growers adopt varying N:P:K ratios, while Pillai et al. (1979) suggest an optimum ratio of 5:5:10 for N, P and K to obtain the maximum yield. As far as the secondary nutrients (Ca, Mg, and S) are concerned, the only published work refers to Ca, but, indirectly through the effect of liming. Pillai et al. (1979), however, did not observe any positive response to lime application in a

AGRONOMY AND ECONOMY OF BLACK PEPPER

293

Table IV Nutrient Deﬁciency Symptoms in Pepper Nutrient Nitrogen

Phosphorus

Potassium

Calcium

Magnesium

Sulfur

Zinc

Manganese

Iron

Copper

Boron

Deﬁciency symptoms General chlorosis and stunted growth. Yellowing of older leaves followed by younger ones. Bottom leaf tips and margins become brown and necrotic Bronzing of older leaves accompanied by necrosis of leaf tips and margins. Stunted growth Browning and necrosis of older leaf tips and margins. Symptoms later spread to younger leaves Young leaves develop tiny, brown necrotic pin-head spots which later spead to older leaves. Interveinal chlorosis can also be seen Pale yellow discoloration of leaf margins and tips, followed by necrosis and defoliation. Major veins remain green and laterals turn yellow Late stage chlorosis in younger leaves, turning to bright yellow color in interveinal areas. Premature leaf fall and die back of growing tip Stunted growth, small leaves, interveinal chlorosis. Leaf margins pucker Interveinal chlorosis, with major veins remaining green. Chlorotic leaves turn yellow or white later and necrotic mature leaves. Symptoms can often be confused with that of Mg deﬁciency Interveinal chlorosis in younger leaves, youngest leaves becoming totally chlorotic Interveinal chlorosis in younger leaves, necrosis on leaf tips and margins Stunted growth, necrosis and interveinal chlorosis. Necrotic lesions are seen on main vein

Cited by De Waard (1969), Nybe and Nair (1986)

De Waard (1969), Nybe and Nair (1986) De Waard (1969), Nybe and Nair (1986) De Waard (1969), Nybe and Nair (1987a)

De Waard (1969), Nybe and Nair (1987a)

Chin et al. (1993), Nybe and Nair (1987a)

Chin et al. (1993), Nybe and Nair (1987c) Chin et al. (1993), Nybe and Nair (1987b)

Chin et al. (1993), Nybe and Nair (1987b) Chin et al. (1993), Nybe and Nair (1987b) Chin et al. (1993), Nybe and Nair (1987c)

294

K. P. P. NAIR

laterite soil growing Panniyur-1. Adzemi et al. (1993) report the highest accumulation of both Ca and Mg in stem (11.5 and 5.98 g vine21, respectively) with a total accumulation of 54.5 and 36.4 kg ha21, respectively. Among the micronutrients, Zinc is the most important followed by Mo and B (De Waard, 1969). Zn deﬁciency is beginning to be seen in many tropical countries in a very signiﬁcant manner adversely affecting crop yields. However, to date, corrective measures still bank on routine soil testing with DTPA as the most commonly used extractant. That such an approach does not lead to satisfactory results uniformly is shown through the work of this author (Nair, 2002), which will be discussed subsequently in this section. A combination of Zn, Mo, and B, along with N, P, and K was shown to result in the highest yield of two locally developed varieties, namely, Subhakara and Sreekara at the Indian Institute of Spices Research experimental farm (IISR, 1977) where Zn, B, and Mo were used at a ratio of 5:2:1. Spraying a solution of Zn SO4 (0.5%) reduced spike (inﬂorescence which eventually develops into pepper berries) shedding by 48.4% (Geetha and Nair, 1990). Organic manuring is a very commonly adopted practice in pepper production in India and part of Asia. This can either be through the use of fresh vegetative matter or through the use of “burnt earth” (Harden and White, 1934; Bergman, 1940). In the former category, freshly chopped materials, such as, leaves, stems, etc., from a number of trees are used. The trees generally used are E. indica, Garuga pinnata, Grevillea robusta, etc. (Sivakumar and Wahid, 1994). Organic manures are widely used in pepper production in Sarawak (Malaysia), which include soybean cake, guano, prawn, and ﬁsh refuse (Purseglove et al., 1981). Holes are dug and about 85 g of manure are placed about 20 cm away from the main stem of the vine. This is done once in two months. Some farmers dig trenches all around the vines and put large quantities of manure once in two months. Of late, sterilized animal meat and bone meal admixture, fortiﬁed with potassium, are gaining popularity as an organic manure. Another method of organic manuring is through the use of “burnt earth”. For this, forests are ﬁrst cleared and on top of the heaped vegetation soil from vacant patches is spread and the heap is set ﬁre to and the ﬁre is on in a slow burning fashion for two to three weeks. At the end of this period, both the wood ash and burnt soil are thoroughly mixed and applied at a rate of 18 kg vine21 year21. The application of burnt ash has many beneﬁcial effects in improving the physical, chemical, and biological properties of soil. Compared to the acidic soil, having a pH of 4– 5, the burnt earth has a pH of 7 –8 and acts as a good soil ameliorant in correcting pH. There is extensive literature on this method (Harden and White, 1934; Bergman, 1940; Huitema, 1941; De Waard, 1969). Owing to environmental hazards, especially because of large-scale burning and the emission of smoke, the Government of Sarawak banned the practice in 1940. Figures 2– 4 represent K, Mg, and Zn deﬁciency, respectively.

AGRONOMY AND ECONOMY OF BLACK PEPPER

Figure 2

K deﬁciency symptom.

Figure 3 Mg deﬁciency symptom.

Figure 4

Zinc deﬁciency symptom.

295

296

K. P. P. NAIR

IV. THE ROLE OF “THE NUTRIENT BUFFER POWER CONCEPT” IN PEPPER NUTRITION Historically, soil testing has been used to quantify availability of essential plant nutrients to ﬁeld grown crops. However, contemporary soil tests are based on philosophies and procedures developed several decades ago without signiﬁcant changes in their general approach. For a soil test to be accurate, one needs to clearly understand the physico-chemico-physiologic processes at the soil –root interface, and an understanding of soils and plant root systems as polycationic systems is as essential. It is this knowledge that leads to sound prescriptive soil management practices in nutrient availability vis-a`-vis fertilizer application, because, of all the factors that govern sustainability of crop production, the nutrient factor is the most important, and it is also the least resilient to management. This section of the review will focus on the buffering of plant nutrients, with speciﬁc reference to Zn, and discuss experimental results that relate to pepper nutrition.

A. THE BUFFER POWER AND EFFECT AVAILABILITY

ON

NUTRIENT

Before being able to understand the signiﬁcance of the nutrient buffer power concept on plant nutrient availability, certain basic concepts must be addressed and the following review starts with this rationale.

B. BASIC CONCEPTS In any nutrient management approach that is sound and reproducible, one must start with a basic understanding of the chemical environment of plant roots. When we consider this, the ﬁrst term that we come across is the “soil solution” because the plant root is bathed in it and is most affected by its chemical properties. The Soil Science Society of America (1965) deﬁnes soil solution as “the aqueous liquid phase of the soil and its solutes consisting of ions dissociated from the surfaces of the soil particles and of other soluble materials”. Adams (1971) has given a simple deﬁnition: “The soil solution is the aqueous component of a soil at ﬁeld moisture content”. Perhaps it is important to emphasize here that much contemporary soil testing has considered a soil extract as synonymous with the soil solution. Since soil extraction is supposed to simulate plant root extraction, it is pertinent to consider the chemical environment of the root, though brieﬂy, from this angle. It is worth noting that the chemical environment of roots in natural soil systems is so obviously complex that both soil scientists and plant physiologists have been unable to provide a precise deﬁnition. If this complex

AGRONOMY AND ECONOMY OF BLACK PEPPER

297

chemical system is to be accurately quantiﬁed, thermodynamic principles will need to be used to evaluate experimental data. Even then, the limitations are obvious, as in the case of K where the thermodynamic investigations are quite often inapplicable under ﬁeld conditions. This is because, although a quasiequilibrium in K exchange can be achieved in the laboratory, these conditions are seldom, if ever, attained under ﬁeld conditions (Sparks, 1987). Agricultural soils are, for the most part in a state of disequilibrium owing to both fertilizer input and nutrient uptake by plant root. It thus appears that a universal and accurate deﬁnition of a root’s chemical environment awaits the proper application of thermodynamics for the root’s ambient solution (Adams, 1974) or even kinetics, as in the case of K (Sparks, 1987), where thermodynamics have been found to be inadequate. Soil extractions with different extractants provide a second approach in deﬁning the root’s chemical environment. This approach has been particularly successful in understanding cases like P insolubility, soil acidity, and K ﬁxation. However, this approach also fails to deﬁne precisely the root’s chemical environment. Though this approach also suffers deﬁciencies, such as the extractants removing arbitrary and undetermined amounts of solid-phase electrolytes and iron (or the extractants causing precipitation of salts or ions from the soil solution) and the soil –plant interrelationship deﬁned in terms of solid-phase component of the soil, even though the solid phase is essentially inert except as it maintains thermodynamic equilibria with the solution phase (Adams, 1974), the latter part could be researched more to understand how the solid phase – solution phase equilibria can be interpreted to give a newer meaning to quantifying nutrient availability. It is in this context that the role of the plant nutrient “buffer power” assumes crucial importance. The close, almost linear relationship in a low concentration range of þ þ 2 22 , 0.5 mM for NO2 3 –N, NH4 – N, K , H2PO4 , and HPO4 , which has been established by numerous solution culture experiments can be quantitively described by the equation: U ¼ 2p r aC r ;

ð1Þ

where U is the uptake of a 1 m root segment, r is the root radius, Cr is the concentration of the ion at the root surface, and a is the root absorbing power (Mengel, 1985). The metabolic rate of the root determines its absorbing power. A high root absorbing power would imply that a relatively high proportion of nutrient ions coming in contact with the root surface is absorbed and vice versa. The nutrient ion concentration at the root surface (Cr) depends on a since a high root absorbing power tends to decrease Cr; it also depends on the rate of movement from bulk soil towards the root (Mengel, 1985). Diffusion and/or mass ﬂow control this movement. But it is now established that nearly 95% of this movement for nutrients such as P, K, and Zn (among heavy metals) and, possibly

298

K. P. P. NAIR

NHþ 4 , is by way of diffusion. When root uptake of an ion species is less than its movement towards it, accumulation of the ion species on the root surface is bound to occur, as has been shown to be the case with Ca2þ where mass ﬂow contributes to this accumulation (Barber, 1995). The diffusive path for ions such as P and K, which plant roots take up at high rates but which are in low concentration in the soil solution near the root, is the concentration gradient. In a sense, the effective diffusion coefﬁcient which quantiﬁes the diffusive path and the buffer power are analogous because the diffusive ﬂux across the root surface is integrally related to the nutrient buffer power. This has been shown to be true in the case of P where a highly signiﬁcant positive correlation between the two was found to exist in 33 soil samples obtained from experimental sites in the USA and Canada (Kovar and Barber, 1988). However, in a routine laboratory set up, it is far easier to measure the buffer power than the effective diffusion coefﬁcient and this review will further focus on the question of how buffer power can be quantiﬁed without recourse to cumbersome analytical techniques and how its integration into routine soil test data will considerably improve predictability of nutrient uptake.

C. MEASURING THE NUTRIENT BUFFER POWER AND ITS IMPORTANCE IN AFFECTING NUTRIENT CONCENTRATION ON ROOT SURFACES The ability to predict the mobility of dissolved chemicals, such as fertilizers, in the soil is of considerable value in managing fertilizer applications. Soil testing in its essence aims to achieve this. While modeling transport and retention of ions from thermodynamic (Selim, 1992), kinetic (Sparks, 1989), and mechanistic (Barber, 1984) angles could be very informative, the importance of translating this information into practically feasible procedures in crop production calls for an understanding not only of the basic concepts but of their intelligent application as well. In a dynamic state of plant growth, the concentration of any nutrient on the root surface is nearly impossible to measure since both the nutrient in the plant tissue and the root absorbing power, which directly affects it, change quickly due to root metabolic processes. The inability of even mechanical mathematical models to accurately predict nutrient inﬂux rate has recently been highlighted (Lu and Miller, 1994). Hence, if an effective soil testing procedure is to be devised for a nutrient, which is an alternative to deﬁning the plant root’s chemical environment, one must resolve the problems of quantifying the nutrient concentration on the root surface indirectly, even if it is impossible to resolve it directly, for the reasons mentioned above. Using Fick’s ﬁrst law, F ¼ 2DðdC=dxÞ

ð2Þ

where F is the ﬂux, dC/dx is the concentration gradient across a particular section, and D is the diffusion coefﬁcient, Nye (1979) has suggested that the formula

AGRONOMY AND ECONOMY OF BLACK PEPPER

299

can be applied to both ions and molecules. The negative sign for D implies net movement from a high to a low concentration. Although for molecules in simple systems like dilute solutions D may be nearly constant over a range of concentrations, for ions in complex systems like soils and clays, D will usually depend on the concentration of the ion, and on that of other ions as well (Nye 1979). Nye (1979) has further suggested that though Fick’s ﬁrst law may be derived from thermodynamic principles in ideal systems, in a complex medium such as the soil, the above equation may be regarded as giving an operational deﬁnition of the diffusion coefﬁcient. Thus, Nye (1979) deﬁnes the diffusion coefﬁcient as D ¼ D1 u f1 ðdC1 =dCÞ þ DE

ð3Þ

where D1 is the diffusion coefﬁcient of the solute in free solution, u is the fraction of the soil volume occupied by solution and gives the cross-section for diffusion, f1 is an impedance factor, C1 is the concentration of solute in the soil solution, and DE is an excess term which is zero when the ions or molecules on the solid have no surface mobility, but represents their extra contribution to the diffusion coefﬁcient when they are mobile. DE can generally be neglected since only in rare instances will it play any role in diffusion of plant nutrient ions in soil (Mengel, 1985). From the point of view of nutrient availability, dC1/dC, which represents the concentration gradient, assumes crucial importance as we shall see below. The term dC1/dC, where C1, is the concentration of the nutrient ion in the soil solution and C is the concentration of the same ion species in the entire soil mass, assumes considerable signiﬁcance in lending a practical meaning to nutrient availability. If we ascribe the term “capacity” or “quantity” to C and “intensity” to C1, we have in this term an integral relationship between two parameters that may crucially affect nutrient availability. Since the concentration gradient of the depletion proﬁle of the nutrient in the zone of nutrient uptake depends on the concentration of the ion species in the entire soil mass (represented by “capacity” or “quantity”) in relation to the rate at which this is lowered on the plant root surface by uptake (represented by “intensity”), it could be argued that a quantitative relationship between the two should represent the rate at which nutrient depletion and/or replenishment in the rooting zone should occur (Nair, 1984a). This relationship has been functionally quantiﬁed by Nair and Mengel (1984b) for P in eight widely differing Central European soils (Table V) and the term dC1/dC has been referred to as the “nutrient buffer power”. Nair and Mengel (1984b) used electroultraﬁltration to quantify C1 while using an incubation and extraction technique to quantify C. For P, C was found to closely approximate isotopically exchangeable P (Keerthisinghe and Mengel, 1979), but in the experiments conducted by Nair and Mengel (1984b) it was estimated by the extraction of incubated soil with an extractant which was a mixture consisting of 0.1 M Ca lactateþ 0.1 M Ca acetateþ 0.3 M acetic acid at pH 4.1. The extractant exchanges adsorbed phosphate and dissolves Ca phosphates except

300

K. P. P. NAIR

Table V Comparison of P Buffer Power of Eight Widely Differing Central European Soils (Determined by Two Different Techniques) Regression Soil Benzheimer Hof Hungen Oldenburg B6 Wolfersheim Obertshausen Oldenburg B3 Klein–Linden Gruningen

r

(1)

(2)

(1)

(2)

Y ¼ 18.8x þ 7.94 Y ¼ 38.2x 2 1.03 Y ¼ 49.8x þ 0.52 Y ¼ 70.3x þ 0.03 Y ¼ 70.5x þ 2.66 Y ¼ 73.6x þ 2.07 Y ¼ 75.0x þ 0.38 Y ¼ 75.4x þ 0.89

Y ¼ 0.23x þ 8.98 Y ¼ 0.25x þ 4.32 Y ¼ 0.26x þ 0.72 Y ¼ 0.27x þ 0.11 Y ¼ 0.30x þ 2.89 Y ¼ 0.31x þ 0.61 Y ¼ 0.32x þ 1.81 Y ¼ 0.36x þ 3.62

0.912 0.967 0.994 0.998 0.966 0.994 0.999 1.000

0.995 0.997 0.999 0.983 0.998 0.997 0.991 0.996

Note: The b values in the regression functions represent the P buffer power of each soil. In regression function (1) (after Nair and Mengel, 1984) Y ¼ CAL-P (Schu¨ller’s method) and in regression function (2) (after Nair, 1992) Y ¼ the author’s method. x in both refers to electroultraﬁltrable P. Note the very high r values in all the cases. The soils are arranged in their sequential increase in P buffer power.

apatites; the method known as the “CAL method”, developed by Schu¨ller (1969), is now widely used in Central Europe. In the case of Kþ and NHþ 4 – N, C denotes the concentration of exchangeable, and to some extent non-exchangeable, fractions (Mengel, 1985). Since very low concentrations in the range of 2.0 mM may be attained on the root surface for both P and K (Claassen and Barber, 1976; Claassen et al., 1981; Hendriks et al., 1981), Nair and Mengel (1984b) had to use electroultraﬁltration to quantify C1. Thus, the nutrient depletion around the roots which is caused by the diffusive ﬂux of the nutrient towards the root surface is related to both the quantity and the intensity parameters, and a quantiﬁable relationship between both represents the buffer power speciﬁc to the nutrient and the soil. A growing root will at ﬁrst encounter a relatively high concentration of P, which is in the range of the concentration of the bulk soil solution (Nair and Mengel, 1984b). As uptake continues, depletion will occur at the root surface. This depletion proﬁle gets ﬂatter with enhanced nutrient uptake (Lewis and Quirk, 1967; Claassen et al., 1981; Hendriks et al., 1981). But it is the capacity of the soil to replenish this depletion which ensures a supply of nutrient ions to the plant root without greatly depressing its average concentration on the root surface. It is the nutrient’s buffer power that decides these depletion and/or replenishment rates. A soil with a high P buffer power implies that the P absorbed from the soil solution is rapidly replenished. In such a case, P concentration at the root surface decreases only slowly and mean P concentration at the root surface remains relatively high. In soils with a low P buffer power, the reverse is true, and mean P concentration at

AGRONOMY AND ECONOMY OF BLACK PEPPER

301

the root surface is rapidly diminished and remains relatively low. This has been proved experimentally for P (Nair and Mengel, 1984b; Nair, 1992). This phenomenon holds true for Zn2þ (Nair, 1984c) K (Nair et al., 1997) and NHþ 4 –N as well (Mengel, 1985).

D. BACKGROUND INFORMATION ON THE IMPORTANCE OF MEASURING Zn BUFFER POWER There is a great paucity of published material on the effect of buffer power on availability of heavy metals. Plants obtain most of their fertilizer Zn from reaction products and not applied sources as such, implying that any source of Zn added to soil has to necessarily conform to a chain reaction involving adsorptive, desorptive, and resorptive processes that govern the maintenance of an equilibrium between adequate Zn concentration in the soil solution nearest to the zone of Zn depletion on the one hand and plant uptake on the other. The Zn buffer power deﬁnes this. As Zn concentration in soil solution is normally very low, the supply to plant roots by mass ﬂow can only account for a very small fraction of plant demand. For instance, with a transpiration coefﬁcient of 300 l kg21 dry matter and a corresponding Zn concentration of 1027 M in the soil solution, approximately 2 mg of Zn can be supplied by mass ﬂow against a demand of 10 – 30 mg Zn kg21 dry weight of plant tissue. In calcareous soils, as the Zn concentration is of a much lower order of approximately 1028 M, the supply by mass ﬂow could be very much lower (Marschner, 1994) indicating that mass ﬂow can only contribute very negligibly to meet plant needs of Zn. Hence, Zn movement to the plant root surface is principally by diffusion and is essentially conﬁned to a zone around the plant root which hardly extends beyond the root hair cylinder (Marschner, 1994). In a review on the mechanism of Zn uptake, Marschner (1994) indicated that ﬂow culture experiments with various species showed adequate ranges of Zn concentration in the range of 6 £ 1028 –8 £ 1026 M, which are concentrations greater than those that would be expected in the solution of most soils. He further pointed out that although work using chelate-buffered solutions has indicated adequate Zn concentrations between 10210 and 10211 M, extremely low adequate Zn concentrations required a concomitant excess of about 100 mM Zn chelate as buffer at the plasma membrane of the root cells. This implies a need for an unlimited Zn pool for replenishment of Zn2þ at the plasma membrane. When plants grow in soil, it is impossible to expect a Zn buffer of this size to exist, and free Zn2þ and chelated-Zn concentrations will be at least threefold lower. Hence, critical deﬁciency or sufﬁciency concentrations obtained through research employing chelate-buffered solutions cannot be applied to soil-grown plants.

302

K. P. P. NAIR

Most of the work on Zn availability to plants is based on chemical extractions, among which DTPA extraction is the most frequently used. The DTPA extraction quantiﬁes a labile fraction of soil Zn comprising water soluble, exchangeable, adsorbed, chelated and some occluded Zn. The critical soil level of DTPA extractable Zn can vary from 0.3 to 1.4 mg kg21 soil which equates to about 900– 4200 g ha21 of Zn in heavy soils and about 600– 2800 g ha21 of Zn in light soils in the plough layer (0 – 20 cm). The crop requirements, on the other hand, are quite small, in the range of 100 –300 g ha21 for a total dry matter production of about 10 tons ha21 (Marschner, 1994). The inadequacy of DTPA extraction to reﬂect plant Zn demand shows that other important factors, such as replenishment of soil solution Zn (Nair, 1984c), mobility, and transport to the root surface (Wilkinson et al., 1968; Nair et al., 1984), and also the activity of the roots themselves (Wilkinson et al., 1968; Marschner, 1994) are involved. Since the Zn buffer power is intricately involved in all the three factors, the focus of this review is mainly on that attribute. As early as three decades ago it was suggested that colloidal Zn was released by some speciﬁc process associated with root activity (Wilkinson et al., 1968). Conditions in the rhizosphere and particularly root-induced changes markedly affect Zn availability. A difference in rhizosphere pH of as much as 2, higher or lower compared to bulk soil, can be expected to occur as a result of imbalance in ionic uptake. For instance, any acidifying fertilizer such (NH4)2SO4 can result in a net excretion of Hþ ions and others, such as, 2 NH4NO3 can result in a net excretion of HCO2 ions. Additionally, 3 or OH secretion of organic acids and enhanced CO2 production, as well, will affect rhizosphere pH, and all of the above-mentioned changes will markedly affect Zn availability. However, the scope of this review is conﬁned to the kinetic/ dynamic aspects of the changes occurring in the rooting zone mirrored in the Zn buffer power rather than changes in soil reaction in the rhizosphere per se on Zn availability. The distribution of Zn between the solid and solution phases can be described by the buffer power. The availability of soil Zn to the plant depends on the initial Zn concentration, Zn buffer power and effective diffusion coefﬁcient (Barber, 1984). The Langmuir equation gives the relation between B and C1 as C1 1 C þ 1 ¼ aB ðx=mÞ B where C1 is the Zn concentration in the soil solution, x/m is the amount of Zn adsorbed per unit of soil, B, the adsorption maximum and a is a constant related to the soil’s bonding energy for Zn. A straight line is obtained when C1/(x/m) is plotted against C1 with a slope of 1/B and intercept of 1/aB. The inverse of C1/(x/m) ¼ b, the Zn buffer power, where C1 and x/m are both expressed in volume units (Barber, 1984). Using this approach, Shuman (1975) estimated

AGRONOMY AND ECONOMY OF BLACK PEPPER

303

the buffer power values varying from 5 to 100 for four soils representing different major physiographic regions of Gerogia. Based on the diffusion model of Drew et al. (1969), Nair (1984) has argued that the c in the equation U ¼ 2paaC t (Drew et al., 1969), where, U is the quantity of Zn absorbed per centimeter root length, a is the root radius in cm, a is the root absorbing power, C 2 is the average Zn concentration on the root surface, and t is the duration of the absorption period, in fact, represents an indirect measure of the Zn buffer power. As we already know, the bulk of Zn uptake is by diffusion (Wilkinson et al., 1968; Elgawahary et al., 1970; Barber, 1984). This diffusive process will maintain a concentration gradient in the root zone. This concentration gradient will directly affect Zn uptake because of its effect on the average Zn concentration on the root surface. The Zn buffer power will affect this concentration gradient, because the rate of Zn depletion and/or replenishment is mirrored by it. In a sense, the effective diffusion coefﬁcient and the buffer power are analogous to each other for nutrients which are principally absorbed by the plant root through diffusive process (Nair, 1989). Hence the crucial question to examine would be the role of Zn buffer power in inﬂuencing Zn availability for plant uptake.

E. QUANTIFYING Zn BUFFER POWER GROWING SOILS

OF

PEPPER

Nair (1984c) has used electroultraﬁltration to quantify Zn intensity in measuring Zn buffer power of European soils. As the procedure is very highly sophisticated and due to its non-availability in most of the developing countries, a simple adsorption – desorption equlibrium technique was developed by the author (Nair, 2002) to quantify the Zn buffer power of pepper growing soils. For this, 200 g of representative soil samples from three locations in the state of Kerala, where pepper is extensively grown, were incubated at 60% of the maximum water holding capacity with graded rates of Zn over a fortnight, by maintaining the water regime to constancy. At the end of the period the soil samples were extracted with 0.01 M CaCl2 over a 24 h period and the extract tested for Zn using atomic absorption spectrometry. This represented Zn intensity. Separate extractions were made with DTPA, which represented Zn quantity. Data in Table VI indicate that the Zn buffer power values of the experimental soils varied widely. Data in Table VI brings to light two very important facts. First, the very highly signiﬁcant correlation coefﬁcients prove that the technique employed allows a precise determination of the Zn buffer power of the soils. Second, the soils varied widely in their Zn buffer power. It is this fact that has to be critically viewed against the existing Zn fertilizer recommendation, emanating from

304

K. P. P. NAIR Table VI Zn Buffer Power of Pepper Growing Soils

Soil Peruvannamuzhi Thamarasseri Ambalavayal

r-value

b-value

0.8337 ppp 0.9304 ppp 0.9604 ppp

0.7824 1.5786 3.0358

Note: ppp Signiﬁcant at a conﬁdence level of 0.1%; b values represent the Zn buffer power.

routine soil tests using the universally employed DTPA extraction to understand the cruciality of Zn buffer power for precise formulations of Zn fertilizer applications for pepper. Data in Table VII clearly indicate that integration of the Zn buffer power has improved the relationship between DTPA routine test versus both Zn concentration and Zn uptake. More remarkably, the overall relationship between DTPA test and dry matter production not only improved with Zn buffer power integration in the computations, but turned from negative to positive. This clearly shows that the commonly employed DTPA test could only be sitespeciﬁc, but on a larger scale it is the Zn buffer power that determines plant performance. The experiment had only monitored dry matter production in the pepper plant without taking it to berry formation as it takes about 3– 5 years for berry formation. But, the data in Table VII clearly substantiate the cruciality of the buffer power concept. Data in Tables VI– VIII have to be viewed in conjunction to precisely understand the cruciality of the buffer power concept. Pepper is the economic mainstay of the state of Kerala and of late widespread Zn deﬁciency has been observed in the state, and the onset of the dreaded disease Foot Rot due to Phytophthora fungi is attributed, to a great extent, to this deﬁciency in the soils. The scientists at the Indian Institute of Spices Research

Table VII Correlation Coefﬁcients (r) for the Inter-relationship between Routine DTPA Soil Test versus Zn Concentration, Zn Uptake and Dry Matter Production without (1) and with (2) Zn Buffer Power Integration in Pepper Details Zn concentration versus DTPA test Zn uptake versus DTPA test Dry matter production versus DTPA test Note: ppp signiﬁcant at a conﬁdence level of 0.1%.

1 0.884 ppp 0.782 2 0.745

2 0.924 ppp 0.862 ppp 0.777 ppp

AGRONOMY AND ECONOMY OF BLACK PEPPER

305

Table VIII Pepper Yield from Farmers’ Fields (kg vine21) Weighted Against Zn Buffer Power Yield Region

Targeted

Actual

Deviation (%)

Peruvannamuzhi Thamaraserri

0.241 0.490

0.401 0.487

þ 66 þ 0.6

Note: Target weighting was done against the highest yield obtained from the Ambalavayal region. Note the remarkable closeness between targeted and actual yields in Thamarasseri region. Peruvannamuzhi soil is an atypical pepper soil.

have made a “blanket recommendation” of soil application of 23 kg ZnSO4 ha21 when the DTPA extractable Zn is less than 1.6 ppm. In terms of monetary input, this would translate to an equivalent of US $25 ha21 in terms of farmer investment. As the soils are variously Zn buffered (Table VI) a single “blanket dose”, such as the one that is made, is totally unscientiﬁc, because, where the farmer needs to apply only lesser quantities of Zn, he would end up applying more, if he were to go by the routine recommendation based on the DTPA extraction. To elaborate, the soils from the Ambalavayal region would only require 25% of Zn needed in the soils of the Peruvannamuzhi region as the former has a Zn buffer power nearly fourfold more compared to the latter. In the case of the soils from the Thamarasseri region the quantity required would only be 50%. In fact, the scientists at the Institute have made a gross underestimation of the Zn supplying power of the soils in the state as their results are all based on the soils of the Peruvannamuzhi region alone where the Institute has its experimental farm. The soils here are obviously Zn impoverished, but that is not the case in the entire state of Kerala. This also shows that, more often than not, recommendations from experimental stations, when extrapolated on a large scale can run into problems, especially when such recommendations are based on routine soil testing procedures. These results (Nair, 2002) substantiate the earlier ones (Nair, 2003) on the importance of measuring Zn buffer; rather than the routine DTPA extractable Zn alone, to make precise Zn fertilizer recommendations for wheat in Central Asia (Turkey) where farmers have been advised to apply as much as 100 kg ZnSO4 ha21 as a “blanket recommendation”, without obtaining any tangible and consistent response in wheat yield. Hence, the economic importance of the buffer power concept, as against the routine soil testing procedures, needs hardly any underscoring. Nair et al. (1997) have conclusively demonstrated the importance of the buffer power concept in the case of K nutrition for another very important perennial crop and spice, cardamom (Elettaria Cardamomum M.) in the state of Kerala. Both pepper and cardamom are the staple spices of the state.

306

K. P. P. NAIR

V. ESTABLISHING A PEPPER PLANTATION A. THE INDIAN EXPERIENCE From a commercial view point, establishing a good pepper plantation is of much signiﬁcance. Unlike in countries such as Thailand, Malaysia, Brazil, etc. where the life span of a pepper plantation is only 10 –15 years, at the end of which replanting is done, in India a pepper plantation can last for more than a quarter century. The prime reason for the shorter duration of the pepper plantations in other countries, referred to above, unlike in India, is because, in the former, pepper is trailed on non-living standards (support), while in India pepper is always trailed on live support, such as E. indica. The starting point to establish a good pepper plantation is the establishment of good planting material. With the exception of India, most other countries use the orthotropic (upward climbing shoot) shoot as the planting material. This is the best planting material as they result in vigorous plants and bear fruiting laterals right from the base itself and yield a lot earlier. In India, however, the nonavailability restricts the use of orthotropic shoots widely. In earlier times, pepper planters used either runner shoots or climbing orthotropic shoots to coincide with the onset of the southwest monsoon in India. These days, pre-rooted cuttings from runner shoots are used instead. When hanging shoots are used as planting material, they result in the formation weak plants. Quite often, pruned material — pepper is pruned ﬁve to six months after planting, and once again, after a year — is also used as planting material. These are pruned stem cuttings. Since pepper is only propagated through the type of planting material described above, world-wide, there is a huge demand running into millions for ready to use planting material. Production of pre-rooted cuttings in light polyethylene bags (PVC bags) is the surest way of mass producing material once the demand for planting material escalated in India. Their production is as follows. Runner shoots from high yielding and healthy plants are kept coiled at the base of wooden pegs ﬁxed at the bottom of the plant so as to prevent the growing shoots spreading on the soil and striking roots. These runner shoots can be separated from the mother plant in winter (January – February) by snapping them using scissors and are disinfected by dipping them in a fungicide solution, such as Oxychloride or Bordeaux mixture for a minute, later surface dried and cut into three nooded strips after clipping the leaves, planted either in raised soil beds or PVC bags ﬁlled with a mixture of soil, sand and some farmyard manure, which will provide enough nutrients in a growth substrate. Application of growth regulators, such as IBA, by dipping the cuttings in the solution has been found to enhance root proliferation (Pillai et al., 1982; Suparman and Zaubin, 1988). Care must be taken to minimize fungal infection arising from the nursery bed.

AGRONOMY AND ECONOMY OF BLACK PEPPER

307

In Sri Lanka growing planting material on split bamboo led to the production of good planting material on a large scale. Popularly known as the “bamboo technique” (Bavappa and Gurusinghae, 1978) the method consists of ﬁrst digging a trench (60 £ 40 cm2) of convenient length, ﬁlling it with a rooting medium — preferably forest soil, sand and farmyard manure in a 1:1:1 ratio mixture and then split bamboo poles (1.25 – 1.5 m long) or PVC pipes — both of about 10 cm in diameter — are sunk into the trenches at 458 using a strong middle support. Rooted cutting, one each per bamboo is planted in the trench and allowed to trail onto the bamboo or PVC poles. The lower portion of the pole is ﬁlled with a rooting medium (of weathered coir dust and farmyard manure in a 1:1 ratio) and as the vine grows it is tied to the support. Vines are irrigated daily and to stimulate rapid growth a nutrient solution consisting of N, P, K, and Mg (1 kg urea, 0.75 kg super phosphate, 0.5 kg muriate of potash (MOP) and 0.25 kg magnesium sulfate all mixed in 250 l of water) can be applied at the rate of 0.25 l per vine every fortnight. It takes three to four months for the vine to reach the top of the pole, and when it does, the terminal bud is nipped off and the stem is crushed about three nodes above the base to activate the axillary buds and after about 10 days each vine is cut at the crushed point and this is ready for transplanting into PVC bags, ﬁlled with the rooting mixture as detailed above. Care should be taken, while planting in PVC bags, not to injure the roots and not let the axil be below the rooting medium. The PVC bags are placed in a cool place and covered with a light PVC sheet (200 gage) to retain high internal humidity. Buds start to develop in about three weeks time and then the PVC bags can be shifted to a place with shade. Regenerated shoots from the stumps can be trained as before, which permits the production of a continuous stream of rooted cuttings as planting material. Four harvests are possible in 1 year, with a multiplication ratio of 1 –40. A proliﬁc root system develops and ﬁeld establishment is very good. In Sarawak (Malaysia), on average, 54 rooted cuttings could be obtained (Ghawas and Miswan, 1984) and the method has been successfully ﬁeld tested in the experimental farm of the Indian Institute of Spices Research. Figure 5 shows the “bamboo technique” developed in Sri Lanka for vegetative pepper propagation. Pepper can be planted on ﬂat surfaces and slopy land. If it is the latter, slopes facing south should be avoided. On slopy land it is advisable to plant along the contour lines to arrest top soil loss due to run off water during monsoon. An important feature of pepper plantations in India is that the vines are trailed on living standards, mainly E. indica or E. lithosperma (a thorny tree and the trailing vines have strong footholds on the thorns) Glyricidia, G. robusta (Silky Oak), G. pinnata, Ailanthus spp. etc. In homestead plantations, coconut and arecanut trees act as the main crop in addition to being the standards. In the southern state of Karnataka, pepper is grown along with coffee, where coffee is the main crop, and the forest trees act as the standards. Glyricidia is the most common standard in Sri Lanka and Malayasia. Pepper planting should precede the onset of monsoon. Three noded 2 –3 rooted cuttings, prepared as explained

308

Figure 5

K. P. P. NAIR

Split-bamboo technique developed in Sri Lanka for vegetative pepper propagation.

earlier, are planted in shallow pits on the northern side of the standard. If the cuttings are unrooted, 4 –5 could be planted. At least one node should be below the soil for the establishment of roots. To accelerate growth, pits can be ﬁlled with top soil, farmyard manure, rock phosphate, and neem cake mixture (Pillai, 1992). Plant density is around 1970 – 2000 ha21 (George, 1982) and can go up to 5000 ha21 when nutrient and carbon stress can be observed (Reddy et al., 1992). In order to support the climbing vine on the standard, it should be tied as frequently as necessary, and for tying jute thread is used. Pruning of the terminal part, though not extensively practiced in India, encourages growth and bearing (Kurien and Nair, 1988). When a pepper plantation is established, starting from the beginning of the year, speciﬁc operations are to be followed. In the ﬁrst two months, starting from January, when harvesting is completed, pruning the hanging shoots, tying vines, and mulching the basin is done. In the following two months, diseased vines are removed and the pits could be ﬁlled with top soil, 1 kg lime and 5 kg farmyard manure mixture. Southwest monsoon establishes in May –June and with this a mixture of 5 kg farmyard manure, 250 g rock phosphate and 1 kg neem cake should be applied. In June – July, a Bordeaux mixture (1%) and drenching the pits with copper oxychloride (0.2%) is done. The most damaging “Pollu Beetle” (Longitarsus nigripennis), a deadly insect that leaves the berries hollow (from which the name “Pollu” in the regional language Malayalam of Kerala state is derived, meaning hollow) can be controlled by spraying endosulfan (1.5 ml l21). Application of 50 g urea, 120 g MOP should also be done. In September – October 50 g urea, 120 g of MOP and 50 g Mg SO4 are applied. Additionally, a Bordeaux mixture spray (1%) is also done. Tying of vines should be continued. In November –December phytosanitary measures,

AGRONOMY AND ECONOMY OF BLACK PEPPER

309

such as removal of infected plants, drenching with oxychloride (0.2%), mulching, and shading the young vines must be done. It is uncommon to irrigate a pepper crop, except in Thailand where it is grown as an irrigated crop. Nevertheless, pepper plants can seriously suffer due to moisture stress in the summer months, especially in India, when daytime temperatures from April to May can soar to 358C plus. A ﬁeld study conducted at the Pepper Research Station in Panniyur, Kerala state, during 8 years (1988 – 96) indicated that irrigation at IW/CPE ratio of 0.25 increased pepper yield, as much as 90%, the positive effect being most manifest in Panniyur-1 and Karimunda (Cultivar) both released from this station (Satheesan et al., 1997a). An interesting aspect of pepper cultivation is the possibility of growing other crops, in addition to pepper, in the same ﬁeld or plot of land. This practice is more in common when it is the case of homestead farming where income from the accompanying crop adds to the proﬁt from the main pepper crop. However, this is not a general practice on large pepper plantations, because raising an additional crop makes demands of its own, in terms of nutrient input, pest management, etc. It is important to note that companion crops to pepper must best be raised in the pre-bearing stage of pepper, which at best can extend from 3 to 5 years (at the most). This is because, when fully grown, pepper canopy could completely shade the accompanying crop and the latter would hardly yield. In an experiment at the Indian Institute of Spices Research (Sadanandan, 1994) reported that an accompanying crop of banana (cv. Mysore poovan) fetched an additional income of approximately US $460 ha21. Pillai et al. (1987) have obtained comparable results. The recent glut in the pepper market and slump in prices have prompted many farmers to look to companion crops to generate additional income. Simultaneously, the falling prices of coconut have prompted coconut farmers to use the coconut trees as a live standard to raise pepper. In this connection, an interesting experiment conducted at the Central Plantation Crops Research Institute (CPCRI) at Kasargode, Kerala State, which has a mandate for research, primarily on coconut, merits mention. Pepper vines (Panniyur-1) were trailed on 60-year-old coconut palms and when the vines reached a height of 4 –5 m, further growth was restricted to enable farmers from climbing the palms to pluck the nuts. On average, 2 kg dry pepper vine21 was obtained. Companion cropping with arecanut has also been tried. An arecanut — pepper combination gave 3832 kg ha21 of dry dehusked arecanut and 1418 kg ha21 dry pepper (Nair and Gopalasundaram, 1993), from 1000 vines ha21. Companion cropping is common in Brazil as well, where an array of companion crops, such as rubber, cocoa, orange, lemon, clove, etc., are extensively tried. Though growing mixed companion crops in pepper gardens is a practice in many situations, it needs to be clearly understood that when inter speciﬁc crops are grown in association, the biological implications could be diverse. This diversity might originate not only in nutrient absorption pattern,

310

K. P. P. NAIR

Figure 6 A ﬁeld view of an Indian Pepper Plantation.

moisture stress relationship, etc., but also can extend to other areas, such as susceptibility/resistance to speciﬁc pests and diseases. These are the areas that deserve thorough scientiﬁc scrutiny before value judgment can be made on the suitability or otherwise of growing companion crops with the main crop of pepper. Figure 6 shows a ﬁeld view of an Indian Pepper Plantation.

B. THE INDONESIAN EXPERIENCE Within Asia, next to India, Indonesia is the major pepper grower. Pepper was the ﬁrst spice Indonesia traded with Europe through Persia and Arabia and prior to the second world war, 80% of world production was controlled by Indonesia. It was the Japanese occupation during the war that left many pepper plantations uncared for with the resulting decline in production. Compared to the pre-war period when the country could boast of more than 20 million vines, after the war it came down to just about a lakh of vines. The most interesting characteristic of pepper cultivation in Indonesia is that most of the pepper gardens are owned by small farmers unlike in India where large plantations could be found. Presently the main producing areas are Lampung and Bangka with Lampung in the lead. During the last quarter of the past century, a perceptible decline in area in Lampung has taken place with shifts to Bangka, both east and west Kalimantan and Sulawesi. In Bangka, pepper was ﬁrst grown in Muntok and the Muntok white pepper had a good world market. In Indonesia, pepper is grown in various types of soil, such as andosol, grumsol, latosol, podsol, regusol, etc. Well-drained alluvium rich in humus, pH above 5.8, is ideal for pepper (Zaubin and Robbert, 1979). In general, pepper soils in Indonesia are poor in fertility, texture, and structure and low in organic matter. The two main

AGRONOMY AND ECONOMY OF BLACK PEPPER

311

pepper-growing areas, Bangka and west Kalimantan have poor, reddish brown, and sandy soil. Soils in Lampung are reddish brown latosols and contain more organic matter than the soils of Bangka. In Indonesia, the planting material is cuttings from orthotropic shoots. Normally cuttings with ﬁve to seven internodes are used, with two to four nodes buried in the soil. Cuttings are planted at an angle of 35– 458 with reference to the standard. For the ﬁrst three to four weeks, the cuttings must be shaded from the sun. Cuttings are planted in pits of 0.6 £ 0.8 £ 0.6 m size. Rooted cuttings are also used. Precautions explained earlier must also be adhered to in order to protect the cuttings. The procedures used in the case of pre-rooted cuttings, except that they have a better ability to withstand heat exposure from sunshine, are the same as above. Such planting materials are selected from the nursery and it should be ensured that only well-rooted and wellgrown cuttings in PVC bags are transplanted in the main ﬁeld. Unlike in India, Sri Lanka, and the Philippians, in Indonesia the types of standards used are both living and non-living. In the former category, we have Glyricidia, Dadap, Kapok, etc., while in the latter category we have hard timber poles, made of iron, wood, or concrete poles, etc. Experience shows that concrete poles result in poor growth and productivity. In Lampung small farmers use living standards while non-living ones are used in Bangka, east and west Kalimantan. There is a gradual shift to living standards as non-living ones, of late have become more expensive. Pepper plantations are usually owned by small farmers, with holdings ranging between 0.2 and 1.5 ha. On average, the size is 0.65 ha. In general, pepper cultivation is intensive in Bangka and south Sumatra, while in Lampung it is extensive, where pepper is planted under live standards and input rates, in terms of fertilizer, pest management, etc., are meager. The Indonesian pepper industry is mostly owned by small farmers and planting pepper on a plantation scale did not last long in Java. The nature of pepper cultivation and the degree of involvement of labor and capital have played a very important role in determining the mode of the pepper farming system in Indonesia. However, it is a tribute to the pepper industry in Indonesia that the IPC ofﬁce is headquartered in Jakarta. Figure 7 shows a ﬁeld view of an Indonesian Pepper Plantation.

VI. PEPPER PESTS AND THEIR CONTROL The pepper plant is attacked by several diseases, caused by fungi, bacteria, virus and mycoplasma, insects, and nematodes. Nutritional disorders aggravate the impact of the pests, be they caused by any one of the above or soil and/or environmentally triggered. However, it is important to make a distinction between, for example, a deadly disease that is caused by a fungus, such as Phytophthora spp.,

312

K. P. P. NAIR

Figure 7 A ﬁeld view of an Indonesian Pepper plantation.

from which speciﬁc symptoms in the plant will originate, and a nutrient deﬁciency in the soil that may lead to an imbalance of the speciﬁc nutrient in the plant that will aggravate the symptoms of the disease subsequently. It is important to recognize that of all the diseases a pepper plant may suffer from, that caused by the Phytophthora fungus is the most devastating. Hence, this review will primarily focus on the “Foot Rot” disease caused by Phytophthora, which is the major disease in India, Malaysia, and Indonesia causing very serious damage to pepper production (Holliday and Mowat, 1963; Kueh and Sim, 1992b; Manohara et al., 1992; Sarma et al., 1992c). In India, although wilt disease was reported as the major cause of pepper crop loss, Phytophthora, as its causal agent, was ﬁrst reported only in 1966 by Samraj and Jose (1966). There are, in fact, 17 diseases that cause loss in yield, but among these Phytophthora Foot Rot and “slow decline”, which were originally referred to as “quick wilt” and “slow wilt”, respectively, cause the most damage (Nambiar and Sarma, 1977; Nambiar, 1978; Das and Cheeran, 1986). Phytophthora capsici is the causative fungus. The infection occurs aerially as well as through soil and the severity of infection depends on the site of infection (Mammootty 1978; Anadaraj and Sarma, 1995). When it is a foliar infection one to many dark spots appear, which have characteristic ﬁmbriate margins, which advance and later coalesce leading to defoliation even before the lesions spread to the entire lamina. The base of the vines can also be infected. On tender shoots, the fungus profusely sporulates forming a white covering and when the infection reaches the stem abrupt wilting of the entire plant takes place. Spikes, when infected, lead to the formation of blackened berries. When the infection is below the soil on roots, rotting and degeneration starts leading to yellowing, defoliation, and drying up of the plants. Feeder root infection reaches the collar through main roots, which causes the characteristic Foot Rot; hence the name (Anandaraj et al., 1994).

AGRONOMY AND ECONOMY OF BLACK PEPPER

313

Since Foot Rot results in sudden wilting followed by death of the plant the disease was earlier referred to as “wilt” or “quick wilt” and the disease causing fungus was identiﬁed as Phytophthora palmivora var piperis by Samraj and Jose (1966) and as P. palmivora by Holliday and Movat (1963). In India, though pepper is traditionally grown in the Western Ghats in the states of Kerala, Karnataka, and Tamil Nadu, it has been introduced to non-traditional pepper growing areas, such as the Andhra state and also the north-eastern states and the incidence of Foot Rot has been reported from these places as well (Sarkar et al., 1985). Up to 30% crop loss has been reported (Samraj and Jose,1966; Nambiar and Sarma, 1977). The onset and spread of Foot Rot has been found to be closely correlated to a wet and cloudy atmosphere prevalent during the southwest monsoon period in the state of Kerala. A combination of factors, such as daily rainfall of 1.6 –2.3 cm, ambient temperature ranging from 22.7 to 29.68C, relative humidity of 81– 99% and daily sunshine duration of 2.8– 3.5 h favor the spread of the disease (Ramachandran et al., 1988c, 1990). Inasmuch as infection through soil substrate is concerned, the prime source of infection is contaminated soil where new planting is done. The inoculum can survive in the soil for upto 19 months in the absence of a host plant (Kueh and Khew, 1982). The pathogen is normally concentrated to a depth of 30 cm (Ramachandran et al., 1986) and the severity of concentration diminishes as one moves deeper and away from the infected vines. Management of Foot Rot is, perhaps, the most challenging aspect of pepper cultivation. There are various aspects to the control measures, such as cultural, phytosanitary, biological, and chemical. The most important among the cultural measures refer to maintenance of good drainage. Stagnant water leads to anaerobic conditions triggering germination of pathogen propagules. A decrease in the production of phenol oxidase, phytoalexin and ﬁxed nitrogen combined with suppressed mycorrhiza leads to host plant susceptibility in anaerobic conditions (Drew and Lynch, 1980). Good drainage blocks the buildup of P. capsici. Phytosanitation assumes an important role in controlling the disease. Since the infected plants serve as the foci of infection (Zadocks and Van den Bosch, 1994), their removal is most crucial in controlling the disease. The initial occurrence and spread of the disease is non-random and tends to cluster around the previously infected plants (Anandaraj, 1997) and this is the reason why such infected plants have to be removed forthwith. Further, as pepper grows on living standards, they develop a canopy of their own generating a microclimate different from that of the ambient situation, with high humidity and low temperature, which is ideal for P. capsici to multiply fast and infect. Branches of the living standards should be regularly lopped, especially during the rainy season, which facilitates the penetration of direct sunlight that leads to the heating up of the soil and thereby reducing the humid surroundings, thus acting as a check for the multiplication of the pathogen. The loppings could be used as surface mulch, which while adding to the organic matter of the substrate soil can also check weed growth. Since the prime source of

314

K. P. P. NAIR

the initial inoculum of P. capsici in pepper plantations is the contaminated soil, infection of the foliage can occur through soil splashed onto tender shoots trailing on the ground and from these shoots onto other parts of the vines through rain splashes (Ramachandran et al., 1990). In order to arrest soil splashes, legume or grass cover mulch is ideal (Ramachandran et al., 1991; Sarma et al., 1992c). However, studies in Malaysia (Ahmed, 1993) indicate that pepper grew better in clean-weeded plots unlike in those with a cover crop of Desmodium trifolium. The Indian experience shows (Anandaraj, 1997) that after the cessation of the southwest monsoon it would be ideal to clean up the pepper plots of the existing weeds and rake up the soil which was seen to check the spread of the pathogen. Chemical control of Phytophthora has revolved around the use of systemic fungicides, such as metalaxyl formulations like Ridomil granules or Ridomilziram (Ramachandran and Sarma, 1985; Ramachandran, 1990; Ramachandran et al., 1988b, 1991; Kueh and Sim 1992a; Sarma et al., 1992c; Kueh et al., 1993) and/or the classical Bordeaux mixture 1% spray and also pasting the collar with Bordeaux mixture and drenching the basin/trench around the vine with oxychloride (Malebennur et al., 1991; Nair and Sasikumaran, 1991; Ramachandran et al., 1991; Lokesh and Gangadarappa, 1995). Owing to the high cost of chemical control, it is not very popular with small farmers. Further, Coffey (1991) has expressed doubts about the efﬁcacy of chemical control of the soil-borne diseases caused by Phytophthora. In addition to the above measures of control, there is also the possibility of the use of organic amendments and biological control. Inasmuch as the addition of organic amendments to the soil are concerned to control the Phytophthora spp., the objective is to raise the growth of saprophytes (Kueh and Sim, 1992a) and it has been observed that the enhanced activity of saprophytes checks the growth of P. capsici dramatically (Anandaraj, 1997). Oil cakes, such as that obtained after expelling oil from neem seed, groundnut, coconut, etc., can beneﬁcially be used to control the growth of the disease causing fungus (Sadanandan et al., 1992; Nair et al., 1993). In addition, even chicken manure has been found to be useful. Applications of organic amendments such as the ones cited above will have a twofold beneﬁcial effect: ﬁrst it enhances the nutrition of the vines, and second, it encourages the growth of saprophytes that check the population of P. capsici. Both complement each other in the overall control of the fungal growth. A recent development in the control of Phytophthora is the employment of biological organisms, primarily from a fungal background. Mention must be made of vesicular arbuscular mycorrhizae (VAM), Trichoderma and Gliocladium (Manjunath and Bagyaraj, 1982; Ramesh, 1982; Datta, 1984; Anandaraj and Sarma, 1994b; Anandaraj and Peter, 1996). Soil-borne pathogens are amenable to biological control (Cook and Baker, 1983). Since P. capsici is a soilborne fungus and the prime source of contamination is infected soil, growth of antagonistic fungi would check the build up of the pathogenic fungus. Interesting work on isolation and mass multiplication of inexpensive carrier media for ﬁeld

AGRONOMY AND ECONOMY OF BLACK PEPPER

315

application of microorganisms having biological control properties has been carried out (Anandaraj and Sarma, 1995; Sarma et al., 1996). An inexpensive carrier medium for the culture of these beneﬁcial organisms is water contained in mature coconuts that is thrown away after shelling. Both Trichoderma and Gliocladium have been found to grow well in this medium (Anadaraj and Sarma, 1997). Interest in VAM has been mainly focused on its ability to solubilize native phosphorus. Its use in the control of pepper diseases was started when it was ﬁrst observed that VAM suppressed the Foot Rot caused by Phytophthora in citrus (Davis et al., 1978; Davis and Menge, 1981). Soil incorporation of VAM can also be done along with Azotobacter and Azospirillum, which, besides controlling the pathogen have beneﬁcial growth effects (Govindan and Chandy, 1985; Bopaiah and Khader, 1989) The positive effect of VAM is through the alteration of the nature of root exudates and the rhizosphere microﬂora — which is generally termed the “mycorrhizosphere effect” — and there is extensive research to substantiate this view (Dehne, 1982; Graham, 1982, 1988; Linderman, 1988). The positive effects of VAM have not only enhancing plant growth, but also, in suppressing disease causing pathogens has been reported by Ewald (1991) and Graham and Egel (1988). In view of the overall positive effect of VAM, it has been recommended that VAM be incorporated into the soil right from the nursery stage and onwards so that the beneﬁcial effects of enhancing growth and suppression of pathogenic effects is obtained right from the early stages (Sarma et al., 1996). Field experience shows that strict phytosanitation in the nursery is a very important pre-requisite in subsequently installing a disease-free pepper plantation. Next in importance to diseases caused by pathogenic fungi are the diseases caused by nematodes. Crop productivity is seriously hampered by plant parasitic nematodes and pepper is no exception. Of the 15,000 nematode species described so far, 2200 are plant parasitic. Quite often nematode damage goes unnoticed. Since nematode infestation starts at the root surface, pepper vines so infected will subsequently suffer from a number of secondary complications and quite often these symptoms can be mistaken for nutritional or other physiological disorders. When susceptible crops are continuously grown, in the year after, they become a good breeding ground for disease causing nematodes. A compilation of plant parasitic nematodes associated with pepper in the major pepper growing countries listed 48 species, that belong to 29 genera (Sundararaju et al., 1979), while 54 species that belong to 30 genera have been listed by Ramana and Eapen (1998). In India, 17 genera of nematodes associated with pepper were listed from the two primary pepper growing states, Kerala and Karnataka (Sundararaju et al., 1980), while in Indonesia 14 genera have been listed (Bridge, 1978; Mustika and Zainuddin, 1978). Among the nematode species, Meloidogyne incognita and Radopholus similis are the most important that cause the most severe damage. The former is commonly known as the “Root Knot Nematode”,

316

K. P. P. NAIR

while the latter as “The Burrowing Nematode”. On a global scale the former is the most devastating. The root knot nematodes have a specialized and complex relationship with the host plant and they are sedentary obligate endoparasites. Their infestation of the vine results in the formation of elongated swellings on the thick primary roots due to multiple infections and typical knots or galls on either secondary or ﬁbrous roots because of hypertrophy and hyperplasia of the infected tissues. The name root knot is derived because of these typical knots on the roots. In thick primary roots a number of adult females with egg masses localize deep below the epidermis and the entire length of the root turns into a gall and appears smooth with infrequent swellings (Mohandas and Ramana, 1987). Since nematodes feed on vascular tissues, a disruption in the arrangement and continuity of the vascular bundles leads to impaired movement of both water and nutrients and as a consequence, the plant is vitially affected. When infestation is severe, a large amount of root mass is lost due to eventual decay of the galled roots, which in turn very adversely affect the entire vine (Mohandas and Ramana, 1991; Siti Hajijah, 1993). When root-knot infestation takes place, yellowing of the foliage occurs resulting in stunted growth and eventual decline of the vine. Dense yellowing of interveinal areas with the deep green veins prominently visible (Ramana, 1992) can often be mistaken for nutritional deﬁciency. When vines are infested with M. incognita, certain impaired physiological reactions, such as lowered absorption and translocation of P, K, Zn, Mn, Cu, Ca, and Mg and their accumulation in leaves have been observed (Ferraz et al., 1988) and also reduction in total chlorophyll content of leaves (Ferraz et al., 1989). These changes lead to stunted growth of the vines. When vines were inoculated with M. incognita inoculum, a high concentration of total phenols without expression of any resistance to the pest was observed (Ferraz et al., 1984). Additionally, several changes in amino acids, organic acids and sugars were also observed in vines infested with the nematode (Freire and Bridge, 1985b). The burrowing nematode (R. similis) is an obligate and migratory endoparasite, is extensively found in both tropical and subtropical regions of the world, is a serious pest of pepper (Holdeman, 1986), and has a wild range of hosts (about 370 plant species). The existence of R. similis was ﬁrst reported in the state of Kerala, India, on the banana host by Nair et al. (1966). The nematode feeds on cortical tissues and produces elongated dark brown necrotic lesions on the roots at the site of infection. The nematode pushes through the cell wall of each cell after draining its contents and this burrowing phenomenon results in the formation of tunnels in the root tissues. The nematode derives its name from this trait of burrowing. When infestation is severe, many lesions coalesce and encircle the root cortex and because of this damage to the cortical cells, the root portion distal to these lesions gradually disintegrates. The vines tend to produce new roots which, in turn, get infected leading to the formation of a bunch

AGRONOMY AND ECONOMY OF BLACK PEPPER

317

of decayed root mass (Mohandas and Ramana, 1987). Yellowing of leaves, defoliation and generally stunted growth result from infestation. Another important disease, popularly known as the “Slow Decline Disease”, caused by a combination of the two nematodes, M. incognita and R. similis, and the fungus Fusarium spp., has been the cause for widespread destruction of pepper in several countries. The disease is also known as “yellows”. It was the cause of major pepper devastation in the Bangka Islands of Indonesia in the ﬁfties (Christie, 1959), and also in Guayana (Biessar, 1969), India (Nambiar and Sarma, 1977), Malaysia (Kueh, 1979, 1990), Brazil (Sharma and Loof, 1974; Ichinohe, 1975), and Thailand (Sher et al., 1969; Bridge, 1978). The disease was ﬁrst observed on the islands of Bangka (Van der Vecht, 1950) in Indonesia and later spread to the islands of Belantung. The disease caused losses of up to 30% annually (Sitepu and Kasim, 1991). The disease is widely prevalent in Sarawak (Malaysia). The total life span of the vines can be reduced to 8– 10 years, as against the normal 25 –30 years (Varughese and Anusar, 1992). In Cambodia, where pepper is predominantly found in the Kampot region, the disease was responsible for bringing down the vine population from 2.5 million in 1942 to 0.5 million in 1953 (Hubert, 1957). In the predominantly pepper growing state of Kerala in India, there is no precise estimation of crop loss, but Menon (1949) reported a 10% loss of vines. It has been observed that plants infested by root-knot nematodes are more susceptible to the fungus Phytophthora (Winoto, 1972). Fool-proof management of nematode infestation is an elusive target. It is important to recognize that the primary source of infestation is nurseries. Solarization and fumigation are effective in controlling nematode infestation in nurseries. Additionally, incorporation of biocontrol agents, such as VAM, to solarized soil has been found to be an encouraging method for the control of nematode infestation (Ananadaraj and Sarma, 1994a,b; Sarma et al., 1996). There are different angles to the control regime. For instance, in plantations it has been observed that chemicals released during decomposition of organic manure, for instance, Azadirachtin from neem cake and Ricin from castor cake, are toxic to nematodes (Stirling, 1991). Mulching has been found to have a positive effect on nematode control (Wahid, 1976; Ichinoe, 1980, 1985). Hubert (1957) obtained very encouraging results in nematode control using Eupatorium mulch. Of the various aspects of developing resistance/tolerance to nematodes, the most reliable is the development of plant resistance. But, to date, there are no varieties or cultivars that are totally resistant to nematode attack. Obviously, the focus of pepper breeding has not been on developing absolute nematode resistant varieties. In fact, polygenic-horizontal resistance is more important than vertical-monogenic resistance (Fassuliotis and Bhatt, 1982). Existence of physiological races or pathotypes in nematodes, in particular Meloidogyne spp., is another factor that must be keenly focused on in any breeding program in pepper for nematode resistance.

318

K. P. P. NAIR

Currently the efforts focus on developing tolerance to nematode infestation rather than developing absolute resistance. Also, efforts are made to screen the currently available cultivars for their reaction to nematodes. For instance, in Sarawak (Malaysia), Kueh (1986) found cv. Uthirancotta to be the most susceptible cultivar, while cultivars such as Balancotta, Belantung, Cheriakaniakkadan, Jambi, and Kalluvally are less susceptible to root-knot nematodes under ﬁeld conditions. Mustika (1990, 1991) found cv. Kuching to be tolerant to M. incognita and R. similis, compared with Kalluvallay, Jambi, and Cunuk. In India, none of the cultivars tested were found to be resistant to M. incognita and R. similis compared with Kalluvally, Jambi, and Cunuk. In India, none of the cultivars tested were found to be resistant to M. incognita (Jacob and Kuriyan 1979a; Koshy and Sundararaju, 1979; Ramana and Mohandas, 1986) or R. similis (Venkitesan and Setty, 1978; Ramana et al., 1987b). However, it has been found that among the Indian cultivars, Pournami, which is a selection from the germplasm collection at the Indian Institute of Spices Research, is tolerant to M. incognita. This has been supported by ﬁeld evidence that shows that where this cultivar is cultivated, only a smaller population of nematodes could thrive. Another possibility in breeding resistance to nematode infestation is the development of resistance from related wild species into the cultivated ones. The related wild species, P. colubrinum and P. aduncum, are highly resistant to M. incognita (Ramana and Mohandas, 1986; Paulus et al., 1993). Additionally, P. hymenophyllum and P. attenuatum showed remarkable resistance to R. similis (Venkitesan and Setty, 1978). Though P. colubrinum has been found to be immune to R. similis (Ramana et al., 1994), crosses between P. nigrum and P. colubrinum have been unsuccessful until now. The solution for cross transfer of resistance from resistant to cultivated species might be found through the biotechnological route, but, as of now, it still is a very challenging route. The gene transfer mechanism in pepper is still very poorly understood. In addition to the above, there is also the possibility of employing biological organisms, as in the case of controlling diseases, to control nematode infestation. Sewell (1965) deﬁnes biological control as “the induced or natural, direct or indirect limitations of a harmful organism or its effects by another organism or group of organisms”. Since a variety of microorganisms inhabit the soil, some of which are either predatory or antagonistic to plant parasitic nematodes, such a course of action could be taken. However, the question of nematode control biologically is still a very open one. Nevertheless, this approach could be a part of the overall integrated pest management in pepper. As of now, the research efforts in the area of biocontrol is, indeed, very sparse. A few examples, given below, illustrate the efﬁciency of biocontrol. In the area of biocontrol, fungal agents are the most important. Among them, an opportunistic fungus, namely, Paecilomyces lilacinus, is an important one that

AGRONOMY AND ECONOMY OF BLACK PEPPER

319

has great potential in the control of root knot and cyst nematodes. The fungus, on contact with the egg masses of the nematodes, colonizes and grows rapidly. The chitinolytic enzymes produced by the fungus help penetration of the eggs and cause the suppression of nematodes. The efﬁcacy of the fungus in controlling root knot nematodes has been reported by Jatala (1986). Inoculation of the fungus into rooted cuttings or seedlings of pepper signiﬁcantly reduced the growth of the root knot nematode, which in turn, reduced damage to the roots (Freire and Bridge, 1985d; Ramana, 1994; Sosamma and Koshy, 1995). However, it was less efﬁcient against the control of the borrowing nematode (Geetha, 1991; Ramana, 1994). The next fungus that is effective against nematodes is Trichoderma spp. It has been found effective in controlling root knot nematode (Eapen and Ramana, 1996), in addition to suppressing the P. capsici, as detailed earlier in this review. Verticillium chlamydosporium is another fungus for control of root knot and cyst nematodes (Kerry, 1990). The fungus colonizes the rhizosphere, infects adult females and egg masses and reduces nematode multiplication by inhibiting the hatching of eggs. Parasitization of root knot nematode eggs by the fungus in Brazil has been reported (Freire and Bridge, 1985d). The occurrence of the fungus in association with Trophotylenchulus piperis in pepper plantations was ﬁrst reported in India by Sreeja et al. (1996). In an in vitro test, the fungus was found to colonize the egg masses of M. incognita and reduce hatching by almost 50% within ﬁve days of inoculation (Sreeja et al.,1996). In addition to these, the VAM also has a suppressive effect on nematode population through the indirect effect of enhancing plant vigor, which has been referred to earlier in this review. Occurrence of VAM fungi on pepper roots was ﬁrst reported in India by Manjunath and Bagyaraj (1982). Bacteria come after fungus as nematicidal agents. Within the bacteria is Pasteuria penetrans, which is an obligate parasite of some nematodes. It infects the nematodes by direct penetration through the cuticle by germinating spores sticking to the body surface of the nematodes. It has been found to suppress both M. incognita and R. similis. In addition, Bacillus pumilis, B. macerans and B. circulans also suppress M. incognita (Sheela et al., 1993). Eapen et al. (1997) reported the inhibitory effect of ﬂuorescent pseudomonas, such as Pseudomonas ﬂuorescens, on M. incognita. It is important to realize that research in the area of biocontrol of nematodes is in its infant stage and a number of studies are done in in vitro conditions and careful evaluation has to be done before large-scale ﬁeld recommendations can be made that are applicable to pepper plantations. The experience, so far, has been promising, though in a limited sense, and much more work needs to be done to probe the area further. Perhaps the most effective means of controlling nematode infestation is through the use of chemicals. Nematicides — chemicals used to kill nematodes — consist of two groups, namely, systemic nematicides and soil fumigants. Of the two, the latter is more effective as nematodes generally inhabit the soil and once

320

K. P. P. NAIR

the nematicide is applied to the soil it is absorbed by the roots of the host plant and translocated to shoots where they act as systemic inhibitors entering into biochemical reactions within the host plant cell. Despite the effectiveness of nematicides, their widespread use is not in vogue because of high costs and the environmental adverse fallout from their use. Various nematicides, such as Phenamiphos (Nambiar and Sarma, 1980), Aldicarbsulphone (Venkitesan and Setty, 1978), Phorate and DBCP (Venkitesan and Charles, 1980) were found to be effective. Among the fumigants, DD, methyl bromide and ethylene di bromide have been found to be effective. With the advent of organic farming and greater awareness among consumers of products of chemical-free agriculture, use of nematicides is being increasingly restricted. Hopefully, with more research to follow, better and safer nematicides will come into the market. Next to the plant pathogenic diseases and plant parasitic nematodes as pepper pests, comes the infestation by insects. Though a number of insect pests attack the pepper plant, the most devastating is the “Pollu” beetle, L. nigripennis Mots, though to a lesser extent others, such as, top shoot borer (Cydia hemidoxa Meyr.), leaf gall thrips (Liothrips karnyi. Bagn.), and scale insects (Lepidosaphes piperis Green and Aspidiotus destructor Sign.) also could be considered as pests of major importance. However, this review will primarily focus on the Pollu beetle. The insect pest derives the name Pollu because of the hollow berry that results from its infestation, the word Pollu in the regional Malayalam language of the state of Kerala meaning hollow. The status and control of major insect pests in India have been reviewed (Devasahayam et al., 1988; Premkumar et al., 1994). The extent of attack by Pollu beetle varies and reports of damage varies from 6 to 21% (Thomas and Menon, 1939), while Rehiman and Nambiar (1967) report damage as much as 30 –40% in the state of Kerala. Survey of one of the major pepper growing districts of Kerala, Kannur district (where incidentally the ﬁrst research station in pepper in the world was established in the town of Panniyur) showed that the single greatest cause for loss in pepper yield was due to infestation by the Pollu beetle amounting to as much as 13% (Prabhakaran, 1994). The beetle measures about 2.5 mm £ 1.5 mm and feed on tender shoots, leaves and spikes. When infestation is severe, leaves and spikes rot and drop down. The grubs do the most damage. They bore into tender berries and empty their contents, and the infested berries, in turn, become chlorotic ﬁrst and then turn black and crumble when pressed. Among the pepper cultivars, wide variation has been observed inasmuch as susceptibility or resistance is concerned. Among the four cultivars in which ﬁeld observations were recorded, Kalluvally and Karimunda were least susceptible to the pest in the northern and southern regions of the state, respectively (Premkumar and Nair, 1988). Inasmuch as control of the insect pest is concerned, both biocontrol and chemical control assume importance in addition to plantation management.

AGRONOMY AND ECONOMY OF BLACK PEPPER

321

Only few natural organisms exist which have a control trait on the Pollu beetle. Among these, an unidentiﬁed entomophagous nematode (Mermithidae family), a predatory spider (Araneae family) and Oecophylla smaragdina Fabr. (Formicidae family) have been indentiﬁed. The extent of parasitization or predation by these on the Pollu beetle is rather limited and their efﬁcacy in controlling the insect pest in the ﬁeld conditions is open to question (Devasahayam and Koya, 1994). Within the ambit of plantation management, application of insecticides is the only viable proposition to effectively control the Pollu beetle. A number of insecticides have been evaluated in the ﬁeld for their efﬁcacy in controlling the pest. On balance, endosulfan spray (0.05%) has been found to be most effective in the state of Kerala (Premkumar et al., 1986; Nandakumar et al., 1987; Premkumar and Nair, 1987a). It must, however, be noted that of late, there has emerged widespread opposition from environmental activists in the state of Kerala to the use of endosulfan because of suspected human health hazards, such as mongolian births, central nervous system (CNS) disorders and even blood cancer. These maladies have been noted in the northern part of the state of Kerala and the insecticide has been currently banned from use. Application of insecticides, like that of fungicides or nematicides, leave pesticide residues in the berries and with increasing health awareness within the country and outside of it, consumers are opting for organically grown pepper. Though biocontrol is a promising avenue, results obtained so far are of a very preliminary nature and no hard and fast recommendations have emerged. A number of byproducts have been used, both in vitro and in situ in the ﬁeld. Among them, mention must be made of neem oil, neem seed extract (Azadirachta indica A. Juss.), nuxvomica (Strychnos nux-vomica L.) and also that of custard apple (Anona squamosa L.) and the success rate has not always been very consistent (Devasahayam and Leela, 1997), though they show potential for further research (Devasahayam and Anandaraj, 1997). As far as Malaysia and Indonesia are concerned, pepper weevil (Lophobaris piperis) is the most serious insect pest. The grubs bore into the nodal region of the climbing and ﬂowering shoots, which result in infested plants. As the infestation advances, the entire aerial portion of the vine wilts and collapses. The infestation of the pest in the lower part of the vine results in the most damage, amounting to a yield loss of around 50% at most and around 5% at the least, in the upper part (Deciyanto and Suprapto, 1992). A number of natural predators have been identiﬁed of which Spathius piperis has been found to be the most effective in controlling the pest both in the laboratory and also in the ﬁeld. As much as 40% control has been reported (Deciyanto and Suprapto, 1992). A number of insecticides have been recommended for the control of pepper weevil. These include the spraying of endosulfan (0.2%), parathion (0.2%), permithrin (0.2%), etc. Removal of affected branches and stems would also help in reducing the level of pest infestation in the ﬁeld (Kueh, 1979; Deciyanto and Suprapto, 1992).

322

K. P. P. NAIR

VII. THE PROCESSING OF BLACK PEPPER ON-FARM The processing of pepper starts from the harvesting operation, which is a long, drawn out process. With the onset of the southwest monsoon in the months of May and June the pepper plants start to ﬂower. The harvest commences in December and spills over into January of the following year when there is ample sunshine and the rains completely cease. Varietal differences and the prevalent climatic pattern decide the harvesting process. In the plains of the state of Kerala, the early maturing varieties reach the harvest stage by November and the late maturing ones by December – January. Harvesting of pepper is important in view of the endproducts, which have a very great industrial signiﬁcance. Govindarajan (1979) has made a detailed and systematic analysis of the harvesting dates against the background of the end-products as detailed in Tables IX and X. Data in Table X give a clear idea how the chemical composition varies in relation to maturity dates.

Table IX Harvest Schedule in Relation to the End Products of Pepper End products

Time to harvest

Black pepper White pepper Canned pepper Dehydrated green pepper Oleoresin Pepper oil Pepper powder

Fully mature and nearly ripe Fully ripe Four to ﬁve month old berries Ten to ﬁfteen days before full maturity Fifteen to twenty days before full maturity Fifteen to twenty days before full maturity Fully mature berries when starch content peaks

Table X Changes in Chemical Composition of Two Important Cultivars in Relation to Maturity Dates Maturity date Months after fruit setting Cultivar End products Volatile oil (%) NVEE (%) Piperine content (%) Starch content (%)

Panniyur-1 3.0 6.4 8.7 1.9 2.5

4.5 7.6 8.8 2.6 3.7

5.5 6.3 8.7 2.7 5.1

Note: Non-Volatile Ether Extract (NVEE).

6.5 2.8 8.1 3.1 10.2

Karimunda 7.5 2.0 7.8 3.5 16.8

3.0 6.8 10.3 1.9 2.6

4.5 10.4 9.7 2.4 4.9

5.5 8.2 8.6 2.4 6.2

6.5 4.4 7.5 3.1 15.3

7.0 3.6 7.4 3.1 15.3

AGRONOMY AND ECONOMY OF BLACK PEPPER

323

Harvesting commences mainly based on the visual appearance of the berries. When a couple of berries in a spike turn orange or red in color harvesting commences. Only the mature spikes are harvested. The immature spikes when dry tend to produce shriveled berries and this in turn will reduce the quality of the produce. However, in Sri Lanka, premature harvesting is done and the produce is generally known as “light pepper” (Purseglove et al., 1981), which has a premium in the market because it contains a higher amount of oleoresin and volatile oils. Harvesting is done manually by climbing up the vine using a bamboo pole and the produce is collected in a jute bag tied to the back of the climber. Only in ﬂat plantations can mechanical ladders attached to trolleys be used for harvesting. In India, pepper plantations are usually on slopy terrain and such harvesting can only be done by manual labor. Once harvesting is completed and the produce brought to the farmyard, postharvesting operations have to begin. Proper post-harvesting handling of the produce is very crucial to maintain the quality of the produce. The ﬁrst step in post-harvest handling is “decorning” or removal of the berries from the stalk. It is synonymous to the threshing of wheat or rice grains. In most of the situations, homestead farming and also large-scale plantations, decorning is still done manually by spreading the produce on a clean ﬂoor in the farmyard and laborers, mainly women, thresh employing their feet to trample the spikes to separate the berries from the stalk. It is not a very hygienic way of doing it, yet, in most Indian situations there seems to be no alternative to it. Mechanical threshers are a rarity, while they are in use in Indonesia and Malaysia. When mechanical threshers are used, the green spikes are slowly fed into the thresher in which a rotating drum with aluminum blades removes the berries from the stalks. Speed of the rotation is adjusted to be neither too slow nor too fast. The spikes are fed between a moving drum with aluminum plates with suitable gaps and due to the mild stretching action the berries are separated from the stalk. A thresher powered by a 3 hp motor can handle 1.5 tons per hour. Before decorning it should be ensured that the spikes are properly washed clean of all adhering dirt. Once the berries are threshed, they should be graded using a mesh and normally three grades are in vogue, namely, . 4.25 mm (large), 3.25 –4.25 mm (medium) and , 3.25 mm (small). The ﬁrst released hybrid Panniyur-1 belongs to the large size. After grading, the berries are blanched and this is done in boiling water by dipping the berries carried in bamboo baskets in boiling water contained in troughs. This is done to clean the produce of all adhering dust particles and any other extraneous particles, such as bird excreta. The produce is then dried to a uniform moisture content (8 –10%) in open sunlight for three days consecutively. While in India, blanching is done in boiling water for 1 min, in Papua New Guinea, Indonesia and Micronesia it is done at 828C for 2 min (Pruthi, 1992). The entire operation of washing, blanching and grading can be mechanized by spraying boiling water onto pepper fruits as they move in a conveyor belt ﬁtted with mechanical brushes and the blanching time adjusted through the speed of

324

K. P. P. NAIR

the conveyor belt. At the end of the conveyor belt the blanched and washed berries pass through sieves of various dimensions to separate “pin head” (good and well sized berries) and light pepper. The graded pepper is later ﬂushed with dry air to remove surface moisture and then transported to the drying yard or to artiﬁcial driers. The phenolase enzyme which imparts the black color to the berries is activated by blanching. It also ruptures the cells and thereby accelerates the escape of moisture from the inner core and simultaneously enhances the black color with the resinoids inside the berry. Therefore, blanched pepper will shine more and dries at a faster rate. The black color which the berries acquire on drying is due to the oxidation of colorless phenolic compounds present in the skin. Polyphenolase (0-diphenol oxidase) present in the fruit wall converts colorless phenolic substrates (3,4-dihydroxy phenyl ethanol glycoside) present in the cells to black polymeric compounds (Variyar et al., 1988).

A. SUN DRYING

OF

PEPPER

Sun drying when done reduces the initial moisture content in the berries from around 65% to within 10%. Most of the pepper growing countries practice sun drying. To check the development of mould, the berry heap, which is spread on a clean ﬂoor for drying, is periodically turned over. On average, sun drying leads to a recovery of 29 – 38%. Non-uniformity and contamination by microorganisms are the major disadvantages of sun drying. Pepper is dried on different surfaces, such as bamboo mats, cement ﬂoors or black low density polyethylene sheets. It has been observed that the black surface of the sheet, due to absorption and retention of heat results in faster drying than when other surfaces are used (Krishnamoorthy and Zachariah, 1992). In the Waynad district, in the state of Kerala, a special type of drying is practiced where after two days of sun drying on bamboo mats or a cement ﬂoor, the dried berries are collected in jute sacks and stacked upright for two days in rooms. During this process of storing a fermentation process sets in which imparts a uniform black color to the berries leading to enhanced ﬂavor. After two days of storage the berries are, once again, spread on the cement ﬂoor or bamboo mats and dried in the conventional manner. It is for this special type of drying that the pepper of this region is highly regarded for its unique quality of color and ﬂavor.

B. SOLAR DRYING

OF

PEPPER

Solar driers have been developed to dry agricultural products in India (Kachru and Gupta, 1993) at the Central Institute for Agricultural Engineering in India and in Germany at the Institute of Agricultural Engineering for the Tropics and

AGRONOMY AND ECONOMY OF BLACK PEPPER

325

Table XI Chemical Composition of Dried Pepper (in Percentage) Chemical constituent

Composition

Moisture Total nitrogen Volatile ether extract Non-volatile ether extract Alcohol extract Starch (acid hydrolysis) Crude ﬁber Crude piperine Piperine Total ash Acid insoluble ash

8.7– 14.0 1.5– 2.6 0.3– 4.2 3.9– 11.5 4.4– 12.0 28.0–49.0 8.7– 18.0 2.8– 9.0 1.7– 7.4 3.6– 5.7 0.03–0.55

After Pruthi (1993).

Subtropics, University of Hohenheim (Esper and Muhlbauer, 1996). In the solar drier the change occurs mainly in the ﬁnal drying phase, whereas sun drying requires several days to reach the desired moisture level. Solar drying accelerates the drying process leading to a considerable reduction in drying time. Additionally, the maintenance of a constant temperature throughout the drying process ensures a clear product free of microbial contamination or any other extraneous matter. Drying of berries can also be done using mechanical driers. Different types are in use, such as the “Copra Drier” which is used to dry coconut shell, husk etc., a convection drier, which is a forced draft drier, a cascade type drier using indirect heating by kerosene or gas, etc. After drying, recovery normally does not exceed 70%. Dry recovery varies from 29 to 38% depending on the variety. Table XI shows the composition of dried pepper.

C. GARBLING, CLEANING,

AND

FRACTIONATION

Dried pepper is cleaned to get rid of dirt, grit, stone, leaves and any other extraneous matter before packing. Pneumatic separators equipped with magnetic separators are used to remove metallic contaminations, such as iron ﬁlings, nails, etc. For destoning of spices a combination of air-classiﬁcation and vibratory conveying using inclined docks is very efﬁcient (Ramanathan and Rao, 1974). Some of the well established processing houses clean and grade pepper with the help of multiple sieve-cum-air classiﬁer type of machines whereby dust, stalks, pinheads, hollows, immature pepper, red pepper and extra-bold pepper are removed (Pruthi, 1992, 1993). Separated pepper is then washed and dried to make it free of adhering fungus and other extraneous matter. Hence, to obtain good

326

K. P. P. NAIR

quality pepper of international standard the steps involved in processing are: (1) drying, (2) separation of various fractions, (3) size segregation or grading, (4) physical cleaning (washing and drying), and (5) packing. In India, farmers as well as intermediaries bring the produce to the market usually in an ungraded form and generally garbling and grading are done at the exporters’ premises. Most of the exporters clean and grade using the garbling machines which remove dust and chaff and grade the pepper according to densities. However, most of the machines do not remove iron ﬁlings and heavy metals. These are generally hand picked. Usually pepper collected by the large merchants (exporters) is ﬁrst cleaned manually by picking out the contaminants by hand, which is mostly done by women. After this the pepper goes for garbling. A spiral separator has been developed by Madasamy and Gothandapani (1995). This ideally removes all extraneous materials. Traditionally, pepper is manually cleaned by winnowing and then packed into jute bags. In some plantations cleaning is done through a funnel-like arrangement in which dried pepper is fed and cleaned using a blower at the outlet point when the pin heads and light berries are blown away. The Directorate of Marketing and Inspection, Government of India, has classiﬁed pepper into the following grades: 1. 2. 3. 4. 5. 6. 7. 8. 9.

Malabar Garbled (MG Grades 1 and 2) Malabar Ungarbled (MUG Grades 1 and 2) Tellicherry Garbled Black Pepper Special Extra Bold (TGSEB) Tellicherry Garbled Extra Bold (TGEB) Tellicherry Garbled (TG) Garbled Light Pepper (GL Special, GL Grades 1 and 2) Ungarbled Light Pepper (UGL Special, UGL Grades 1 and 2) Pin Heads (PH Grade Special and PH Grade 1) Black Pepper (Non-speciﬁed) (NS Grade X)

Graded pepper fetches a higher price than ungraded, and in the international market TGSEB fetches a premium. Good grabled pepper should have a bulk density of 500 –600 g l21. Light berries should be less than 10% and pin heads (unfertilized berries) less than 4%. When bulk density is low it indicates the presence of more lightweight berries, with consequent less starch content, leading to poor milling quality. Only when the end product has a good aroma and a biting taste will it be considered good. The end product should contain at least 1.5% volatile oil and 3% piperine (Lewis, 1984). Table XII contains data pertaining to the physico-chemical composition of different grades of black pepper. There is another grade of pepper known as “Half Pepper” which is placed in between normal pepper and light pepper. The berries are slightly under mature and therefore contain more of the active principle piperine. Due to immaturity the berries may appear slightly wrinkled and this is ideal for extraction of oleoresin.

AGRONOMY AND ECONOMY OF BLACK PEPPER

327

Table XII Physico-chemical Composition of Different Grades of Black Pepper (in Percentage) Grade

Moisture

Volatile oil

Piperine

NVEE

Starch

Crude ﬁber

MG MUG TGSEB TGEB Light pepper Tellicherry ungarbled Pin heads High range ungarbled

13.0 12.0 10.0 13.0 13.0 12.0 13.0 12.0

3.7 2.8 3.2 2.2 2.9 4.0 0.6 2.6

5.0 5.0 4.9 4.4 4.1 6.3 0.8 4.0

12.3 11.4 10.3 9.1 13.5 13.5 7.1 11.1

39.7 41.8 40.9 39.7 14.6 39.3 11.5 41.8

11.8 12.5 9.2 11.8 27.8 11.0 27.4 10.5

Note: Non-volatile ether extract (NVEE).

The concentration of piperine in oleoresin is more important than the aroma, which is determined by the oil content. Using appropriate sieves the pepper is graded into different sizes. Very small and underdeveloped berries are classiﬁed as “Pin Heads”. The bulk of the pepper belongs to the average size which is known as “Malabar Garbled”. The larger-sized ones are classiﬁed as “Tellicherry Bold”, “Tellicherry Extra Bold”, “Tellicherry Special Extra Bold” and “Giant” (Kachru et al., 1990; Mathew, 1992). In normal pepper starch accounts for 34%, while it is 56.5% in white pepper and 63.2% in decorticated white pepper. Of the 12% water-soluble nitrogen present in the berries non-protein nitrogen constitutes about 82% and of this more than 50% is made up of simple amino acids which act as a nutrient for humans when pepper is consumed as a culinary material. One of the most worrisome aspects of pepper processing is the adulteration practiced by some of the unscrupulous merchants involved in the pepper trade. When such adulterated pepper is exported and the adulteration subsequently detected at the port of entry, it takes not only a heavy toll on the credibility of the country exporting, but, with the recent provisions in the World Trade Organisation (WTO) it can lead to international disputes. One of the most common practices of adultering pepper is mixing it with papaya seeds, which look exactly like pepper (Hartman et al., 1973; Sen and Roy, 1974).

D. PACKAGING

AND

STORING

One of the most important precautions to be taken in packaging pepper is that, since it is hygroscopic in nature, it can easily absorb ambient moisture and when such moisture-laden pepper is stored without care it will lead to the formation of mould and insect infestation will follow as pepper contains a good

328

K. P. P. NAIR

amount of starch. Mould and insect infestation can lead to loss of aroma and caking will result along with hydrolytic rancidity. Whole pepper is generally packed and transported in jute bags (burlap bags) or polyethylene lined double burlap bags. Dried pepper having 10 –11% moisture can be stored without the mould infestation in jute bag with polyethylene lining or in laminated highdensity polyethylene bags (Balasubramanyam et al., 1978). Storing of pepper should be done with great care and moisture should be in the range of 10 –11% before storage. Store houses should be damp proof and should be free of rodent attack, with controlled ventilation and devices for both humidity and temperature control. It is important to fumigate the room before storage. The room walls should be whitewashed with slaked lime and should be exclusively used for storing pepper. Another aspect which will impact the processing of pepper is the microbiological contamination. Investigations conducted in Malaysia have indicated that pepper berries collected from various farms had a total viable count (TVC) between 105 and 1072g sample (Apun et al., 1993). The TVC dropped on drying. TVC was found to be less when the produce was mechanically threshed as compared to when threshing with feet. Washing the spikes before threshing and blanching in boiling water minimizes the microbial load. White pepper is one of the major pepper products consumers all over the world prefer. Retting ripe berries in running water for about 7 –10 days is done to make white pepper. The skin of the berries undergoes bacterial degradation and the softened skin is removed by either decortication or trampling with feet as explained earlier (Pruthi, 1993). An investigation done in Sarawak indicated that pepper berries which have a speciﬁc gravity . 1.12 are the best for conversion to white pepper and are not as good for making ordinary pepper (Anon, 1995).

VIII. AN ACCOUNT OF INDONESIAN PEPPER PROCESSING With Indonesia being a market leader in pepper, it will be informative to review, though brieﬂy, the processing of pepper in that country, more so because the IPC Ofﬁce is located in Jakarta, the capital of the country. In Indonesia most of the pepper is grown by subsistent farmers. To improve pepper processing technology, the Research Institute for Spice and Medicinal Plants has carried out research on equipment design meant for pepper processing which has relevance to on-farm processing. Immediate post-harvest treatment of pepper in Indonesia is quite similar to that practiced in India, but the ﬁnal cleaning, grading and packing of the dried pepper is carried out by exporters. This is mainly because the small-scale farmers do not have the infrastructure to carry out the various aspects of processing. Most of the exporters have a complete set of

AGRONOMY AND ECONOMY OF BLACK PEPPER

329

winnowing and shaking machine (for removal of extraneous materials), washing machine, hot air dryer and screw type separation machine for size-separation. The entire unit can process around 5 tons of black pepper per hour. Grading is normally done according to requirements of overseas buyer and double-lined jute bags are used for packing. Packed bags are stored in “go downs” prior to shipment. In the traditional method of making white pepper, the hygienic and quality aspects are the major problems. The berries are usually soaked in rivers or ponds with poor quality water. Microbial contamination is a major problem, including contamination with waste and the product acquires a swampy odor that is difﬁcult to eliminate later. In addition, soaking in water for a long time results in the loss of volatile oil and deterioration in aroma and ﬂavor. The quality of white pepper in the whole and ground forms is imparted by appearance, aroma and pungency. Blanching or steaming fresh green berries for 10 –15 min for decortication and then removing the rind in a fruit pulping machine has been tried (Purseglove et al., 1981). Decortication by soaking in boiling water for 15 –25 min has been tried and this led to the white pepper having a better aroma than when prepared by the traditional method (Risfaheri and Hidayat, 1996). Prolonged soaking in boiling water will lead to loss of volatile oil and adversely affects the aroma of white pepper as prolonged soaking in boiling water is known to damage the ﬂavor proﬁle of pepper (Pruthi, 1992). Soaking in boiling water softens the pepper fruit wall and makes decortication easy. Pepper oil extraction is an important aspect of pepper processing. It is produced by steam distillation of pepper berries. Pepper oil does not have piperine, which imparts pungency, and so is valued for its aroma in the fragrance industry as well as the ﬂavor industry. Pepper oil is used in high-grade perfumes and the toiletry industry. In the ﬂavor industry, it is primarily used in foodstuffs requiring a high pepper aroma. Volatile oil derived from steam distillation is almost colorless to slightly greenish. Mild in taste with no pungency at all, Indonesia has put much emphasis on pepper oil production and its export. Oleoresin, another by-product of pepper having odor, ﬂavor and pungency, is obtained by extraction of the berries using organic solvents. The organoleptic properties of oleoresin are determined by its volatile oil and piperine contents and their abundance primarily depends on the raw material used for extraction. Ground pepper can be extracted with pure organic solvents, such as acetone, ethanol, dichloroethane, etc. When freshly made, pepper oleoresin is a dark green, viscous heavy liquid with a strong aroma, but, on standing, crystals of piperine appear and the oleoresin requires mixing before use to ensure uniformity and consistency. The components and the quality of the extract depend on the raw material used and research in Indonesia indicates that light black berries are better than pin heads for oleoresin, non-volatile ether extract (NVEE) and piperine (Mapiliandri, 1989). Oleoresin offered for sale by some of the principal producers contains 15 – 20% volatile oil and 35 – 55% piperine.

330

K. P. P. NAIR

Purseglove et al. (1981) have reported that as much as 1 kg of oleoresin can be obtained from 8 kg of black pepper. From a review of research on extraction technology it can be concluded that further research in this area is needed to minimize oil loss and fractionate oil for further blending to produce oil and oleoresin with speciﬁc quality characteristics.

IX. INDUSTRIAL PROCESSING OF BLACK PEPPER Industrial processing of pepper has to be followed subsequent to on-farm processing before the ﬁnal product is ready for human consumption. In the pepper processing plants, the pepper that has earlier gone through one round of on-farm processing, where drying and grading to different standards has been done, goes through further processing. The dried pepper is passed through mechanical sifters for removal of pin heads, vegetable seeds, ﬁne dust, sand etc., all of which are extraneous materials most likely to be present, before winnowing and destoning for removal of stalks, dust, light foreign matter and stones. Multiple sieve-cumair classiﬁer type of machines and gravity separators are used for this purpose. The pepper that is separated is then washed in mechanical washers ﬁtted with brushes to remove mould and dust and this gives a luster to the pepper berries. Then the product is centrifuged to remove adhering water and then dried in electric or diesel-ﬁred indirect dryers. The dried pepper is then sent through spirals for ﬁnal cleaning followed by sterilization either by steam or gamma radiation before being packed. When a pneumatic conveyor-cum-dryer-cumgrader is used, the entire operation of drying and grading can be done in sequential steps, consequently saving a lot of time. There are different grades of pepper. Grading is done on the basis of size and shape of the berries. As per the International Standards Organisation (ISO) the following three grades of pepper exist: 1. Non-processed (NP) 2. Semi-processed (SP) 3. Processed (P) There are permissible limits, set by ISO, for contaminants or light berries, pin heads, broken berries, etc., in the ﬁnal product and its bulk density according to ISO speciﬁcation (ISO/DIS 958-1, 1996). However, individual countries have their own grades and these have been detailed earlier. Sterilization is important to ensure good quality and freedom from microbial contaminants. Several methods of sterilization are available, such as hot air/steam sterilization, extrusion, hydrostatic/pressure sterilization, ozone sterilization, compressed carbon diodixide treatment, irradiation, microwave heating, alcohol treatment, etc. In hot

AGRONOMY AND ECONOMY OF BLACK PEPPER

331

air/steam sterilization the product is preheated to 50– 558C and subsequently sterilized by a combination of indirect heating and direct steam injection. The process substantially reduced microbial contaminants while not adversely affecting the ﬂavor, aroma or oil content (Schneider, 1993). Yet another process of steam sterilization involves exposure to high pressures and temperatures for a pre-determined time period in a series of chambers. The material to be sterilized is placed in the ﬁrst chamber and high pressure exerted and later transferred to subsequent chambers following a similar procedure for speciﬁc time periods and ﬁnally the sterilized material is depressurized in a chamber (Dudek, 1996). There is a modiﬁed steam sterilization process that consists of a steam-jacketed pressurized vessel with three temperature envelopes, namely, 50, 100 or 2008C for different products. Since the system is pressurized, there will be no loss of oil or moisture and the appearance of the product remains unchanged. Steam, compressed air and nitrogen gas may be used depending on the end product desired (Darrington, 1991). A continuous steam sterilization process developed involves subjecting the pepper to a rapid ﬂow of superheated steam during a predetermined time period followed by drying, re-humidiﬁcation and packaging. Both enzyme and microbial activities are signiﬁcantly reduced to low levels and no loss of ﬂavor or aroma or oil was noted (Uijil, 1992). Leife (1992) describes a system of steam sterilization where pepper is subjected to rapid pulses of steam, and the steam condenses on the surface of the berries and contaminants are removed. Steam sterilization is the ideal method in situations where, due to national safety restrictions, both chemical and irradiation methods are prohibited. Yet another method of sterilization is gamma irradiation. This method is particularly useful against infestation by insects. As the process is a cold treatment, loss of volatile oil, ﬂavor and aroma are practically nil (Nair, 1993). A low dose of 1 kGy (kiloGray) is sufﬁciently effective against insect infestation whereas a 10-fold increase is needed to control microbial infestation. The process does not lead to accumulation of any harmful residues unlike fumigation. Depending on the source geometry and the conveyor system used, the size and shape of the containers will be decided. Codex General Standards for irradiated food and recommendations of the International Code of Practices for the operation of irradiation facilities gives the guidelines for irradiation procedures. It is important that Codex Standards are followed to facilitate international trade. Pepper subjected to a radiation level of up to 10 kGy does not lose its aroma and ﬂavor, and there are no signiﬁcant changes in the oil or piperine contents. When the produce is heavily contaminated, irradiation up to 30 kGy is needed. Countries such as USA, France, Austria, Germany, etc., have approved this irradiation procedure for sterilization and a commercial application of the technology is followed. More than 95% of the gamma irradiation facilities operating in the world use Cobalt60. Gamma irradiation facilities are capital intensive and the unit cost of processing is highly sensitive to the scale of operation. To minimize operational costs, facilities need to operate round the clock at the maximum rated throughput

332

K. P. P. NAIR

that the facility is designed for. A gamma irradiator is simple to operate and maintain. Chemiluminescence, thermoluminescence, and free radial dosimetry are employed as routine methods for the detection of irradiated pepper. Other methods in use are colored indicators on the containers which are to be irradiated, viscosity measurements of gelatinized pepper and intensity measurements of reﬂected signals of the electron spin resonance (ESR) spectrum. Two other methods of sterilization are chemical and through the use of microwave. Fumigation with ethylene dibromide to disinfect insect contamination followed by fumigation with ethylene oxide or propylene oxide for eliminating microbial contamination is the most commonly used chemical sterilization process. Moisture content in the berries, temperature, time of contact, and concentration of the fumigating gas decide the effectiveness of the sterilization process. Fumigation results in the reduction of both volatile and nonvolatile oil contents. The main problem with the use of these fumigants is their explosive, toxic, and irritant nature, and as such, unsafe when used without adequate safety measures for the workers who are involved. Additionally, harmful residues are left behind following fumigation. On account of these problems many countries have banned the use of these fumigants. The main advantage with microwave sterilization is that the microbial count is drastically reduced. Microwave heating is done at 2450 MHz and high frequency heating at 27.12 MHz. With an increase in moisture in the samples, microbial efﬁciency is increased, but there are losses of both volatile and non-volatile oils. Both chemical and microwave fumigation are comparable in their efﬁciency. A number of value added pepper products of industrial origin exist. The ﬁrst is dehydrated green pepper. This product is prepared from immature green pepper fruits of appropriate varieties by industrial processing. The fruit should be uniformly sized with pungency, ﬂavor, and color of green pepper. The fruits are blanched in boiling water for a few minutes, the water drained, then cooled and soaked in SO2 solution to ﬁx the green color followed by drying in a cabinet drier at 508C. Upon rehydration, the product will reconstitute to a good quality product possessing the characteristic pungent spicy taste, color, and ﬂavor of green pepper, when one part by mass of dehydrated green pepper is cooked for 20 min in the presence of 10 parts by mass of NaCl solution of 1% concentration. To conform to international standards (ISO/DIS 10621, 1996), the product should have a moisture content of less than 8%. The use of SO2 is a health hazard and efforts are on way at different research institutions to ﬁnd an alternative (Sankarikutty et al., 1994). The next in line among the value added products is canned green pepper. Of late, there has been a surging demand for canned green pepper. The pepper fruits after removal from spikes are washed in running water and kept soaked in water with a Cl concentration of 20 ppm for an hour. They are then covered with 2% hot brine of 0.2% citric acid warmed to 808C, sealed airtight and kept in boiling water for 20 min. After this, they are immediately cooled in a stream of running cold water. A better color was obtained using acetic acid

AGRONOMY AND ECONOMY OF BLACK PEPPER

333

instead of citric acid, but, as per international standards, only citric acid is permitted not exceeding 0.6% by mass of the packing medium and the covering brine must be 1 –2% by mass of edible common salt. For making canned pepper, berries harvested a month prior to the actual date of maturity are the best. Bottled green pepper, dry packed green pepper, and freeze dried pepper are among the other directly edible pepper products. Bottled green pepper is made by ﬁrst despiking fresh green pepper fruits of uniform size and maturity, immediately after harvest, followed by cleaning, washing, and steeping in 20% brine solution containing citric acid. This is then allowed to cure for 3–4 weeks. The liquid is then drained off and fresh brine of 16% concentration together with 100 ppm SO2 and 0.2% citric acid is added. The resulting product is stored in containers kept away from direct sunlight. As per international standards (ISO/DIS 11162, 1996), the product will have the characteristic odor and ﬂavor of fresh green pepper, the color varying from pale green to green. Dry packed green pepper is produced in the same way as bottled green pepper, with the only exception being that the liquid at the ﬁnal stage is drained off and packing is done in ﬂexible pouches. Dry packed green pepper, which has similar qualities to that of canned or bottled green pepper, can replace the latter. Freeze-dried pepper retains the original color and shape of green pepper and fetches a premium price in the market. Its processing is a manufacturer’s secret.

A. WHITE PEPPER White pepper is produced from fully ripened fruit after removal of the outer pericarp either after or before drying. White pepper is preferred for use in food products, such as colored sauces, salad dressing, soups, and mayonnaise, where dark colored (black pepper) is undesirable. In some of the European countries it is white pepper that is in common use. It is made by any of the following methods: (1) steeping and retting in water, (2) steaming, (3) boiling, (4) chemical treatment, (5) the simple decorticating process of ripened fresh or dry berries. The water steeping and retting process is started when one or two fruits in the spike start yellowing when the crop is harvested and ready for the treatment. The produce after harvest is threshed and heaped in tanks through which water is allowed to run for 7–10 days. Light-weight pin heads and light berries accumulate on the surface, are removed and the remaining mass is rolled over at least three times a day during the retting stage. On the eleventh day, the outer skin of the berries is removed by gentle rubbing and the deskinned fruits (the pepper seeds) are transferred to another tank which contains a bleaching solution. The produce is left to stand in the bleaching solution for 2 days after which the solution is drained, the seed washed and sun dried. The boiling or steaming processes involve the steaming or boiling of the mature green fruits for 10 –15 min when the outer skin of the fruits gets softened during the steaming process which is then

334

K. P. P. NAIR

removed by passing through a pulping machine. The deskinned fruit (pepper seeds) are washed and treated with SO2 or bleaching powder solution, subsequently washed in water and dried in the sun. The skin of the berries recovered in this process can later be used for the recovery of pepper oil by steam distillation though it may not be economical. The process has been developed by the Central Food Technology Research Institute in Mysore, India. There are a number of chemical processes to prepare white pepper by treating the berries with acids or alkalis. As of now, no commercially viable production unit exists. Manilal and Gopinathan (1995) have described a method by which skin of both dried and fresh pepper is removed by microbial decortication. By using decorticating machines, whole dried black pepper fruits can be processed to white pepper. Since loss due to breakage will be high, production of white pepper by this method will be expensive. The characteristic ﬂavor and aroma also could be missing in the ﬁnal product and, as such, this method of making white pepper is not preferred. Among all the processes described above, the traditional retting process is the most popular and the white pepper obtained by this method is preferred by consumers. According to international standards (ISO/DIS 959-2, 1996) the product should have a maximum of only 1% permissible extraneous matter, 4% of broken berries and 15% of black berries. The minimum bulk density should be 600 g2l. On a comparative basis, while 100 kg of mature green pepper will yield 33 kg dry pepper, in the case of white pepper it will only be 25 kg. From this it can be inferred that the cost of white pepper should be at least 35% above that of black pepper. The total world demand for white pepper is around 38,000 metric tons per annum, which is a quarter of the world-wide black pepper production. There is yet another value added pepper product known as ground pepper. It comprises both white and black pepper. Ground pepper is obtained by grinding the normal pepper without adding any external material. Pepper has to be ground to a speciﬁc size and for grinding, hammer mills with copper tipped hammers are preferred to silica tipped plate mills as ground pepper obtained from the latter can contain silica in excess of the permissible limit. The ground materials after sieving, is packed and any above the speciﬁed size is sent back to the mill for further grinding and size reduction. When pepper is processed in modern spice grinding units, it is at ﬁrst cleaned of any extraneous matter and then passed through a magnetic separator for the removal of metal particles. After this, it is passed through a vibrating screen for further removal of extraneous matter and then sent to the hammer mill. The ground product from the hammer mill is fed into a cyclone separator to recover pepper powder. It is further sieved using appropriate sieves and the overﬂow is recycled. Bag ﬁlters are used after the cyclone separators to prevent escape of ﬁne particles outside. The ground pepper is packed into air-tight containers prior to shipping and the packages are made to speciﬁcations provided by the customer on size, pungency, moisture content, ash content, crude ﬁber, volatile oil, piperine,

AGRONOMY AND ECONOMY OF BLACK PEPPER

335

starch content, etc. White pepper is processed in the same manner as that described above except that, instead of black pepper, the starting material is white pepper. The ground pepper has a characteristic ﬂavor, very aromatic, slightly sharp and must be free of any extraneous odors and ﬂavors including mouldy and rancid odors and the international standards conform to ISO/DIS 959-2 (1996). White pepper can also be made from black pepper using selective grinding followed by sieving. The skin of the berries is a by-product and contains essential oil and oleoresin, both of which can subsequently be extracted.

B. CRYOGROUND PEPPER To cut down oxidative losses of essential oils and to assist in the ﬁne grinding of pepper by making the raw product brittle at low temperatures a new technique known as cryogrinding at temperatures very much lower than the normal 1008C is followed. The product disperses more uniformly in spice formulations. The process involves injecting liquid nitrogen into the grinding zone. A temperature controller maintains the desired product temperature by suitably adjusting the liquid nitrogen ﬂow rate. The exhausted gas is re-circulated for the pre-cooling process.

C. PEPPER OIL

AND

OLEORESIN

Pepper oil and oleoresin are the two most important ingredients of pepper and have a great economic value. The ﬁrst imparts the aroma to the pepper while the second gives the pungency and ﬂavor. The essential oil can be recovered by steam or water distillation. The essential oil contains mainly a mixture of terpenic hydrocarbons and their oxygenated compounds having a boiling point in the range of 80 – 2008C. Depending on the pepper cultivar, the agro climatic situations in which the crop is grown, the management aspects and the grades, the content of the oil varies. The pin heads contain the most sesquiterpenes. In order to meet different ﬂavor requirements it is possible to suitably blend oils from different grades of pepper. In industrial processing to recover essential oils, the product is ﬁrst ﬂaked using roller mills and then steam distilled in stainless steel extractors. The product to be distilled is ﬁrst heaped into the distillation unit and then compacted near the walls to prevent any steam escape. Through the bottom of the still, dry steam is passed. The still also gets heated through the jacket provided for this purpose. When the steam comes into contact with the pepper the temperature is raised and the oil present in the oil cells of the berries vaporize, rising along with the steam through the still. It is then condensed and since the oil is lighter than water it ﬂoats on the surface of the condensing water. Using an oil or water

336

K. P. P. NAIR

separator the oil is then recovered. The oil recovered by steam distillation is allowed to remain on anhydrous Na2SO4 to exclude traces of moisture. When exposed to direct sunlight and air ﬂow the hydrocarbons and the sesquiterpenes present in the oil undergo oxidative changes especially during long storage. Hence, it is always desirable to store the material in airtight containers, especially since the end use occurs after a long time because of the process of export by shipping to different countries. Oleoresin is usually obtained by extraction of ground pepper with acetone, ethanol, ethylene dichloride, ethyl acetate, etc. The main advantage of oleoresin is that it is uniform in composition and strength. Contaminants such as mould and fungus are absent in oleoresin and can be directly added to food stuff after adjusting the level of concentration. Oleoresin comes in oil soluble, water dispersible and dry forms. Currently, oleoresin is recovered either by a singlestage or two-stage process. The process involves size reduction of the ground pepper prior to solvent extraction done in stainless steel extractors. In the singlestage process the oil is recovered along with the resins by solvent extraction whereas in the two-stage process the ground pepper is ﬁrst subjected to steam distillation for the recovery of the essential oil. The composition of the oil and the oleoresin content obtained in both the processes differ slightly from each other. Due to the moisture present in the wet cake obtained after steam distillation, there is the likelihood of oleoresin yield obtained in the two-stage process to be less compared with that obtained in the single-stage process. Drying prior to solvent extraction can prevent such loss in oleoresin yield. In the case of the single-stage process, pepper is ﬂaked to 1.0 –1.5 mm thickness and packed in stainless steel extractors for extraction with organic solvent. Normally a solid to solvent ratio of 1:3 is maintained at an extraction temperature of 55 – 608C. The solvent is continuously kept re-circulated to ensure efﬁcient solid to solvent contact and after 3 h of extraction the miscella is ﬁltered and evaporated. The solids are further contacted with lean solvents from other extractors and the entire extraction process is completed after six or seven stages of extraction. The miscella sent from the extractors are evaporated in a shell and tube evaporator at temperatures not exceeding 808C in a vacuum of 250 mmHg. As the temperature starts to rise and no further solvent is recovered from the concentrated miscella, it is pumped into a high vacuum stripper. Final desolventization is done at a vacuum of less than 20 mmHg and at no stage of the operation is the temperature allowed to rise above 1008C. Initially, during vacuum distillation, the condensate recovered will mostly be the organic solvent and towards the end of the distillation the essential oil will also start coming out along with the organic solvent. At this stage, some entrainer, such as alcohol or the monoterpene fraction of the essential oil is added to the desolventiser. This is done to remove the ﬁnal traces of the organic solvent as well as to supplement the monoterpene fraction of the pepper oil which might have been lost during the ﬁnal solventization. The product is pumped into storage tanks after conﬁrming that the residual

AGRONOMY AND ECONOMY OF BLACK PEPPER

337

solvent levels are within limits. Suitable blending is used to meet the customer requirements in terms of both oil and piperine contents. Some customers specify the homogeneity of oleoresin through a homogenizer which may be a colloid mill or sand mill. The oleoresin, which is dark in color, is bleached using activated carbon to obtain decolorized oleoresin. All over the world, oleoresin extraction is done mostly in batch extractors. Investigations conducted at the Regional Research Laboratory, Trivandrum, in the state of Kerala, India, showed that extraction can be done in ambient conditions and at low residence time (Sreekumar et al., 1993). Present day extraction, which takes more than 18 h to process one batch, can be dispensed with employing continuous counter-current extraction. Simultaneously, the quantity of solvent needed for extraction and its loss in the extraction process can be substantially reduced. Research in the laboratory has shown that extraction by enzymatic breakdown of spice cell walls of the dried pepper can be accomplished. In this process, the product is mixed with water, adjusting the pH by the addition of citric acid, treated with enzymes, either individually or in combination, which is followed by centrifugation. The enzymes used are those which are commercially available, such as cellulase pectinase, hemicellulase, and liquefaction enzyme preparations. Extracts having good sensory and compositional properties were obtained with some of the enzyme combinations. Optimum results were obtained using a combination of cellulase and pectinase preparation. The addition of hemicellulase did not result in any improvement of ﬂavor. This solvent-free extraction procedure is an alternate route for the recovery of ﬂavor compounds which is not commercially possible at the moment. Another solvent-free extraction procedure which has shown much promise is the super-critical carbon dioxide extraction. Low energy consumption, high purity of resultant extracts, environmentally acceptable, and possible fractional separation of the components are some of the major advantages of this process. The current demand for piperine is on the increase. Piperine can be produced from the oleoresin in the concentrated form by centrifuging the oleoresin in a basket centrifuge. Part of the oil along with some resin can be collected after centrifugation and the centrifuged cake, which contains as much as 60% piperine, is obtained. By washing the centrifuged cake with pepper oil and further centrifuging, the piperine concentration can be further enhanced. A number of secondary products have been developed from oleoresin to improve solubility in food substances which are marked under various trade names by the manufacturers with their ﬂavor strength indicated on the respective label. Standardized seasonings, which are able to withstand almost all processing conditions, are some of the secondary products. Emulsions, solubilized spices, dry soluble spices, encapsulated spices, heat resistant spices, fat based spices, etc., are some of the additional secondary products. Emulsions are liquid seasonings that are prepared by emulsifying blended pepper oil or oleoresins with gum acacia or other permitted emulsifying agents. A stabilizer is added to check

338

K. P. P. NAIR

creaming and the products have only limited shelf life. Solubilized pepper is blended pepper oil and/or oleoresin which is mixed with one of the polysorbate esters in such a concentration to give a clear solution when mixed with fresh water. Dry soluble pepper is prepared by dispersing the standardized oleoresin onto an edible carrier-like salt, such as dextrose, rusk, etc., in order to give a product which has a ﬂavor strength equal to that of good or average quality ground pepper. In addition to the above-described, industrially produced value added byproducts of pepper, there are a few others, such as heat resistant pepper, fat based pepper, extruded spices, and micro encapsulated ﬂavors. Of these, micro encapsulation is the most important. It is a process by which the ﬂavor material is encapsulated in a solid matrix and is ready for release as and when required. Encapsulation can be achieved by a number of techniques, such as spray drying, coacervation, polymerization, etc. Of these, spray drying is the most popular. The process involves homogenization of the oil/water mixture in the presence of the wall material and later removal of water under controlled conditions in a spray drier. The advantage of the spray drying process over the others is that the product, though in contact with existing gas at a higher temperature, will never reach this high temperature within the short residence time. The oil/water emulsion is atomized through an atomizer in the spray drier during spray drying. Commonly used wall materials for encapsulation are selected from among vegetable gums, starches, dextrins, proteins, sugars, and cellulose esters. The wall material is selected so as to meet, as closely as possible, the properties, such as low viscosity at a high solid state, ability to emulsify or disperse the active material, non-reactivity with the material to be encapsulated both during processing and on prolonged storage, uniform ﬁlm forming property, ready availability, etc. Investigations indicate that the addition of surfactants during the process of encapsulation prior to spray drying will reduce the oil loss (Anon, 1989). Another process for microencapsulation is the CR-100 process (Anon, 1995; Findlay-Wilson, 1995; Mos, 1995) and this process overcomes the limitations of the spray drying process such as reduction in ﬂavor quality and yield. The heat resistant pepper is the double encapsulated product in which the capsules are rendered water insoluble by a suitable coating and the contained ﬂavor is released only at high temperatures such as in the case of baking. Fat based pepper is a blend of pepper oil and/or oleoresin in a liquid edible oil or hydrogenate fat base formulated for use in products such as mayonnaise. Extruded spices are spices sterilized, ground and encapsulated in a single step. The process involves a combination of pressure changes, temperature shock, and shear. The process, accomplished in a twin screw extruder, helps in retaining the color and ﬂavor in the original form because of the very short processing time. When fresh spice is fed into the twin screw extruder the starchy materials get gelatinized and form an encapsulated product. This process substantially reduces

AGRONOMY AND ECONOMY OF BLACK PEPPER

339

contamination by bacteria, mould, and yeast. The product emerges from the extruder as a “spice rope” which is then cut into pellets. Lucas Ingredient, Britain, markets the products under the brand name “Master Spice” (Scott, 1992).

X. THE FUTURE OF THE GLOBAL PEPPER ECONOMY The global pepper economy has different facets. The ﬁrst is the supply side, the second the demand side, the third the price projections, both supply and demand included simultaneously, and the fourth referring to the prices and equilibrium in the market. Bade and Smit (1992) have resorted to precise models to study the above facets. This review does not provide an elaborate scope to study these models, but the essential conclusions stemming from these models will be discussed in this section. The various models are all mathematical in nature.

A. THE SUPPLY SIDE

OF THE

PEPPER ECONOMY

The supply side of pepper is approached on a national basis, examining the area, production and exports from each country. The normal production is obtained by multiplying the number of pepper vines by the average yield of a vine in a country. Although measurement by vines must be considered much better, a similar line of reasoning can be applied with area and average yield per hectare. The major drawback of this method in comparison with using vines is that it adds to a source of variation. Not only the per vine yield ﬂuctuates, but also the number of vines per hectare. In India this is a very signiﬁcant variation, but in other countries, for example in Thailand, the number of vines per hectare is almost the same all over the country. If the intensity of cultivation is relatively constant and there is no signiﬁcant change in cultivation patterns, it is possible to estimate a nation-wide normal average yield per hectare. In case the average number of vines per hectare cannot be expected to be constant an assumption is needed about the change. In case the intensity of cultivation increases or a superior variety is introduced, it can then be assumed that there will be an upward trend in the normal yield per hectare some 2 years after the start of the intensiﬁcation when the planted vines become productive. If normal yield can be used as a basis for forecasting, the actual yield will deviate from it because of the inﬂuences of weather, amount of fertilizer applied and the level of maintenance. Unfortunately, these effects are not conﬁned to only 1 year. When a year is very wet, the crop will be small, but, most importantly, there will be also the incidence of dreaded diseases such as Foot Rot. Invariably, these diseases will determine the yield in

340

K. P. P. NAIR

the succeeding year. Also, the effects of neglect or exceptionally good maintenance will spread over more than 1 year. All of the above illustrates that even when data on productive vines and average yield are available, and quite reliable, there would be sufﬁcient scope for simulation and expert interpretation. However, global information is very expensive to gather and gathering information on agriculture can even be a lot more expensive and time consuming task. This is not a priority of the developing countries. Planning depends largely on information and a model cannot compensate for the lack of quality data. Inasmuch as quality of information is concerned, it can only interpret and detect inconsistencies. Thus, in conclusion, it can be said that the modeling of production and supply presented is only a preliminary step on the road to a more sophisticated model analysis based on superior data. In the following paragraphs, a brief review of the pepper economy of important pepper producing countries, such as India, Brazil, Thailand, Malaysia, and Sri Lanka is given.

1.

India

It is a very important observation that the area under pepper in India would be enough to meet the entire global demand if only average yields were a third of what is obtained in Sarawak (Malaysia). This points to very important lacunae in data collection, their precision and reporting. Except in large pepper plantations in the state of Karnataka and Northern Kerala, pepper is still grown in homestead gardens as an intercrop between coffee, cardamom and quite often trailed using arecanut and coconut trees as standards. Up to 1986 a survey was conducted by ﬁeld extension agents in Kerala in randomly chosen parts of Kerala in such a way that within 5 years every part was visited once. The total area under pepper from the population was then multiplied by the inverse of the sample area divided by state area ratio. The question asked was to estimate area on the basis of 560 vines ha21. The method was applied, asking the same people, to get production estimates. From 1987 onwards, the Department of Economics and Statistics of the state is attempting to introduce a more sophisticated system especially to estimate production. Though the method used is rather elementary, other countries probably do not use better systems. A similar equation as that used in India is used in Brazil as well. On the market side, it is seen that an increase in price is an incentive for enhanced acreage as more farmers resort to use of fertilizers and adopt better maintenance practices followed by more picking rounds. The previous year’s price is the trigger for these added incentives. Farmers also keep pepper in stock, hoping to sell for a better price in the following year if a price drop is encountered in the current season. Alternatively, when prices escalate and farmers release

AGRONOMY AND ECONOMY OF BLACK PEPPER

341

their stocks it will seem as though production has gone up, which is not the case. Indian exports depend largely on the current season’s crop. The other factor is the inﬂuence of price changes on stocks of traders. When there is a sudden increase in price, the total amount of exports may even exceed production. The year 1985 was exceptional. Again, 1991 and 1992 were also exceptional on account of the restructuring of the Indian economy when globalization and liberalization processes set in.

2. Brazil In Brazil there are no time series on area by cultivation system, although there are different cultivation systems in the country. To be sure, it is still totally unclear how the Brazilian Pepper Exporters Association gathers the data which are presented at the IPC meetings. Apparently, data on area and production do indicate that pepper yields have gone up over the years in Brazil. But this is not the case. If production is divided by area it would appear that there is a decreasing trend over time. In fact, between 1978 and 1979 yields plummeted despite area expansion suggesting that in those years data on area did not include area with immature vines. After 1979 the correlation between area and production was strong, although 1984 was an extraordinary year which had a much higher yield than could be explained. This must be due to positive price correlation. Pepper consumption in Brazil is negligible compared with production and presumably kept out of production statistics and as there are some other small pepper producing regions outside of Para state, it can be expected that almost all pepper produced will be exported. However, there is a rather constant amount of about 1.5 thousand tons that is apparently not exported. Price is the most important variable in the current productive area under pepper.

3.

Indonesia

Indonesian data on aggregate area under pepper is rather scanty. The ofﬁcial records claim that the total area did not change from 1983 until 1987. However, records for Lampung and Bangka show substantial changes over these years. The time series on area of Lampung and Bangka are still too short. Further, there is hardly any information on area in Kalimantan. For future modeling research, regional disaggregation of supply of Indonesia is important, especially because of the special status of Bangka, which produces only white pepper. Prices have played only a minor role in area expansion. Clearing of new land for pepper production will play a predominant role in pepper production especially in

342

K. P. P. NAIR

Kalimantan and Sulawesi. For Indonesia the inﬂuences of price on stocks is far less important than on maintenance, a similar conclusion to that for India. The year 1993 was a bad year because of a serious set back in production and thus on exports reﬂecting a serious situation for white pepper in Bangka.

4.

Thailand

Thailand is a relative newcomer to pepper production. Until 1986, area was quite steady, thereafter it expanded, but once again dropped between 1991 and 1993 when low prices ruled the market. At this time the government encouraged farmers to diversify from pepper. Changes in prices were inﬂuenced by prices prevailing in the previous year, while area was held steady. Productivity per hectare is very high and the production was inﬂuenced by the previous year’s prices. Over recent years, production has been on the decline. The country has a high domestic consumption, which is around 5000 tons per year. Exports take a ﬁxed share of production, which is inﬂuenced by the previous year’s shortage or surplus.

5.

Malaysia

Malaysian data on area and production of pepper, of which nearly 98% refers to that of Sarawak, are rather unreliable. Foot Rot strongly inﬂuences yield. Data on production is based on exports and as pepper farmers from Sarawak speculate on the pepper market, export is not a reliable index to show production trends as these farmers who speculate in the pepper trade are rich. Hence, large differences between production and exports are bound to arise. From 1989 onwards there is a decline in area arising from shortage of labor and its high cost, which did not allow for proﬁtable production to the extent as was the case in the past. The price of the previous 2 years greatly inﬂuenced production per hectare, though this inﬂuence is through the planting of new vines.

6.

Sri Lanka

A traditional pepper producer, there is only a slight increase in area over the years. Productivity is low and despite an increase in area, production did not correspondingly increase, which is rather surprising. The production pattern in 1993 is just the opposite to that of Indonesia, a year that showed an incredible increase in production and exports. Export in 1990 was exceptionally high, exceeding production. Export can be explained on the basis of production, the change in production and domestic prices.

AGRONOMY AND ECONOMY OF BLACK PEPPER

343

7. Countries Other Than Those Named Above In addition to the countries named above, countries such as Madagascar and those on the African continent, such as Benin, Cameroon, Congo, Gabon, Ethiopia, Nigeria, Zambia, Zimbabwe, etc., also produce pepper, of which Madagascar with a production of about 2500 tons leads the pack. As long time series and information on area and production are needed, as well as information on internal markets, no tangible conclusions can be drawn from the existing information.

B. THE DEMAND SIDE

OF THE

PEPPER ECONOMY

To make a clear analysis of the demand side of the pepper economy, precise data on end-use of pepper is needed. To date, precise data on differentiation of pepper use in food industries, institutional catering and household consumption is unavailable and, as such, only rough estimates can be made. Further, a distinction between black and white pepper is needed. The following countrywise groupings will provide an insight into the demand side of the pepper economy.

1.

European Union

As a shift in the food consumption pattern takes place, for example consumption of more pre-fabricated food, fast food, etc., the demand scene will ﬂuctuate. Western Europe is price conscious as far as stock formation is concerned. Also, the level of stocks at the end of the previous year has a reasonably signiﬁcant inﬂuence.

2.

Rest of Western Europe

Starting from 1998, Switzerland has import data on pure pepper, whereas until then these ﬁgures also included pimento and capsicum. Price changes inﬂuence stock formation.

3. Eastern Europe and CIS The economic problems that overtook Eastern Europe and the CIS have had a very negative effect on import since 1990. Price changes inﬂuence stock formation.

344

K. P. P. NAIR

4.

North America (USA and Canada)

There is a negative relationship between price levels and starting stocks. When prices are low, traders will keep more stock expecting prices to increase again sometime in the near future and if stocks are high, traders tend to sell.

5.

Japan

Demand for pepper in Japan is pegged to the Gross Domestic Product (GDP) in more than one way. A shift towards outdoor, ready to eat food, in particular Western food, goes along with a rising GDP. Changes in stocks paralleled trends in North America.

6. Australia and New Zealand Stock levels are related to consumption and no price inﬂuence is found on the demand side.

7.

Latin America

Per capita import of pepper did not show any increase. Between 1976 –80 and 1985 – 88 signiﬁcantly lower imports took place and stock formation depended on prices and lagged stocks.

8.

Asia and Paciﬁc Countries, Singapore, Australia and New Zealand (Excluding China)

Stock formation depended on prices and lagged stocks.

9. China China was a small importer of pepper and the import spurted in the latter part of the 1970s. Of late, imports have dwindled.

AGRONOMY AND ECONOMY OF BLACK PEPPER

10.

345

Middle East and North Africa

These regions have a rather irregular import pattern. Stock formation was inﬂuenced by price changes.

11.

Rest of Africa

Per capita import of pepper did not show any increase as was the case of Latin America. Since 1990 imports surged. Stock formation was inﬂuenced by price changes.

C. PRICES

AND

WORLD MARKET

As far as the world market is concerned Singapore has a special place for pepper import and its commercial consequences. Of the total pepper exported from Malaysia, 69% moves via Singapore. The function of Singapore as an entreport is accounted for by the variable total exports of producing countries less estimated world consumption, which is merely an estimate of the change of stocks outside producing countries. Singapore is expected to import part of these stocks and keep the major part of these stocks as carryover stocks. Some part of pepper imports is consumed, but no statistics are available on pepper consumption in Singapore. It is estimated that consumption is approximately 1% of imports. The price pattern indicates that pepper traders are presumably more interested in trade if prices are high, which does not seem unrealistic as margins will deﬁnitely be correlated with the price peaks.

D. THE PEPPER PRICE OUTLOOK

BY

2020

Perhaps the most important aspect of the price structure is the impact of the size of the Indian supply responses to price ﬂuctuation. This aspect has not been sufﬁciently captured. For all countries the investment side is not yet represented adequately for long-term analyses. Stock formation at various levels needs further work as well. Price cycles of around 7 years will continue in future showing that the 10 year cycles in income have very little impact. There will not be an increase in average real prices. Obviously, with inﬂation around 4% per year, nominal prices do increase on average.

346

K. P. P. NAIR

E. THE PEPPER SUPPLY OUTLOOK

BY

2020

Projections indicate that production and exports are expected to decline in Brazil, Malaysia, and Thailand because of the increase in labor costs. Brazil shows a modest decline on average, while India and Indonesia show strong growth. The same conclusion as for Brazil can be drawn for Malaysia and Thailand. These countries are becoming too rich to grow labor intensive agricultural crops such as pepper. Sri Lanka is expected to level off on account of limited area. Exports from Vietnam and China, especially the former, is expected to grow rapidly reaching an aggregate of over 60,000 tons. A large quantity of pepper is imported into India from Vietnam, often clandestinely, in the garb of value addition. Since Indian pepper fetches a premium in the world market, unscrupulous traders mix the low quality and cheap Vietnamese pepper with Malabar pepper and re-export as Malabar pepper to the US and Europe. On detection by the consumers, this adversely affects Indian pepper in the world market, price-wise and also future imports. The gap left behind the relatively rich countries such as Brazil, Malaysia and Thailand in production and exports is ﬁlled up by Vietnam and China and not by the traditional major producers such as India and Indonesia. Exports from Madagascar and other African countries are expected to be steady. This inevitably leads to the important conclusion that almost all of pepper export will originate in Asia.

F. THE PEPPER DEMAND OUTLOOK

BY

2020

On the consumption side, North America, Japan, and the European Union will show a steady growth. Developments in Eastern Europe will impact the consumption pattern, so will the increased consumption in the Paciﬁc countries and other up coming Asian regions, such as Philippines, Korea, Cambodia, etc. Another major consuming area is the Middle East and North Africa.

G. COUNTRY-WISE ECONOMIC GROWTH IMPACTING PRODUCTION AND CONSUMPTION It is informative to analyze the GDP/growth rates that impact pepper production and consumption. The details of the estimates are presented by Burger and Smit (1997). Historical developments have impacted consumption. For instance, the late eighties have been years of political turmoil for Eastern Europe and the CIS and this has adversely affected growth rates impacting consumption negatively. On the other hand, countries in eastern, far eastern and south eastern

AGRONOMY AND ECONOMY OF BLACK PEPPER

347

Asia have enjoyed very buoyant growth rates which positively impacted consumption. However, such very high growth rates cannot be maintained for decades, as has been clearly shown by other countries with high growth rates in the past. In the long run, the USA will average a growth rate of just below 2% per annum, while Japan, despite its recovery, may not reach the growth rate of the USA, after the period of high growth up to 1990. In Western Europe, growth rate will hover around 1– 3%. Eastern Europe, CIS and former Yugoslavia are expected to recover, but, on average, at rather moderate levels of economic growth. Growth rates in Latin America will be somewhat higher than in Europe. In south Asia, India is maintaining a growth rate around 5%, though expectations were higher. Pakistan has a somewhat higher rate. In southeast Asia, Philippines will show a moderate growth, while in the other countries, though showing higher rates, a declining trend would be observed. The same applies to China as well. Vietnam will grow around 4%. With the exception of Nigeria, Africa on the whole is showing only very moderate growth rates. The above growth scenario indicates that potential for pepper export from producing countries to the USA and Europe remain bright in the long run. This is primarily because of the fact that the processed and fast food cultures will unfold further in the years to come. Southeast Asian countries are also potential markets. There is not much scope for export into Africa. Latin America, which will have a higher growth rate, is a potential market, but countries such as Brazil are already important pepper producing countries and, as such, export from other pepper producing countries has only limited scope for export to Latin America.

XI. PEPPER ECONOMY IN INDIA As India is the most important producer of black pepper in the world, a discussion on pepper economy and marketing patterns in the country is relevant. Currently, India accounts for 49.5% of the total acreage in the world, but in 1951 it was 70%. Though the country holds the largest pepper acreage in the world and contributes up to 29.8% of world production, its productivity is the lowest in the world, an abysmally low ﬁgure of 294 kg ha21. In fact, the total contribution of India to world production dropped from 66% in 1951 to the current 29.8%. This has been the case for exports as well. From 1951 to 1991, the share of India in the global pepper market fell from 56 to 23%. It is interesting to note that while India’s position of pre-eminence in pepper production dwindled, other countries’ performance was on the rise. The last decade of the past century witnessed a global downtrend in pepper production. In India, pepper is primarily cultivated for export. From around 16,000 tons in 1950– 51 export has scaled up to 32,000 tons, a 100-fold increase, by the turn of

348

K. P. P. NAIR

the last century. The nineties witnessed a ﬂuctuating market because of price ﬂuctuations. The latter half of the nineties witnessed a sharp increase in unit price peaking at about US $50 kg21 in 1997 and though export per se witnessed a drop, it was more than compensated by the escalating unit price. In the latter half of the nineties global pepper production was around 189,000 tons of which more than 118,000 tons were exported as per the data provided by the IPC headquartered in Jakarta. In these estimates India tops the list of producing countries with 65,000 tons followed by Indonesia. India exports spices to more than 120 countries around the world and of these countries, export to 23 countries account for about 82% of the hard currency earned (in US $) by the country. In the case of pepper, India is the major supplier to 83 countries and among the importers USA tops the list with an annual growth rate of 5% in quantity and 3% in value. Of the total spices imported into the USA, pepper accounts for about 15% of which 50% orginates in India. USA is also the major importer of spice oils, primarily oleoresin, and India supplies close to 86%. However, the recent trend shows a decline in import into the USA. In the beginning of the last decade pepper import from India to the USA was 56%, which declined to 30% by the middle of the decade. Though USA tops the list of pepper importers from India, it is interesting to note that per unit value realization is lowest from Russia, Canada, Italy, and Poland, with Russia topping the list. It is also interesting to note that there is a visible shift in the food habits of Americans, where the preference for hot spices like pepper in the last two decades accounts for nearly 75% consumption. The US trade association projected spice consumption to peak to 1000 million pounds by the turn of the century (Cheriankunju, 1996a). Maintenance of quality assumes the greatest importance in this regard. The United States Food and Drug Administration (USFDA) detained 1164 consignments of Indian spices due to contamination with seeds of noxious weeds in 1995. In fact, there has been a steep escalation of detection of contaminated consignments from 140 in 1991 to 757 by 1994 (Sivadasan, 1996), which, indeed, is a very poor reﬂection of Indian trade. With the WTO tightening its noose on sub-standard quality of imported food articles from the developing countries, India would do well to clean its own stables lest a very adverse fallout on the pepper market follows. It is, however, encouraging to note that the Spices Board of India, the apex body attached to the Ministry of Commerce, Government of India, has initiated a number of measures to ensure that quality of pepper is guaranteed both at the producer as well as the exporter levels. The former USSR was the largest pepper importer from India, with a total tonnage of 19,400 in 1989 –90, which came down dramatically to 6060 tons by 1994– 95 primarily due to political turmoil in the region (Cheriankunju, 1996a). In spite of the economic instability, Russia continues to be a major consumer of pepper from India. Russia with an import of about 3200 tons in 1994 – 95 accounted for US $6.05 million followed by Italy at US $3.62 million and Canada at US $2.54 million. Traditional export of pepper from

AGRONOMY AND ECONOMY OF BLACK PEPPER

349

India to such countries as Hungary, North Korea, Maldives, Trinidad, Yemen, United Arab Emirates, and Yugoslavia has considerably decreased and in their place new markets, such as Chile, Kazhakistan, Tajikistan, and Venezuela have emerged. Of these, Kazhakistan imports the most. Apart from these, Japan is fast emerging as a potential market for Indian pepper. This is mainly due to the fast changing food habits of the country where ready-to-eat “fast foods” are emerging along the lines of western food habits, where pepper forms an important culinary component. In 1994– 95 Japan imported 41,911 metric tons of spices of which 17.8% was pepper. Most of the pepper goes into industrial uses followed by domestic consumption and other services sector. The annual rate of increase in spice import by Japan is estimated to be 5% (Cheriankunju, 1996b) and share of Japan’s import from India is stagnant at 5.5% while that from Malaysia, the bulk exporter to Japan, stays at 68%. Indonesia contributes 31%. Indian pepper has a premium in the world market and if strategic advantages are exploited following the establishment of WTO, India can scale up its export to Japan. Additionally, Israel is a potential market for Indian pepper. In 1994– 95 Israel imported 6902 tons valued at US $7.42 million (Anon, 1996). Among the potential markets for Indian pepper, USA, Canada, Russia, UK and Japan are on the top rungs as these countries have a lower level of instability in pepper imports (Jeromi and Ramanathan, 1983). According to the International Trade Center, Geneva, world trade in spices during the last decade averaged 450,000 metric tons valued at US $1500 million of which pepper is the major component, which accounts for as much as 30– 35% in quantity and 20 – 25% in value (Peter, 1996b).

A. PEPPER PRODUCTION SCENARIO

IN INDIA

In 1950 –51, nearly 80,000 ha of cultivated area in India was under pepper of which 98% was in Kerala State. An analysis of the trends in growth with regard to area, production, and productivity is given in Table XIII. It can be surmised from the table that during the last ﬁve decades, overall acreage, production, and productivity have declined all over India and, in particular, in Kerala state. The steepest fall was in productivity. On an all India basis, while acreage slipped by only 22% from the ﬁfties to the nineties, production and productivity declined by 54 and 89%, respectively. For Kerala, the respective ﬁgures for decline in area and productivity are 25 and 96% while production increased by 70%. In fact, both on an all India basis and for Kerala state as well, productivity showed very dramatic negative trends. The reason for the steep fall in productivity, especially in Kerala, is due to the prevalence of diseases, of which the Phytophthora Foot Rot is the most serious and there is, as yet, no permanent remedy for this dreaded disease. As the

350

K. P. P. NAIR Table XIII Area, Production and Productivity Growth Rates for Black Pepper in India and the State of Kerala

Details Area Production Productivity

1950s India Kerala India Kerala India Kerala

0.87 0.94 1.36 0.40 0.44 0.46

1960s 1.12 0.19 2 0.50 2 0.88 2 1.40 2 1.07

1970s 1.05 2 0.19 0.34 2 0.40 1.27 2 0.22

1980s 1.81 1.85 2.43 2.26 0.63 0.41

1990s 1.11 1.00 2 0.32 0.70 2 1.40 2 0.01

Overall 0.68 0.70 0.63 0.68 2 0.65 2 0.02

Note: Date given as log Y ¼ a þ bt in percent per annum. Source: Economic Preview, Government of the State of Kerala.

question of area expansion has only very limited possibilities, if there is to be a breakthrough in pepper production in Kerala, it will have to be through breeding superior disease resistant varieties and good agronomy to match. Soil management is still based on bookish knowledge and there is a very clear dearth of fresh ideas. Pepper cultivation in Kerala is largely concentrated in certain districts. Kannur, Kozhikode, Kollam, and Thiruvananthapuram districts, in sequence, contribute more than 80% of the total pepper area in the state and 75% production of the state. By the early nineties, the share of these four districts declined to 33 and 29%, respectively, in acreage and production and others took over. Currently, the districts of Idukki, Wayanad, and Kannur have emerged as the major pepper producing districts of the state contributing 60% of area and 67% of the production. Among the three, Idukki district leads with 21% area and 27% production. It is followed by the Wayanad district, which during the past decade has contributed 25% of pepper produced in the state. A very interesting development of late is the emergence of the Idukki district as the front-runner in the production of “organic pepper”, much valued in the USA and Europe. Based on the cool climate of the region and relatively lower impact by the chemical agriculture, the district holds out much promise to emerge as a global leader in the production of organic pepper. Pepper as a mono crop is conﬁned to the district of Idukki and Wayanad which constitute only 3% of the total area (George et al., 1989). There is ample scope to cultivate pepper as an inter crop in coffee, tea and cardamom plantations. The high range districts of Kerala, such as Idukki and Wayanad, where currently most of these plantations are located, are potential areas where pepper can be grown as an inter crop. Silver oak, used as a shade tree in these plantations can be used as standards (support for the pepper vine to trail). There is a strong case to initiate systematic ﬁeld trials to develop a clear-cut package of practices in growing

AGRONOMY AND ECONOMY OF BLACK PEPPER

351

pepper as an inter crop in these plantations. Isolated studies have been carried out. However, no efforts on a large scale have been initiated. This is a very promising area for enhancing pepper production in India, until now not tackled in a very systematic and scientiﬁc manner. Pepper is an important component of “homestead farming” in Kerala and the vines are trailed on coconut trees, arecanut trees, mango trees, jackfruit trees, and tamarind trees, which are generally grown in small farms owned by small scale and marginal farmers. The level of inputs, namely fertilizers, pesticides, etc., in these situations is low. Transformation of these homestead gardens due to various socio-economic reasons largely reduced this practice. Coupled with this, the incidence of the dreaded disease Foot Rot swept away a substantial number of pepper plants in homestead farms. There is a strong case for developing a package of practices at low input levels in these homestead farms. There is a need to re-introduce pepper in homestead farming, both in urban and rural areas. For this purpose, bush pepper offers much promise.

B. BUSH PEPPER Bush pepper holds out much promise in homestead farming. The plagiotropic lateral fruiting branches of pepper exhibit sympodial growth and these branches when rooted and planted grow into a bush. More fruiting branches are produced by these bushes and ﬂowering commences in the ﬁrst year of planting itself. A common method developed at the Indian Institute of Spices Research (IISR) involves the following steps: (1) selecting healthy lateral shoots of the previous year’s growth with two or three leaves, (2) dipping the cut end in a commercially available rooting hormone, and (3) planting them in moist rooting medium, which consists of weathered coir dust, in a 200 gauge PVC bag of about 40 cm £ 25 cm size, and later the bags are kept under shade. Prior to closing the bag, air is blown into it. In about a month, the cuttings start producing roots. After about two months the bags are opened and allowed to remain in that state for a few days before transplanting into clay pots of convenient size. Almost 80% success can be ensured provided the steps described above are taken meticulously. Bush pepper can be planted at a minimum density of 2500 plants ha21 spaced 2 m £ 2 m and the population can be doubled if spacing is reduced. Record yields of 1960 kg ha21 have been obtained using Panniyur-1 variety (Annual Report of IISR, 1997) and yields in the range 1600 –1700 kg ha21 are quite common. Bush pepper is commonly grown in Indonesia, Malaysia and the Philippines. It is still very far from catching on in India, and the home state of pepper, Kerala. Figure 8 shows Bush Pepper in India.

352

K. P. P. NAIR

Figure 8 Bush Pepper in India.

C. ECONOMICS OF PEPPER PRODUCTION STATE OF KERALA

IN THE

In India, Kerala is the most important pepper producing state and the following discussion pertains to the state. Vinod (1984) and Santhosh (1985) have attempted to make a dependable estimation of the cost structure of pepper production in the Idukki and Kannur districts, respectively, which are at the forefront of pepper production in the state of Kerala. Pepper production in the state of Kerala is highly labor intensive and labor costs account for more than 50% of the total cost of production. An important reason for the decline in pepper production in the state could be due to the non-availability of timely labor and its very high wage structure on an all India basis. Among all the states of India, Kerala’s labor wages are the highest. Though Idukki district leads in cultivation, proﬁtability is higher in Kannur district. The cost-beneﬁt ratio was 1.09 for Idukki district and 1.16 for Kannur district as per the ﬁndings of the above mentioned authors. The pay back period of pepper cultivation was estimated as 10 years in Idukki district and 11 years in Kannur district. The corresponding net present worth is approximately US $93 for Idukki district and US $148 for Kannur district and the internal rate of return 13.48 and 17.22%, respectively.

D. MARKETING Village traders dominate the local market inasmuch as small pepper farmers are concerned in Kerala. The trade is based, in many ways, on faith

AGRONOMY AND ECONOMY OF BLACK PEPPER

353

and long-term personal bonds, which is absent in the case of whole-sale businessmen who deal in large quantities. Additionally, transportation to the city or town market is quite often out of the reach of small farmers and this also leads to the small farmer going to the village trader. By tradition, village traders collect the product directly from the farmers’ yard or if not possible from the neighboring village markets. Wholesale businessmen, though, offer higher prices compared to the village trader. Village and internal wholesale merchants are important market intermediaries (Vinod, 1984). Small and marginal farmers generally affect village sales in the nature of a pre-harvest contract whereby the farmer is obliged to sell his product at a pre-determined price prior to harvest and if there arises any ﬂuctuation in price subsequently, both parties are bound to go by the earlier contract of sale agreed upon. The up country wholesale merchants transport the produce to Kochi, which is the port city of Kerala and the prime business center of the state, in the central part of the state. Once the produce reaches Kochi, the commission agents negotiate with the exporters or wholesale merchants or brokers. The brokers act on behalf of the internal wholesale merchants or exporters and negotiate with the commission agents. At the exporters’ premises the produce is processed and graded. Prior to the ofﬁcial clearance of the produce for export, it is subjected to check sampling and all the tests for grading. The consignment thus cleared for export is then handed over to the clearing or forwarding agency after completion of the paper work. Subsequent work is carried out by the forwarding agency. Four major marketing channels were identiﬁed in Idukki District, while there were ﬁve in Kannur district. The share of the producer on the free on board price (FOB price) of pepper was estimated to be 86.06%. Compared to this, the producer is the beneﬁciary of a higher share, which is 88.80%, when the produce is internally consumed moving through the domestic consumption channel. In either case, it is clear that middlemen corner a good percentage of the value chain.

E. PEPPER FUTURES MARKET The origin of futures trade is almost a century old in India. Futures trade is an insurance against price ﬂuctuations of commodities handled by traders. It was the Indian Pepper and Spice Trade Association (IPSTA) who heralded the futures trade in pepper in India in Kochi (Sethuram, 1995). The move had a very positive effect on the transfer of risk and recovery of price. From hedgers the risk is transferred to the speculators, who in turn provide the liquidity to the market. Future markets are standardized with regard to trading speciﬁcations and the terms of delivery. The international pepper price shows a trend which is unfavorable to most of the pepper producing countries. This is largely attributable to supply situations rather than production scenarios. The IPC in an effort to

354

K. P. P. NAIR

ensure an efﬁcient price management system conducted a study, sponsored by the United Nations Commission for Trade and Development (UNCTAD), on the viability of pepper futures contract. Based on this a working group of representatives from the IPC member countries was set up to examine the requirements of a global pepper futures contract. Consequently, the IPSTA at Kochi initiated action for the establishment of an International Pepper Exchange (IPE) at Kochi. The expectations are that it will provide an incentive in pepper trading, both in the domestic and international markets, ensure better prices to producers and also extend a hedging facility to exporters of all the member countries. Further, it is also proposed to link the Kochi pepper exchange with the Kuala Lumpur Commodity Exchange at a future time, when a second trading ﬂoor would be started. The globalization process, in general, and the increasing volatility of pepper prices necessitated the establishment of a global pepper market to trade pepper futures contract. This also facilitates a free trade regime in pepper, as the product can be freely moved among the countries without tariff barriers and the restrictions of quotas. On account of the additional premium for Indian pepper, its prices in the international market are relatively higher than that of pepper produced by other countries. Low yield combined with high unit cost of production combined with better prices in the domestic market leads to a higher international price. Also, the improved stock holding capacity of the farmers inﬂuence price, because, in times of price decline, farmers can hold on to their produce without unloading it to the market. Both corporate procurement and speculation play a signiﬁcant role in pushing up the price. Futures trade ensures a constant supply and steady price for the produce. One of the main criticisms against the exchange is how it can best translate the beneﬁts to the actual producers. Pepper production in Kerala is mainly scattered among small farms and the market awareness of these small and marginal farmers is indeed scanty. As much as 72% of the farmers belong to this ignorant category (George et al., 1989). These small and individual farmers neither have sufﬁcient quantity to trade nor the technical expertise to beneﬁt from the pepper exchange. This has resulted in the futures trading in Kochi being handled mainly by representatives of exporters. The history of futures trading in pepper has clearly shown that producers have almost no say in its operations, including the large ones. The Primary Agricultural Credit Societies (PACS) have also not attempted to enter the futures market. The State Marketing Agencies, such as the State Trading Corporation and the State Marketing Federation, both governmental agencies, are wary of their association with the exchange set up as they have met with large ﬁnancial losses. This is only natural, because governmental agencies in Kerala rarely have true accountability, which is a legacy of the earlier socialistic pattern of governance in the state. If fair trading has to be ensured, the possibilities of manipulative tactics, which create an artiﬁcial market to the detriment of the interests of the producers, by the powerful members of the exchange, have to be fully scotched. But this invariably does not

AGRONOMY AND ECONOMY OF BLACK PEPPER

355

happen and the speculator calls the shots rather than the toiling producer, however big or small he may be. Despite these shortcomings, the IPSTA is hopeful of bridging the disparity, often quite wide in certain instances, between farm gate price and market price, by eliminating needless middlemen involved in transactions, which would ensure better prices for the producers. Liquidity of traders is expected to increase as banks are now willing to advance as much as 80– 90% of the value of the produce when it is hedged against adverse price ﬂuctuations through the medium of futures contract. Reckless speculations will be prevented by the imposition of limits on open position, daily price band ﬂuctuations bound both up and down and collection of special margin deposits to curb speculative activity through ﬁnancial restraints. An important component of the economics of pepper growing in Kerala is the availability of reliable market information to the farmers. As of now, most of the pepper farmers have either no dependable access to reliable market information or are illiterate so cannot make use of it. Hence, the need to educate farmers on the importance of acquiring reliable market information and marketing processes in pepper is of paramount importance. One of the prime reasons for pepper farmers’ inability to realize a good price, especially in the case of small and marginal farmers, is the lack of adequate storage facilities. Holding capacity of the farmers can be enhanced through the provision of credit support. There is good scope to attempt a co-operative marketing system as in the case of other plantation crops in the states, such as rubber, coconut and arecanut. In the case of these crops, there are Co-operative Marketing Societies spread across the state and a similar attempt for pepper could be initiated also.

XII. PEPPER PHARMACOPOEIA The use of black pepper as a drug, apart from its wide spectrum use as a food additive, both in the Indian and Chinese systems of medicine is well documented (Atal et al., 1975; Nadkarni, 1976; Kurup et al., 1979). Pepper is elaborately described for its medicinal values in the ancient Indian system of medicine Ayurveda. It is described as katu (pungent), tikta (bitter), and ushnaveerya (potency). Further mention of pepper is also made in the control of the principal sources of diseases, as described in Ayurveda, namely, vata and kapha, in the ancient Indian language Sanskrit, the biological processes in the human system controlled by the CNS and the later formation and regulation of all ﬂuids that have a preservative quality, such as mucus, synovia, etc. Pepper is described as a drug which enhances digestive power in the body, improves appetite, cures colds, coughs, dyspnoea, diseases of the throat, intermittent fever, colic, dysentery, worm infestation, and also hemorrhoids. Pepper is also prescribed to relieve toothache, muscular pain, inﬂammation, leucoderma, and even

356

K. P. P. NAIR

epileptic ﬁts (Ayier and Kolammal, 1966; Kirtikar and Basu, 1975). Black pepper is called maricha or marica in Sanskrit, which implies its ability to dispel poison. In the Chinese system of medicine, pepper is used as an antidote against snake and scorpion bites. The above descriptions explain the diverse actions of pepper being used in the Indian system of medicine, either as such, or through its active ingredients in many formulations. In sharp contrast to the use of pepper in the Indian system of medicine, pepper ﬁnds no use in allopathy or homeopathy. The beneﬁcial effect of pepper in medical formulations can be attributed to piperine and other phenolic amides and essential oils. But in Ayurveda, the active ingredient based speciﬁc activities are not taken into consideration. Hence, the pharmacological and toxicological aspects of pepper and its constituent secondary metabolites were not studied. It must be, without doubt, mentioned that many of the clinical uses of pepper are time-tested, over several generations, and there is implicit faith among the Indian masses on the unquestionable medicinal value of pepper. However, to scientiﬁcally establish the value of pepper as a unique medicinal product, it is crucial to investigate, based on well deﬁned experimental protocols currently used in modern drug research and development, the pharmacological aspects of pepper and its active ingredients. Some of the preliminary investigations carried out over the last two decades in India on the above lines provide valuable information. Antipyretic, analgesic, antiinﬂammatory, antimicrobial and antineoplastic activities were reported after in vitro and in vivo investigations. Of late, there is a surge of interest in environmental friendly use of products to control crop pests and pepper has been found to possess insecticidal as well as insect repellent properties. Developments of new pepper based products as insecticides could be a boon to human use as they would be free from the toxic nature of most of the insecticides currently in use. Piperine is the major alkaloidal constituent of pepper. Systematic pharmacological studies on piperine have revealed its analgesic (pain relieving), antipyretic (fever relieving) and antiinﬂammatory properties, in addition to rejuvenation of the CNS. The antimicrobial activities are provided mainly by the essential oils. In addition to the ancient Indian system of medicine, Ayurveda, there are other native systems of medicines in India, such as Siddha and Yunani. Pepper is used in these as well. In the Chinese system of medicine, pepper and chenghan (Dichora febrifuga L.) are used in the treatment of malaria (Das et al., 1992). The active ingredients of chenghan are febrifugin and iso-febrifugin, both of which show a 50% higher antimalarial property than the conventional quinine. As an antipyretic, pepper is also used in the treatment of malaria (Nadkarni, 1976). The curative effects claimed in the cases mentioned above in the case of pepper can be attributed to the antipyretic and analgesic actions of the active ingredients. A strong antipyretic effect on typhoid vaccinated rabbits at a dose of 30 mg kg21 body weight after oral administration has been reported by Lee et al. (1984). In these studies, acetaminophen was used as the reference drug. The antipyretic action of

AGRONOMY AND ECONOMY OF BLACK PEPPER

357

piperine was found to be stronger than that of acetaminophen. To determine the analgesic action, acetaminophen and aminopyrine were used as reference compounds. Piperine showed a strong activity with an LD50 value of 3.7 mg kg21 dose using the writhing method and 104.7 mg kg21 using the tail-clip method.

A. ANTIINFLAMMATORY AND CENTRAL NERVOUS SYSTEM (CNS) DEPRESSANT ACTIVITY OF PEPPER Several researchers have reported the antiinﬂammatory property of pepper. Piperene showed a signiﬁcant inhibition of the increase in edema volule in a carragenin induced test with an oral intake of 50 mg kg21 body weight (Lee et al., 1984). The effect of different acute and chronic experimental models were studied by Mujumdar et al. (1990a). These researchers evaluated the mechanism of antiinﬂammatory activity by biochemical processes and concluded that piperine acted positively in the case of early acute changes in inﬂammatory processes. These studies have been further corroborated by those of Kapoor et al. (1993). In the case of epileptic patients, pepper is used in Ayurveda to induce sleep (Kirtikar and Basu, 1975). In the pharmacological investigations pepper and piperine were found to have a CNS depressant effect. Both pentetrazole seizure and maximal electroshock seizure were inhibited by pepper and piperine (Won et al., 1979). The LD50 value of 287.1 mg kg21 i.p. was found for piperine and a corresponding oral value was 1638.8 mg kg21. In another investigation Shin et al. (1980) showed that piperine at one-tenth its LD50 value had strong potentiating effect on hexobarbitol induced hypnosis in mice. Decreased passivity, ptotic symptoms and decrease in body temperature were also observed. Protection against electroshock seizure and a muscle relaxant effect were observed at a relatively low intraperitonial dosage range of an LD50 of 15.1 mg kg21 (Lee et al., 1984). This dosage appeared almost equipotent to the reference compound phenytoin. Petrol extract of pepper leaves was found to potentiate pheno barbitone induced hypnosis in mice (Sridharan et al., 1978). Majumdar et al. (1990b) have also reported that a high dosage of piperine potentiates the phenobarbitone sleeping time by inhibiting the liver microsomal enzymes. It also acted partially through stimulation of the pituitary adrenal axis. The above investigations clearly show that piperine has the capacity to be an effective ingredient of pepper in the treatment of “petit mal” (minor ailments). Antiepilepsirine (AE) isolated from white pepper is a compound of very great pharmacological interest. AE is 1-(3-benzodioxol-5yl)-1 (oxo-2-propenyl)piperidide. Perhaps this is the only compound derived from pepper which is clinically used in the treatment of human epilepsy. This is an alternative to dilantin therapy which is used in Chinese hospitals for the treatment of epilepsy (Ebenhoech and Spadaro, 1992). It is probable that AE may release 5-hydroxy

358

K. P. P. NAIR

tryptamine (5-HT) from nerve endings which can intensify the anticonvulsive state (Liu et al., 1984). Studies using rats as test animals show that AE increases 5-HT concentration in their brains which intensify the anticonvulsive state. AE also raises the tryptophan level in the brain which causes elevation of serontin and monoamine levels which leads to control of seizures.

B. EFFECT

ON

HEPATIC ENZYMES

While pepper or piperine positively affects liver function, as evidenced by the research of many workers, they do in no way cause hepatic toxicity. Piperine functions mainly as a chemopreventive substance through modulating enzyme function. Hepatic toxicity of piperine in rats was studied by estimating the hepatic mixed function oxidases and serum enzymes as speciﬁc markers of hepatotoxicity (Dalvi and Dalvi, 1991). An intragastric dose of 100 mg kg21 body weight resulted in an increase in hepatic microsomonal enzymes 24 h after treatment. In the experiment, cytochrome p-450, cytochrome b5, NADPH-cytochrome C reductase, benzphetamine N-demethylase, aminopyrine N-demethylase and aniline hydroxylase were estimated. An intraperitonial dose of 100 mg kg21 body weight did not produce any effect on the activity of the drug metabolizing enzymes. However, when the dose was enhanced an eightfold to 800 mg kg21 body weight as intragastric and 100 mg kg21 body weight as intra peritonial, a signiﬁcant decrease in the levels of the enzymes was noted. But these treatments did not affect those serum enzymes which are speciﬁc markers of toxic liver conditions. Piperine has been found to provide signiﬁcant protection against chemically induced hepatotoxicity. Both in vitro and in vivo lipid peroxidation and prevention of depletion of GSH and total thiols were observed (Kaul and Kapil, 1993). Liquid peroxidation causes the production of free radicals which, in turn, causes tissue damage. GSH conjugates xenobiotics which are excreted by subsequent glucuronidation. In the above cited study the hepato protective action of piperine was compared with a reference compound, silymarin, a known hepato protective drug, and it was found that piperine has slightly lower activity. In a feeding experiment using Swiss albino mice fed with a diet containing 1, 2, and 3% black pepper, w/w, for 10 and 20 days, a dose dependent increase in the levels of the hepatic biotransformation enzymes, namely, glutathione-stransferase, cytochrome p-450, cytochrome b-5 and acid soluble sulfhydril-SH were observed (Singh and Rao, 1993). Reen et al. (1993) in an in vitro study found lowered levels of glucuronidation due to the inhibition of the enzyme UDP-glucose dehydrogenase. Tripathi et al. (1979), while investigating the hypoglycaemic action of several plants, found that pepper fruits are devoid of any signiﬁcant hypoglycaemic action in rabbits. The aqueous extract of pepper leaves at a dose of 10 – 20 mg kg21 body weight led to a moderate increase in the blood pressure of dogs (Sridharan et al., 1978). Both piperine and AE have detoxifying

AGRONOMY AND ECONOMY OF BLACK PEPPER

359

qualities which may lead to an increase in the bioavailability of other drugs, as a result of which they alter the pharmokinetic parameter of the epileptic (Bano et al., 1969).

C. CARCINOGENIC AND MUTAGENIC EFFECTS BLACK PEPPER

OF

Positive effects of pepper as an anticarcinogenic and non-mutagenic source have been found by investigators. The Ames test shows its non-mutagenic quality. Chemical carcinogenesis has been found to be prevented by pepper because of its stimulatory effect on xenobiotic biotransformation enzymes. The antioxidant properties of piperine and associated unsaturated amides play a preventive role in carcinogenesis. Dietary intake of natural antioxidants can be a crucial aspect of building up the human body’s natural defense mechanisms against the onslaught of diseases, which can be caused by degradative changes brought about by mutagens. Additionally, the essential oil constituents in pepper inhibit DNA adduct formation by xenobiotics. This observation shows that pepper has anticarcinogenic potential. However, pepper extracts showed enhanced incidence of tumor formation in mice and an elevated level of DNA damage caused by piperine in cell culture investigations. Hexane, water and alcohol extracts of pepper were tested for mutagenecity on Salmonella typhimurium strains TA 98 and TA 100 by Ames assay and the results indicated that the extracts were non-mutagenic. This investigation provides evidence that water extract exerts an antimutagenic action on carcinogen induced mutagenesis (Higashimoto et al., 1993). The chemoprotective role of pepper as well as its constituents has been shown because of their positive effect on the activity of biotransformation enzymes in the liver in a dose-dependent manner (Singh and Rao, 1993). Nakatani et al. (1986) report a very signiﬁcant antioxidant activity due to the phenolic amides present in pepper. It is clearly established that antioxidants exert a preventive role in carcinogenesis. The modulating effects of the essential oil constituents of pepper were investigated by Hashim et al. (1994) who showed that they have an inhibitory effect on carcinogenesis. The volatile oil and its constituents suppress the formation of DNA adducts with aﬂatoxin B1. The microsomal enzymes modulate this action. Feeding powdered pepper to mice at 1.66 w/w dose showed no positive impact on carcinogenesis, while feeding and painting of the mice body with solvent extract of pepper, 2 mg for 3 days per week for three months, showed an increased incidence of tumor formation (Shwaireb et al., 1990). The activity of methylcholanthrene, a potent carcinogenic compound, was found to be reduced by a pepper terpenoid d-limonene (Wuba et al., 1992). Two minor constituents of pepper, safrole and tannic acid, have minor carcinogenic activity. In a tissue culture study using V-79 lung ﬁbroblast cell lines it was found that piperine treated cell lines showed

360

K. P. P. NAIR

increased DNA damage compared with untreated ones (Chu et al., 1994). Piperine treatment lowered the activity of the enzymes glutathione-s-transferase and uridine diphosphate glucuronyl transferase indicating cytotoxic potential. The in vivo formation of n-nitroso compounds from naturally occurring amines and amides contribute to the carcinogenic potential of certain foods and food additives. Piperine and other phenolic amides present in pepper are also known for their conversion to n-nitroso compounds in acidic conditions and, hence, are treated as carcinogenic (Lin, 1986). However, it can be inferred that the presence of a conjugated unsaturated system in the phenolic amide prevents the oxidation of the amide nitrogen to n-nitroso compounds to a great extent. Additionally, the essential oil constituents in pepper would also ensure DNA stability because of their anticarcinogenic potential. All of the above cited research results point to an ambivalent nature of pepper, both as an anticarcinogenic agent, and as a procarcinogenic agent. But, the overwhelming evidence indicates that it is more of the former than the latter. It is because of this belief that pepper has been a crucial ingredient of the Indian systems of medicine, primarily Ayurveda and to a lesser extent Siddha and Yunani, for centuries, and continues to be so even in the modern day.

D. PEPPER

AS AN

ANTIOXIDANT

Antioxidants are one of the most crucial biochemical compounds in the human system ensuring good health. They scavenge free radicals, which trigger many untoward biochemical reactions in the human system, and control lipid peroxidation in mammals. Of late, there is considerable interest in antioxidants. Lipid peroxidation is a chain reaction that is a constant source of free radicals, which initiate further peroxidation, which trigger deterioration of food, but also damage tissue, both of animal and plant origin, which in human beings cause many inﬂammatory diseases, ageing, atherosclerosis and also cancer. Many investigations reveal that pepper and its phenolic constituents, such as amides, possess good antioxidant properties. Tocopherol and vitamin-C are two important natural antioxidants. Chiapault et al. (1955) have investigated the antioxidant property of spices in a two-phase aqueous fat system. The investigations indicated that pepper has an antioxidant index of 6.1 while turmeric and clove show values of 29.6 and 103.0, respectively. The antioxidant property of pepper has been attributed to its tocopherol content (Saito and Asari, 1976). Revankar and Sen (1974), however, attribute the antioxidant property of pepper to its polyphenolic content. These authors investigated the effect of pepper oleoresin on ﬁsh oil and arrived at this conclusion. Abdel-Fattah and E1-Zeany (1979) observed that the autoxidation of unsaturated fatty acids and proteins was delayed by the addition of pepper and substantial protection against oxidative degradation was obtained. Nakatani et al. (1986), while investigating the family Piperaceae, established the

AGRONOMY AND ECONOMY OF BLACK PEPPER

361

fact that all the ﬁve phenolic amides present in P. nigrum possess very good antioxidant properties, which were found to be superior to the synthetic antioxidants such as butylated hydroxy toluene and butylated hydroxy anisole.

E. PEPPER

AS AN

ANTIMICROBIAL AGENT

Volatile oils, which are an active constituent of most spices, of which pepper is the most important, impart antiseptic, antibacterial and antifungal properties. The positive effect of volatile oils can be attributed to their terpenoid constituents. Pepper possesses both bactericidal and bacteriostatic properties. These properties aid in enhancing the shelf life of foods to which pepper has been added. Even the leaf extract of pepper possesses the antibacterial activity (Subramanyam et al., 1957). However, the extract was not inhibitory to the growth of E. coli, Aerobacter aerugenosa, Lactobacillus casei, Staphylococcus faecalis, S. aureus, and S. sonnei (Subramanyam et al.,1957). The essential oil obtained from pepper was found to be inhibitory to the growth of a penicillin-C resistant strain of S. aureus. Jain and Kar (1971) documented the inhibitory action of pepper oil on Vibrio cholerae, S. albus, Clostridium dipthereae, Shigella dysenteriae, Streptomyces faecalis, S. pyogenes, B. pumilis, B. subtlis, Micrococcus sp., Pseudomonas pyogenes, P. solanacearum and Salmonella typhimurium. The antibacterial action was determined by the agar well diffusion technique using cephazolin as standard (Pever and Anesini, 1994). The mycelial growth and aﬂatoxin synthesis by Aspergillus parasiticus were inhibited by pepper oil at a concentration of 0.2 –1% (Tantaoui and Beraoud, 1994). Antifungal activity against Candida albicans (Jain and Jain, 1972) was exhibited by pepper leaf oil. Rao and Nigam (1976) reported a similar effect of pepper leaf oil on A. ﬂavus. The antibacterial effects of pepper extract, essential oil and isolated piperine in vitro against sausage micro ﬂora, L. plantrum, Micrococcus specialis and Streptococcus faecalis have been reported by Salzer et al. (1977) in which the authors noted that only isolated piperine displayed microbial growth inhibiting effects at a normal dose. Pepper powder and extract were active only at high concentrations. An alcoholic extract of pepper was found active against the deadly food borne bacterium C. botulinum (Huhtanen, 1988). Another potential activity was revealed by a study carried out using tincture of pepper by Houghton et al. (1994). These authors investigated the antibacterial activity against nine strains of Mycobacterium tuberculosis and found that growth of all the nine strains was inhibited. The mode in which pepper acts positively is through its effect on enhancing bioavailability of administered medicaments, when they contain pepper or its active ingredients and uptake of proteins and amino acids from ingested food. In an investigation using Trikatu, which is an Ayurvedic preparation containing P. nigrum, P. longum and turmeric (Zinjiberus ofﬁcinale, another spice with great therapeutic value), it was observed that the pepper

362

K. P. P. NAIR

containing preparation enhances the bioavailability of other medicaments (Johri and Zutshi, 1992). At a dose of 25– 250 m mol21, piperine enhanced the uptake of 1-leucine, 1-iso leucine and 1-valine and increased lipid peroxidation in an in vitro study using intestinal epithelial cells of rats (Johri et al., 1992). Probably, piperine interacts internally to enhance the permeability of intestinal cells. Protein uptake from pulses was enhanced by up to 1.5% with the addition of pepper (Pradeep and Geervani, 1994). The study involved mainly two pulses in the Indian vegetarian diet, namely, winged bean and horsegram. The results clearly suggest that the Indian vegetarian diet, where spice is a common ingredient, has a decided advantage in human health.

F. THE PHARMACOLOGICAL EFFECT HUMAN HEALTH

OF

PEPPER

ON

The precise role of pepper on human health can only be understood through well-structured studies, which are only very limited in number to date. However, from the above it can be concluded that where extrapolations and comparisons can be made, pepper, as a whole product, or its active ingredients have been found to have a very positive role on human health. A brief description of the limited number of studies is given here. When used in high doses, gastric mucosal injury caused by pepper is comparable with that of aspirin. This observation was made in a double blind study of intragastric administration of pepper to human volunteers (Mayore et al., 1987). In this investigation, healthy human volunteers were given meals containing 1.5 g of pepper; 655 g of aspirin and distilled water were used as positive and negative controls. This is a short term study and it must be pointed out that long term effects of daily pepper intake, through food or medicaments, are unknown. Vezyuez et al. (1992) investigated the effect of intestinal peristalsis by measuring the orocaecal transit time (OCTT) utilizing a lactulose hydrogen breath test on healthy human subjects. They were given 1.5 g of pepper in gelatine capsules and the OCTT was measured on several days and it was found that OCTT increased signiﬁcantly after administration of pepper. This ﬁnding has great clinical importance in the management of various gastrointestinal tract disorders. An equally important investigation reveals the effectiveness of pepper extract volatiles in the treatment of cessation of smoking. Results of the investigation on human subjects reveal that cigarette substitutes, which deliver pepper volatile compounds, alleviated smoking withdrawal symptoms. The results of both the investigations detailed above have very great potential for further experimentation and, perhaps, pharmacological exploitation for further therapeutic beneﬁts for human health.

AGRONOMY AND ECONOMY OF BLACK PEPPER

G. CLINICAL APPLICATIONS

OF

363

PEPPER

It is only in the case of the Indian system of medicine, primarily Ayurveda, and to a lesser extent Siddha and Yunani, that pepper has been used in many clinical applications. No such use is made of pepper in allopathy. When clinically used, none of the active pepper ingredients are used as such. Dried black pepper with other medicinal plants, also in the dried form, are used in the preparation of speciﬁc formulations. The most widely used formulation in Ayurveda is Dasamulakatutrayadi Kashayam, a formulation that is used in the treatment of asthma. Then comes Ashtacurnam, which is used to get over dyspepsia, ﬂatulence, etc. The formulation also has stomachic and carminative effects. Amritharishtam is used as an antipyretic and also in the case of women with excessive menstrual bleeding. Muricadi Thailam is used as an antiinﬂammatory agent and also to relieve rheumatic pain. To cure coughs and bronchitis, Dasamularasayanam is used and Gulgulutiktaka-Ghrtam has both analgesic and antiinﬂammatory properties. Though pepper as such is widely used in the preparation of Ayurvedic medicines, its antipyretic, analgesic and antiinﬂammatory properties merit further research for use in allopathy, since, to date, pepper is not used in the preparation of any of the allopathic medicines. Its antioxidant and antimicrobial activities are also worth investigating further for possible use in the manufacture of allopathic medicines. Additionally, pepper has a good dietary value. It has a high ﬁber content (15 – 33%), iron (54 –62 mg g21), calcium (1 – 1.5%) and also appreciable amounts of essential amino acids (Uma Pradeep et al., 1993). As a good digestive, pepper enhances the secretion and ﬂow of salivary enzyme amylase.

H. TOXICOLOGICAL EFFECTS Published reports do not detail any toxic effects of pepper. This may be due to the relatively small amounts used in many medicinal formulations. Since, in most cases, the total contents of piperine and associated phenolic amides used would add up to just about 7 – 9%, w/w, and that of the volatile oils 2 –4%, which are negligible amounts, no untoward health hazard has been encountered so far. At this level, the actual doses of the different constituents available from the quantity of pepper powder, oleoresin or extractive used, will be very unlikely to trigger any toxic side effects in human body. In fact, the very pungency of piperine and strong ﬂavor of the volatile oils, act as deterrents against excessive human consumption. The FAO and WHO Experts Committee on Food Additives do not prescribe any limit to acceptable intake of piperine and the volatile oils. The major untoward consequence of pepper use is the gastric mucosal injury at a dose of 1.5 g kg21 of food. It enhances the DNA adduct formation. The extract of pepper produces enhanced cancer incidence in mice. When mice were fed with

364

K. P. P. NAIR

extract of black pepper it resulted in the formation of heptocellular carcinoma, lymphosarcoma and ﬁbrosarcoma (El Mofty et al., 1991). However, a large number of investigations show that pepper has anticarcinogenic attributes. Before conclusive evidence can be arrived at, whether or not pepper and its active ingredients trigger anticarcinogenic or procarcinogenic reactions in human beings, much more detailed scientiﬁc scrutiny needs to be carried out. As of now, such studies are conspicuously lacking. The observation that pepper enhances the liver microsomal enzyme activity must also be investigated further. This will help decide whether or not constant use of pepper in the human diet leads to production of non-speciﬁc enzyme induction. Pepper is a universal dietary component on the Indian subcontinent and the quantity used is much more than used anywhere else in the world. Such information has very vital implications of health for a population that currently stands at more than a billion.

I. THE INSECTICIDAL ACTIVITY

OF

PEPPER

The primary advantage of using plant derived insecticides in agriculture is that while the product is toxic to the pest in question, it is non-toxic to human beings, including those who handle it in the ﬁeld, unlike most other chemically produced commercial insecticides, which have come under attack from the environmentalists. Plant derived products have been used since time immemorial both in agriculture and for domestic purposes. The extract of pepper, volatile oil components, and the different phenolic amides present in pepper have shown insecticidal, insect repelling, and ovicidal activities to various plant insects and pests harmful to human beings. The major alkaloid of pepper, namely piperine, is more toxic than pyrethrine, a standard insecticide for house ﬂies (Harvill et al., 1943). In a number of instances the insecticidal or insect repellent activity was obtained at low concentrations. Hence, this aspect needs to be further probed for positive ﬁeld application. The volatile oil of pepper was found to cause mortality of the cigarette beetle (Lasioderma serricorne) (Samuel et al., 1984). Several groups of researchers investigated the effectiveness of the alcohol extract of pepper and found it effective against cotton boll weevil, rice weevil and Drosophila (Su, 1977; Scott and McKibben, 1978; Barakat et al., 1985). The LD50 value for the tropical application of the crude and puriﬁed extracts, obtained after 24 h showed a mortality rate of 3.4 and 4.8 mg, respectively, per adult insect of Sitophilus oryzae and 4.5 and 7.2 mg, respectively, per adult insect of Callosobrunchus maculatus. The oleoresin, piperine and other amides in pepper were found to be lethal to culex mosquito (Culex pipiens) and also to pulse weevil (Su, 1977). Pepper amides, such as piperonal, piperine, piperoline, pellitorine, pepercide and dihydropipercide, also were investigated by Miyakado et al. (1979), Masakazu et al. (1980) and Deshmukh et al. (1982) for their effectiveness against insects and were found to be effective against pulse beetle, rice weevil,

AGRONOMY AND ECONOMY OF BLACK PEPPER

365

and Drosophila, where larval growth inhibition was found. Desai et al. (1997) showed that 100% mortality within 24 h at a dose of 80 ppm was obtained using acetone extract of pepper in the case of Anopheles subpietus larvae.

XIII.

CONSUMER PRODUCTS OUT OF BLACK PEPPER

Of all the spices used in the world, black pepper, is, perhaps, the most widely used one in the kitchen, perfumery, medicine, and industry. Most of these consumer products that have their origin in black pepper are the “value added” ones. For instance, when pepper is used in the food processing industry, it is not pepper as such, but the oleoresin extracted through solvent extraction and its pungency and ﬂavor are the ones that add to the value of the end product made by the food processing industry.

XIV. VALUE ADDITION IN PEPPER There has been a dramatic advancement in the ﬁeld of value addition of black pepper and diversiﬁcation of processed pepper products. The value added pepper products are classiﬁed as follows: 1. Green pepper based products 2. Black and white pepper based products 3. Pepper by-products The above may be further classiﬁed as follows (Pruthi, 1997).

1. a. b. c. d. e. f. g. h. i.

Green Pepper Based Products

Canned green pepper in brine, Bottled green pepper in brine, Bulk packed green pepper in brine, Cured green pepper (without any covering tissue), Frozen green pepper, Freeze-dried green pepper, Sun dried or dehydrated green pepper, Green pepper pickles in oil, vinegar or brine, Green pepper mixed pickle in oil, vinegar or brine,

366

j. k.

K. P. P. NAIR

Green pepper ﬂavored products, and Green pepper paste. 2.

a. b. c. d. e. f. g. h.

Black and White Pepper Based Products

Black pepper powder, White pepper powder, White pepper whole, Pepper oleoresin, Pepper oil, Other pepper products, By products from pepper waste, Miscellaneous forms of utilization, such as in medicine, culinary, and industry.

The Pepper Marketing Board of Malaysia (PMBM), as well as many privately owned industrial houses have brought to the market a number of innovative end products of pepper and have facilitated planning and product development of black pepper. The following structure has been proposed by PMBM (Abdulla, 1997). The pepper products are whole black pepper in retail packs for table use, ground pepper in retail packs and dispensers. Pepper based products are ﬂavored ground pepper, such as lemon pepper, garlic pepper, sauces, and marinades, which have pepper as the primary component and also pepper paste. Spice mixtures and blends are curry powders and spice blends for speciﬁc cuisines, such as, ﬁve-spice powder, soup blends, etc. In addition to the above, there are pepper ﬂavored products, which are pepper mayonnaise, pepper tofu, pepper cookies, pepper keropok (which are prawn or ﬁsh crackers). Products that use pepper extracts are pepper candy, pepper perfume, etc. Pepper is also put to other auxiliary uses, such as in the case of pepper stalk, which is used as a substrate for growing mushrooms. There are a number of common black pepper products of which the most important is ground pepper. This is the exclusive form in which black pepper is used on the table. It is also the only spice served along with food on board and also used in restaurants specializing in fast foods. Pepper is very commonly used for seasoning food at different stages of preparation. The following factors have to be kept in mind in the case of ground pepper use, which is the easiest to manufacture and market. Freedom from bacteria and mould is the most important, as otherwise it is a potential health hazard. During the grinding process, volatile oil content can be adversely affected. It should be ensured that it is not so. Since high moisture content in the berries will adversely affect shelf life, moisture during storage should be kept to a minimum. Optimum particle-size should be ensured for free ﬂow during the period of storage and active use.

AGRONOMY AND ECONOMY OF BLACK PEPPER

367

Another very important aspect is packaging. Without good packaging, the product will not only not have a good market value but also can deteriorate during storage. Above all, marketable pepper should be free from all extraneous matter. Flavored pepper is an important product with good market for culinary purposes. Lemon and garlic pepper are examples. The former contains lime powder, pepper powder and salt and the latter is a blend of dehydrated garlic powder and dried pepper powder. Both are table condiments, especially in nonvegetarian dishes, containing chicken, ﬁsh, etc. The PMBM has developed these products. Pepper sauce is another product of PMBM which can be used as a marinade especially in meat-based dishes, such as steak, lamb, etc., and is also used in stir-frying. The preparation of this product involves the use of black pepper, soybean extract, garlic and other ingredients to obtain a relatively mild dark sauce. Pepper is also used extensively in the preparation of a variety of sauces including Worcestershire sauce. Spice mixtures and blends are extensively used in continental kitchens as well as in regional kitchens. Pepper based spice mixes, available in departmental stores and supermarkets, are widely used in different meat based preparations. There are also soup mixes which contain pepper. Private industrial houses and PMBM have developed a variety of pepper ﬂavored products, such as pepper ﬂavored mayonnaise, egg tofu, savory pepper cookies, traditional biscuits with pepper ﬂavor, pepper ﬂavored prawn and ﬁsh crackers, etc. Many traditional dishes and snacks, when dressed with pepper will result in enhanced taste and ﬂavor. Pepper ﬂavored vinegar and pepper salts are excellent taste enhancers. Pepper is widely used in industrially manufactured food items. Almost all ﬁsh and meat based products use pepper for seasoning, which enhances both ﬂavor and pungency. Industrially manufactured foods, such as soups, pickles, etc., use pepper extracts or powdered pepper. Additionally, there are pepper ﬂavored beverages, such as, pepper tea, pepper coffee, pepper ﬂavored milk, etc. Some brands of brandy contain piperine and pepper oil, which impart a pungent taste and an “exotic” ﬂavor. In the perfumery industry, volatile oil from pepper is used to impart an exclusive spicy “oriental” touch. Some of the more popular brands of perfume that contain pepper volatile oil are Revlon’s “Charlie” and Christian Dior’s “Poison” and Malaysian perfumes, such as “Sensai” and “Amila” (Ng, 1993).

3.

Green Pepper Based Products

Pepper fruits that are not fully mature are used in the manufacture of green pepper based products, such as pepper in brine, pepper in oil, pepper in vinegar, desiccated green pepper, freeze-dried green pepper, pepper paste, etc. Almost all of the green pepper products are used by the catering sector to be served with meat dishes, such as steak and pork. The food industry also uses green pepper in the production of certain specialized cheeses. The other green pepper based products

368

K. P. P. NAIR

are pickled green pepper and pepper spike. Tender pepper spikes and fruits alone or in combination with tender cardamom, pickled in vinegar and salt or sugar make very delicious dishes which can be served with dinner. Green pepper paste is another product. The green pepper paste, in polypacks or bottles, are now a common sight in supermarkets. This can be used as a substitute to pepper powder and it gives a refreshing taste with ﬂavor and a unique “bite”, which enhances the palatability of ﬁsh and meat dishes, and is a common choice of well trained chefs.

4. Other Ancillary Pepper Based By-products An industrial area worth greater research and development is the ancillary pepper-based by-products, which are of uncommon use. For instance, Malaysia has succeeded in producing paper and board from pepper stalks. These materials, especially paper made of pepper stalk make good invitation cards, having a unique texture and color pattern (Ng, 1993). Pepper stalk makes a good substrate for oyster cultivation when grown in a 1:1 mixture of pepper stalk and shredded paper (Siti Hajijah and Bong, 1993). In Malaysia, the remnants of the pepper processing industry, which consists mainly of stalk and pericarp, are powdered and used as an organic manure. This is called pepper dust. It can enrich soil fertility when used in combination with other organic agricultural wastes (Ng, 1993). In addition to the above, there are a number of other pepper-based commercial formulations. Though a number of commercial formulations, such as canned green pepper, green pepper in brine, pepper oil, oleoresin, etc., are readily available in the market, they still are to ﬁnd widespread acceptability by the consumer. The most recent use of the spice processing technology is the encapsulation of the ﬂavoring components, namely, spice oils and oleoresin. These can be fully exploited commercially on a large scale, provided capital and entrepreneurship are readily available. This is yet to happen. There is now a changing consumer preference for white pepper from black pepper in some western countries. White pepper is used both as table pepper and also for enriching the ﬂavor and taste of crab soups, sea food salads, casseroles, chicken, ﬁsh, and egg preparations. Sauces like mayonnaise also contain white pepper as an ingredient. One of the primary reasons for a shift in preference to white pepper from black pepper in western countries is because the black specks from the skin of black pepper are unseemly to the eye on a dining table. The pepper product that is becoming increasingly popular in European countries is the green pepper. As the piperine and pepper oil levels are high in slightly immature corns, they are preferred in the preparation of green pepper. This gives a characteristic ﬂavor and is soft with a “bite” — a unique feel on the palate. These qualities make it ideal for garnishing meat dishes. The major products are green pepper in brine, canned green pepper, dehydrated green

AGRONOMY AND ECONOMY OF BLACK PEPPER

369

pepper, freeze-dried green pepper, green pepper paste, etc. While an increasing number of consumers prefer green pepper in brine and green pepper paste, and their demand is on the increase, demand for canned pepper is not much as its cost of production is quite high. India, Brazil, and Malaysia are the major producers of these value added pepper products. The catering industry is the main source of demand, followed by the food manufacturing industry for these products. Among the ﬂavoring agents, the most important are oleoresin, pepper oil, and encapsulated pepper. The most extensive use of oleoresin is in the ﬂavoring of meat. The other end uses of oleoresin are for preparing pickles, sauces, gravies, dressing “chutneys” (unique Indian spicy eatables), soups, and snacks. In most meat preparations pepper is used to impart ﬂavor. The same task is accomplished by oleoresin as well. One of the most recent advances made by the Regional Research Laboratory in Trivandrum, Kerala State, India, is in developing the spray drying encapsulation technique for oleoresin. The encapsulated ﬂavor powder is used in a variety of products, such as cake mixes, dry beverage mixes, desserts, soup mixes, dusting on potato chips, and nuts. These have an emerging market all over the world.

XV. CONCLUSIONS AND A PEEP INTO PEPPER’S FUTURE Pepper has certainly had a checkered history. Within the past, present and future, pepper can indeed be termed the “King of Spices”. Though beset by many problems, both economic and agronomic, it is a safe bet that pepper will emerge as the world’s most sought after spice. It has been estimated that by the year 2020 global demand for pepper would be about 28,0000 metric tons, which is projected to escalate to 360,000 metric tons by the year 2050. This would entail almost doubling the present production in the ﬁrst half century of the current millennium. Where will the additional output come from? Increase of pepper area could only be marginal on a global basis. India, the major producer, will, perhaps, experience only marginal area increases. On the Asian subcontinent, it is only in China and Vietnam, the latter already emerging as a key producer, where marginal area increases can be expected. Like India, only a marginal increase in area can be expected in Indonesia. Malaysia and Thailand, the other two pepper producers of Asia, are already experiencing a decrease in area. This is also true for Brazil, a key producer on the Latin American continent. Labor is becoming increasingly difﬁcult to obtain and very expensive when found and, hence, countries such as Malaysia and Thailand on the Asian continent and Brazil on the Latin American continent are moving away from the labor intensive pepper production. The impact of globalization is clearly seen in these countries, where market forces are shying away from such labor intensive crops such as pepper in

370

K. P. P. NAIR

preference to more mechanically managed crops. By the very nature of the pepper plant and the pepper canopy, mechanization has only very limited possibilities. From the foregoing review, it can clearly be seen that one of the most pressing needs of pepper production is to have a pepper ideotype that combines many positive attributes in terms of boosting production potential, at the same time, resisting many ravages of nature, especially those of diseases, among which the most predominant is Foot Rot, caused by the Phytophthora fungi. There is yet no single pepper variety that is universally resistant to this dreaded disease, though some lines have been identiﬁed which have shown regional promise. Conventional breeding, which will take years to accomplish, since pepper is a long duration perennial, will not be the answer to resolve this major problem. Biotechnology is a promising area, but there are vast gaps in knowledge. Behavior of the pepper plant at the molecular level is far from clearly understood. Though tissue culture is a promising area in terms of cutting down the long gap in producing a desirable ideotype, this will not give ﬁnal answers because there are far too many gray area variables in the technique. Gene sequencing, currently a very highly sophisticated technology with very high levels of investments, opens up a door in the case of pepper as well. At the current level of global scientiﬁc knowledge and available technical expertise in the case of pepper, it is a safe bet to say that, for whom pepper is a great fascination among crops, it is a very long road to traverse. Sophisticated instrumentation is only part of the answer, with the availability of superlative technical skills being of greater importance, which the pepper producing countries, India included, clearly lack. What one might expect on the pepper front, at least for the few decades in the current century to come, is the following. As science advances and global populations become more and more health conscious, the demand for organically grown food would enlarge. One of the most promising areas for future pepper development will be in the area of producing “organically grown” or what one might term “clean pepper”. The environmental scare generated by the so-called “green revolution”, the hall mark of which has been extensive and indiscriminate use of chemical fertilizers and pesticides, is driving more and more people to organically grown food and the demand for organically grown pepper is no exception. The potential market for organically grown food products, including pepper, is most promising in the USA followed by Germany. The rigors of growing organic foods are encompassed in the Codex Alimentaris stipulations developed by France and the USA has its own stipulations. In a country like India where pepper production on the farm and its further processing is not subject to the rigors of quality control, export of organically grown pepper can only be put on sound footing if production and processing methods are well streamlined and rigorously controlled for quality, sanitation being the most important. The same would hold good inasmuch as the prevalent

AGRONOMY AND ECONOMY OF BLACK PEPPER

371

situations in other pepper producing countries like Indonesia, Vietnam, Malaysia, Thailand, etc., are concerned. Another area which has much promise is the development of pepper derived food additives. The potential of this is gradually unfolding on an industrial scale. Heavy investments are necessary to make the various products price competitive. Some regional success stories do exist, though. For instance, the pepper based products, which have a ready regional market. Yet, these products do not have a global reach. The potential market for these products would be Europe and the USA, where a gradual change in food habits is taking place. Both post-harvest technology and product development would undergo remarkable changes in the future. Imaginative product diversiﬁcation would lead to the development of an array of novel products for the consumer, currently unheard of. Pepper with high ﬂavor, but low pungency, will begin to be increasingly used in the manufacture of liquors, fruit juices, bakery products and even choice perfumes. A possible challenge to the development of pepper varieties with high ﬂavor will be the creation of the in vitro pepper ﬂavor from bioprocessors. If this comes about, the only way natural pepper can be salvaged is a preference of the consumer for the clean, natural spice, rather than the “laboratory created spice”. Another important area that needs to be thoroughly researched is the management of soil fertility of pepper soils. Almost the entire approach to fertility management in pepper soils is based on classical textbook knowledge, where empirical recommendations generated from micro plots are extrapolated to large-scale ﬁeld conditions. These experimental micro plots are nothing but artefacts and many of the recommendations which emanate from such studies, when applied to large-scale plantations, turn out to be quite off the mark in reproducing accountable results. This often shatters farmers’ conﬁdence. A very signiﬁcant departure is being made after the concerted efforts of the author, who has developed an entirely new approach to soil testing and fertilizer management, based on nutrient buffering. The concept is now known, universally, as “the nutrient buffer power concept”. A detailed discussion on the concept and its relevance in pepper nutrition is given in Chapter IV. In India, in the state of Kerala, where pepper is grown to the largest extent, it has been observed that fertilizer input can be signiﬁcantly reduced by taking into consideration the buffer power of the nutrient concerned. Experimental and ﬁeld results (Nair, 2002) show that, as against the routine soil testing and fertilizer recommendations, addition of Zn fertilizer, a very crucial input in pepper production, can be reduced by almost 75% in some soils. There is an urgent need to extend the concept to other crucial nutrients, such as phosphorus and potassium, and possibly nitrogen as well. Of the ﬁrst two, phosphorus is more important in view of the fact that soils of Kerala State, the home of pepper, are lateritic and much of the applied phosphate fertilizer,

372

K. P. P. NAIR

based on routine soil testing is not available to the crops. It is in this context that this new concept holds out much hope for pepper farming. However, it must be emphasized that the success of a new approach, to a great extent, rests with the ingenuity of those applying it to suit the demands of a new situation. This principle is no exception to making the nutrient buffer power concept succeed in the case of pepper production, as has been the case with other crops, such as, maize, rye, white clover, and even a perennial crop like cardamom, tested by the author (Nair et al., 1997). The fact that pepper is a perennial crop makes it all the more important because, unlike an annual one, where a mid course correction can be effected in the following season, the fertilizer regime has to be correct from the very beginning, because pepper grows for upwards of 25 years. Unlike routine soil testing, the new approach calls for an accurate determination of the buffer power of the nutrient in question at the very start. Once this is accomplished, the buffer power factor can be integrated into the computations with the routine soil test data and accurate fertilizer recommendations can be made on the basis of this new information. This implies that, in addition to obtaining routine soil test data, one also needs to know the buffer power. The author has obtained very encouraging results with the new concept in pepper production in Kerala with regard to Zn fertilization, which signals a very promising change in the fertilizer practices. Hopefully, the new concept could very successfully be extended to other important plant nutrients. Agriculture is the engine for development, nationally and globally, as is being increasingly proved against the unfolding scenario of globalization. A nation with a strong agricultural base can be expected to be in the forefront of economic development. Pepper is the crop of the tropics and, in that sense, is very much a part of the development of the third world economy. A lot has been achieved in pepper production, but, before it moves to center stage, as a commanding crop of third world economy, much more needs to be done. This calls for the concerted efforts of not just agronomists, soil scientists or plant breeders dealing with pepper, but a whole range of visionary thinkers and planners, who can really make the crop, not just the king of spices, but the monarch of third world agricultural economy.

ACKNOWLEDGMENT With love I dedicate this chapter to my wife, Pankajam, my all, who sustains me in this very difﬁcult journey, that life is.

AGRONOMY AND ECONOMY OF BLACK PEPPER

373

REFERENCES Abdel-Fattah, I. E., and EI-Zeany, B. A. (1979). Effect of spices on the autooxidation of fatty foods. Rev. Ital. Soztanze Grasse. 56, 441 –443. Abdulla, A. (1997). Product development and promotion of pepper products. In “Proceedings of the 22nd Pepper Tech. Meet”, pp. 1–6, Cochin, Kerala State, India. Adams, F. (1971). Ionic concentrations and activities in soil solutions. Soil Sci. Soc. Am. Proc. 35, 420 –426. Adzemi, M. A., Hamdan, J., and Sjahril, J. S. (1993). Micro-nutrient removal studies on black pepper (Piper nigrum L.) in Sarawak. In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. P. Ipor, Eds.), pp. 104– 109. Universiti Pertanian Malaysia, Sarawak. Agarwal, S. (1988). Shoot tip culture of pepper and its micropropagation. Current Sci. 57, 1347–1348. Ahmed, K. (1993). The effect of Nitrogen on the growth of Piper nigrum L. planted with Desmodium trifolium cover. In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. P. Ipor, Eds.), pp. 97–105. Universiti Pertanian Malaysia, Sarawak. Anandaraj, M. (1997). Ecology of Phytophthora capsici (Leonian 1922, emend A. Alizadeh and P.H. Tsao) In “Causal Organism of Foot Rot Disease of Black Pepper (Piper nigrum L.)”, p. 154. Ph. D. Thesis, University of Calicut, Kerala State, India. Anandaraj, M., and Peter, K. V. (1996). In “Biological Control in Spices”, p. 52. Indian Institute of Spices Research, Calicut, Kerala State, India. Anandaraj, M., and Sarma, Y. R. (1994a). Effect of Vesicular Arbuscular Mycorrhiza on rooting of black pepper (Piper nigrum L.). J. Spices Aromatic Crops. 3, 39– 42. Anandaraj, M., and Sarma, Y. R. (1994b). Biological control of black pepper diseases. Indian Cocoa, Aracanut Spices J. 18, 22–23. Anandaraj, M., and Sarma, Y. R. (1995). Diseases of black pepper (Piper nigrum L.) and their management. J. Spices Aromatic Crops 4, 17–23. Anandaraj, M., and Sarma, Y. R. (1997). Mature coconut water for mass culture of biocontrol agents. J. Plantation Crops. 25, 112 –114. Anandaraj, M., Sarma, Y. R., and Ramachandran, N. (1994). Phytophthora root rot of black pepper in relation to age of the host and its culmination in foot rot. Indian Phytopath. 47, 203–206. Annual Report. (1997). “Indian Institute of Spices Research-Vision 2020”. IISR Perspective Plan, Calicut, Kerala State, India. Anon. (1989). “Technology transfer document of microencapsulated spice ﬂovours Regional Research Laboratory”, pp. 9–10. Council of Scientiﬁc and Industrial Research, New Delhi, India, Trivandrum, India. Anon. (1995). Research Branch, Department of Agriculture, Sarawak, Malaysia, Annual Report. Anon. (1996). Israel — A promising market for Indian spices. Spice India 9, 2–3. Anon. (1997). Spice export grows to an all time record. Spice India 10, 2–4. Apun, K., Siti Hajijah, A. S., and Yusof, I. M. (1993). Microbiological properties of pepper during processing. In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. P. Ipor, Eds.), pp. 337 –341. Universiti Pertanian Malaysia, Sarawak. Atal, C. K., Dhar, K. L., and Singh, J. (1975). The Chemistry of Indian Piper. Lloydia 38, 250–264. Ayier, K. N., and Kolammal, M. (1966). “Pharmacognosy of Ayurvedic Plants”, 9, p. 63. Kerala University, Trivandrum, Kerala State, India. Bade, J., and Smit, H. P. (1992). An annual model of the world pepper economy, p. 63. Seminar on Commodity Analysis, Jakarta, Indonesia, ESI-VU, Amsterdam. Balasubramanyam, N., Mahadevan, B., and Anandaswamy, B. (1978). Packaging and storage studies on ground block pepper (Piper nigrum L.) in ﬂexible consumer packages. Indian Spices 15(4), 6 –11.

374

K. P. P. NAIR

Bano, G., Alma, V., Raina, U., and Zutshi, V. (1969). The effect of piperine on the pharmokinetics of phenytoin in healthy volunteers. Planta Medica 12, 566–569. Barakat, A. A., Family, H. S. M., Kandil, M. A., and Ebrahim, N. M. M. (1985). Toxicity of the extract of black pepper, cumin, fennel, chamomile and Lupin against Drosophila ceratetis and Spodoptera. Indian J. Agric. Sci. 55, 116 –120. Barber, S. A. (1984). “Soil nutrient bioavailability: a mechnistic approach”. Wiley, New York. Barber, S. A. (1995). “Soil nutrient bioavailability: a mechnistic approach”. Wiley, New York. Bavappa, K. V. A., and Gurusinghae (1978). Rapid multiplication of black pepper for commercial planting. J. Plantation Crops 6, 92– 95. Bergman, A. (1940). De pepper culture en-handel op Bangka. Landbouw 16, 139–256. Biessar, S. (1969). Plant parasitic nematodes of crops in Guyana. PANS 15, 74–75. Bopaiah, B. M., and Khader, K. B. A. (1989). Effect of bio-fertilizers on growth of black pepper (Piper nigrum L.). Indian J. Agric. Sci. 59, 682–683. Bridge, J. (1978). Plant Nematodes Associated with Cloves and Black Pepper in Sumatra and Bangka, Indonesia. In “O.D.M. Technical Report”, p. 19. U.K. Ministry of Overseas Development, U.K. Broome, O. C., and Zimmerman, R. N. (1978). In vitro propagation of black pepper. Hortic. Sci. 43, 151–153. Burger, W. C. (1972). The evolutionary trends in the Central American species of Piper. Brittonia. 24, 356–362. Burger, K., and Smit, H. P. (1997). The Natural Rubber Market, Woodhead Publishing Limited, Cambridge, p. 354. Chaudhary, R., and Chandel, K. P. S. (1994). Germination studies and cryo preservation of seeds of black pepper (Piper nigrum L.), a recalicitrant species. Cryo-Lett. 15, 145–150. Cheriankunju, N. E. (1996a). Know your market for spices — Japan. Spice India 9, 5–7. Cheriankunju, N. E. (1996b). Know your market for spices — USA. Spice India 9, 2–5. Chiapault, J. R., Mizuna, G. R., Hawkins, J. M., and Lundberg, W. O. (1955). The antioxident properties of spices in oil-in-water emulsions. Food Res. 20, 443–451. Christie, J. R. (1959). “Plant Nematodes — Their Bionomics and Control”, p. 256. University of Florida, Gainesville, Florida, USA. Chu, C. V., Chang, J. P., and Wang, C. J. (1994). Modulatory effect of piperine on benzopyrene cytotoxicity and DNA adduct formation to V-79 lung ﬁbroblast cells. Food Chem. Toxicol. 39, 373–377. Chua, B. K. (1981). Studies on in vitro propagation of black pepper (Piper nigrum L.). MARDI Res. Bull. 8, 155 –162. Claassen, N., and Barber, S. A. (1976). Simulation model for nutrient uptake from soil by growing plant root system. Agron. J. 68, 961–964. Claassen, N., Hendriks, K., and Jungk, A. (1981). Rubidium — Verarmung des wurzelnahen Bodens durch Maispﬂanzen. Z. Pﬂanzenern. Bodenk. 144, 533 –545. Coffey, M. D. (1991). Strategies for the integrated control of soil borne Phytophthora species. In “Phytophthora” (J. A. Lucas, R. C. Shattock, D. S. Shaw, and L. R. Cooke, Eds.), pp. 411– 432. Cambridge University Press. Collins, L., and Lapieree, D. (1976). Freedom at Midnight. Vikas Publication House, New Delhi, India. Cook, R. J., and Baker, R. (1983). “The Nature and Practice of Biological Control of Plant Pathogens”, p. 53. American Phytopathological Soc., St. Paul, Minnesota, U.S.A. Dalvi, P. R., and Dalvi, P. S. (1991). Comparison of the effect of piperine administered intragastrically and intraperitonially on the liver and liver mixed function oxidases in rats. Drug Metabol. Drug Interact. 9, 23–30. Darrington, H. (1991). Herbs and spices. Food Manuf. 66(11), 21– 23. Das, T. P. M., and Cheeran, A. (1986). Infectivity of Phytophthora species on cash crops in Kerala. Agric. Res. J. Kerala 24, 7– 13.

AGRONOMY AND ECONOMY OF BLACK PEPPER

375

Das, P. C., Das, A., Mandal, S., and Chatterjee, A. (1992). On the validity of the ethnic use of P. nigrum and M. indica L. In “Proceedings of the International Seminar on Traditional Medicine”, p. 84. Calcutta, India. Datta, P. K. (1984). “Studies on two Phytophthora diseases (Koleroga of arecanut and black pepper wilt) in Shimoga District, Karnataka State, India”, p. 121. Ph.D. University of Agricultural Sciences, College of Agriculture, Dharwad, Karnataka State, India. Davis, R. M., and Menge, J. A. (1981). Phytophthora parasitica inoculation and intensity of vesicular–arbuscular mycorrhizae in citrus. New Phytol. 87, 705 –715. Davis, R. M., Menge, J. A., and Zentmyer, G. A. (1978). Inﬂuence of vesicular arbuscular mycorrhiza on Phytophthora root rot of three crop plants. Phytopathology 68, 1614–1617. De Waard, P. W. F. (1964). Pepper cultivation in Sarawak. World Crops 16, 24–30. De Waard, P. W. F. (1969). “Foliar diagnosis, nutrition and yield stability of black pepper (Piper nigrum L) in Sarawak”, p. 150. Comm. No. 58, Royal Tropical Institute, Amsterdam, The Netherlands. De Waard, P. W. F. (1979). Yellow disease complex in black pepper on the island of Bangka, Indonesia. J. Plantation Crops 7, 42 –49. De Waard, P. W. F. (1984). Piper nigrum L. In “Plant Resources of South East Asia” (E. West Phal and P. C. M Jansen, Eds.), pp. 225–230. Rijksherbarium, The Netherlands. Debrauwere, J., and Verzele, M. (1975a). Constituents of pepper VI. Oxygenated fraction of pepper essential oil. Bull. Soc. Chem. Belg. 84, 167 –177. Debrauwere, J., and Verzele, M. (1975b). New Constituents of the oxygenated fraction of pepper essential oil. J. Sci. Food Agric. 26, 1887–1894. Debrauwere, J., and Verzele, M. (1976). The hydrocarbons of pepper essential oil. J. Chrom. Sci. 14, 296 –298. Deciyanto, S., and Suprapto (1992). Research progress on important insect pests of black pepper in Indonesia. In “Proceedings of the International Workshop on Black Pepper Diseases 3–5 December, 1991, Bandar Lampung, Indonesia” (P. Wahid, D. Sitepu, D. Deciyanto, and U. Suparman, Eds.), pp. 237 –244. Research Institute for Spice and Medicinal Crops, Bogor, Indonesia. Dehne, H. W. (1982). Interaction between vesicular–arbuscular mycorrhizal fungi and plant pathogens. Phytopathology 72, 1115–1119. Desai, A. E., Ladhe, R. U., and Deorag, B. M. (1997). Larvicidal activity of acetone extract of Piper nigrum seeds to a major vector, Anopheles subpictus from Nashik region. Geobios 24, 13– 16. Deshmukh, P. B., Chavan, S. R., and Renapurkar, D. M. (1982). A study of insecticidal activity of twenty indigenous plants. Pesticides 16, 7. Devasahayam, S., and Anandaraj, M. (1997). Black pepper (Piper nigrum L.) research. In “Neem in Sustainable Agriculture” (S. S. Narwal, P. Tauro, and S. S. Bisla, Eds.), pp. 117 –122. Scientiﬁc Publishers, Jodhpur. Devasahayam, S., and Koya, K. M. A. (1994). Natural enemies of major insect pests of black pepper (Piper nigrum L.) in India. J. Spices Aromatic Crops 3, 50–55. Devasahayam, S., and Leela, N. K. (1997). Evaluation of plant products for antifeedant activity against pollu beetle (Longitarsus nigripennis Motschulsky), a major pest of black pepper. In “Abst. First Natn. Sym. Pest Management in Horticultural Crops”, p. 85. Environmental Implications and Thrusts, Bangalore. Devasahayam, S., Premkumar, T., and Koya, K. M. A. (1988). Insect pests of black pepper Piper nigrum L. in India — a review. J. Plantation Crops 16, 1 –11. Drew, M. C., and Lynch, J. M. (1980). Soil anaerobiosis, microorganisms and root function. Ann. Rev. Phytopath. 18, 37– 67. Drew, M. C., Nye, P. H., and Vaidyanathan, L. V. (1969). The supply of nutrient ions by diffusion to plant roots in soil. I. Absorption of potassium by cylindrical roots of onion and leek. Plant Soil 30, 252 –270.

376

K. P. P. NAIR

Dudek, D. H. (1996). “Sterilization method and apparatus for spices and herbs”, US Patent US 5523053 [US 260068 (940615)]. Eapen, S. J., and Ramana, K. V. (1996). Biological control of plant parasitic nematodes of spices. In “Biological Control in Spices” (M. Anandaraj and K. V. Peter, Eds.), pp. 20 –32. Indian Institute of Spices Research, Calicut, Kerala State, India. Eapen, S. J., Ramana, K. V., and Sarma, Y. R. (1997). Evaluation of Psuedomonas ﬂuorescens isolates for control of Meloidogyne incognita in black pepper (Piper nigrum L.). In “Biotechnology of Spices, Medicinal and Aromatic Plants” (S. Edison, K. V. Ramana, B. Sasikumar, K. Nirmal Babu, and S. J. Eapen, Eds.). Indian Society for Spices, Calicut, Kerala State, India. Ebenhoech, A., and Spadaro, O. (1992). Antiepilepsirine: a new Chinese anticonvulsant herb drug. J. Eco. Tax. Bot. 16, 99–102. El Mofty, M. M., Khudoley, V. V., and Shwaireb, M. H. (1991). Carcinogenic effect of force-feeding an extract of black pepper in Egyptian toads. Oncology 48, 347–350. Elgawhary, S. M., Lindsay, W. L., and Kemper, W. D. (1970). Effects of EDTA on the self-diffusion of zinc in aqueous solution and in soil. Soil Sci. Soc. Am. Proc. 34, 66–70. Esper, A., and Muhlbauer, W. (1996). Solar tunnel dryer for fruits. Plant Res. Develop. 44, 61–79. Ewald, S. (1991). Vesicular – Arbuscular Mycorrhiza Management in Tropical Agrosystems. Technical Cooperation, Federal Republic of Germany, 371. Fassuliotis, G., and Bhatt, D. P. (1982). Potential of tissue culture for breeding root knot nematode resistance in vegetables. J. Nematol. 14, 10–14. Ferraz, E. C. A., Orchard, J. E., and Lopez, A. S. (1984). Reactions of black pepper to Meloidogyne incognita in relation to total phenol. Revista Theobroma. 14, 217–227. Ferraz, E. C. A., Lordello, L. G. E., and de Santana, C. J. L. (1988). Nutrient absorption of black pepper vine (Piper nigrum L.) infested with Meloidogyne incognita Kofoid & White (1919), Chitwood (1949). In “Boletin Tecnico Centro de Pesquisas do Cacau, Brazil, No. 160”, p. 34. Ferraz, E. C. A., Lordello, L. G. E., and Gonzaga, E. (1989). Inﬂuence of Meloidogyne incognita (Kofoid and White 1919) Chitwood 1949 on chlorophyll content of black pepper (Piper nigrum L). Agrotropica 1, 57–62. Findlay-Wilson, I. Z. F. L. (1995). Microencapsulated ﬂavours for a great taste. Internationale Zeitschrift fu¨r Lebensmittel — Technik, Marketing, Verpackung und Analytik 46(10), 76 –78. Fitchet, M. (1988a). Establishment of black pepper in tissue culture “Information Bulletin No. 189.” Citrus and Subtropical Fruits Research Institute, South Africa. Fitchet, M. (1988b). Progress with in vitro experiments of black pepper. In “Information Bulletin No. 196”. Citrus and Subtropical Fruits Research Institute, South Africa. Freire, F. C. O., and Bridge, J. (1985b). Biochemical changes induced in roots and xylem sap of black pepper by Meloidogyne incognita. Fitopatologia Brasileira 10, 483 –497. Freire, F. C. O, and Bridge, J. (1985d). Parasitism of eggs, females of juveniles of Meloidogyne incognita by Paecilomyces lilacinus and Verticillium chlamydosporium. Fitopatologia Brasileira 10, 577 –596. Fresco, L. O., and Kroonenberg, S. B. (1992). Time and spatial scales in ecological sustainability. Land Use Policy. July, 1992, 155–167. Gamble, J. S. (1925). Flora of the Presidency of Madras. In “Bot. Sur. India”, Vol. II. Calcutta (Repr.). Geetha, S. M. (1991). In “Studies on the Biology, Pathogenicity and Biocontrol of Different Populations of Radopholus similis”, p. 196. Ph. D. Thesis, Kerala University, Trivandrum, India. Geetha, C. K., and Nair, P. C. S. (1990). Effect of plant growth regulators and zinc on spike shedding and quality of pepper. Agric. Res. J. Kerala 14, 10–12. Geetha, S. P., Manjula, C., and Sajina, A. (1995). In vitro conservation of genetic resources of spices. In “Proceedings of the Seventh Kerala Science Congress, January 27 –29”, pp. 12 –16, Science Technology and Environment Committee, Kerala Government. George, C. K. (1982). Pepper cultivation in Malaysia. Indian Cocoa, Arecanut and Spices J. 5, 75–76.

AGRONOMY AND ECONOMY OF BLACK PEPPER

377

George, P. S., Nair, K. N., and Pushpangadan, K. (1989). “The Pepper Economy of India”. Oxford and IBH Publishing Co. Pvt. Ltd., New Delhi, India. Ghawas, M., and Miswan, J. (1984). Satu Kaedah memprecepatkan Pembiakn lada hitam (Piper nigrum L.). Teknol. Pertanian. MARDI Res. J. 5, 50–54. Gopalakrishnan, M., Menon, N., Padmakumari, K. P., Jayalakshmi, A., and Narayanan, C. S. (1993). GC analysis and odor proﬁles of four new Indian genotypes of Piper nigrum L. J. Essential Oil Res. 5, 247–253. Govindan, M., and Chandy, K. C. (1985). Utilization of the diazotroph, Azospirillum for inducing rooting in pepper cuttings (Piper nigrum). Curr. Sci. 54, 1186–1188. Govindarajan, V. S. (1977). Pepper — Chemistry, Technology and Quality Evaluation. Crit. Rev. Food. Sci. Nut. 9, 115– 225. Govindarajan, V. S. (1979). Pepper — chemistry, technology and quality evalutation. CRC Crit. Rev. Food. Sci. Nut. 9, 1–115. Graham, J. H. (1982). Effect of citrus root exudates on germination of chlamydospores of the vesicular–arbuscular mycorrhizal fungus Glomus epigaeum. Mycologia 74, 831 –835. Graham, J. H. (1988). Interactions of mycorrhizal fungi with soil borne plant pathogens and other organisms, an introduction. Phytopathology 78, 365–366. Graham, J. H., and Egel, D. S. (1988). Phytophthora root rot development on mycorrhizal and phosphorus-fertilized non-mycorrhizal sweet orange seedlings. Plant Dis. 72, 614–615. Guenther, E. (1950). Essential oils of the plant family Piperaceae. In “The Essential Oils”, vol 5, pp. 135–161. Robert Krieger Pub. Co., New York. Harden, H. J., and White, J. T. H. (1934). Quoted by De Waard (1969). Hartman, C. P., Divakar, N. G., and Nagaraja Rao, U. N. (1973). A study of identiﬁcation of papaya seeds in pepper. J. Food. Sci. Technol. 10, 43– 45. Harvill, E. K., Hartzell, A., and Arthur, J. M. (1943). Toxicity of piperine solution to house ﬂies control. Boyce Thompson Inst. 13, 87–92 (Merck Index, 1976, 1266). Hashim, S., Aboobaker, V. S., Madhubala, R., Bhattacharya, R. K., and Rao, A. R. (1994). Modulatory effects of essential oils from spices on the formation of DNA adducts by aﬂatoxin B in vitro. Nutr. Cancer 21, 169 –175. Hasselstrom, T. F., Hewitt, E. J., Konigsbacher, K. S., and Ritter, J. J. (1957). Composition of oil of black pepper. J. Agric. Food Chem. 5, 53 –55. Hendriks, L., Classen, N., and Jungk, A. (1981). Phosphatverarmung des wurzelnahen Bodens und Phosphataufnahme von Mais und Raps. Z. Pﬂanzenern. Bodenk. 144, 486 –499. Higashimoto, M., Purinatrapiban, J., Katoka, K., Kinouchi, T., Vinitkekumnuen, U., Akimoto, M., and Dinishi, M. (1993). Mutagenicity and antimutagenicity of extracts of three medicinal plants. Mutat. Res. 303, 135– 142. Holdeman, Q. L. (1986). “The Burrowing Nematode, Radopholus similis sensu lato”, p. 52. Division of Plant Industry, Department of Food and Agriculture and the Agriculture Commissioners of California, California, U. S. A. Holliday, P., and Mowat, W. P. (1963). Foot rot of Piper nigrum L. (Phytophthora palmivora). In “Phytopathological paper no. 5”, p. 62. Commonwealth Mycological Institute, Kew, Surrey. Hooker, J. D. (1886). “The ﬂora of British India”, vol. 5, pp. 79 –95. L. Reeva and Co., London. Houghton, P. J., Astaniou, A., Grange, J. M., and Yates, M. (1994). Antibacterial effects of extracts and constituents of Piper nigrum. L. Pharm. Pharmacol. 46(Suppl 2), 1042. Hubert, F. P. (1957). Diseases of some export crops in Indonesia. Plant Dis. Reptr. 41, 55–63. Huhtanen, C. N. (1988). Inhibition of Clostridium botulinum by spice extracts and aliphatic alcohols. J. Food Prot. 43, 195 –196. Huitema, W. K. (1941). Quoted from De Waard (1969). Humboldt, Alexander von., Bonpland., and Kunth. (1815). Cited by Burger (1972). Ichinohe, M. (1975). Infestation of black pepper vines by the root knot nematode, Meloidogyne incognita at Tome-Acu, Para, Brazil. Jap. J. Nematol 5, 36–40.

378

K. P. P. NAIR

Ichinohe, M. (1980). Studies on the root knot nematode of black pepper plantation in Amazon. In “Annual Report of the Society of Plant Protection of North Japan, No. 31”, pp. 1–8. Ichinohe, M. (1985). Integrated control of root knot nematode, Meloidogyne incognita on black pepper plantations in the Amazonian region. Agric. Ecosystems Environ. 12, 271–283. IISR. (1997). “Annual Report of Indian Institute of Spices Research 1996–97”, pp. 50– 51. Calicut, Kerala State, India. Ikeda, R. N., Stanley, W. L., Vannier, S. H., and Spitler, F. M. (1962). The monoterpene composition of some essential oils. J. Food Sci. 27, 455– 458. Jacob, J. A., and Kuriyan, K. J. (1979a). Screening of pepper varieties for resistance against root knot nematode (Meloidogyne incognita). Agric. Res J. Kerala 17, 90. Jain, S. R., and Jain, M. R. (1972). Antifungal studies on some indigenous volatile oils and their combinations. Planta Medica 22, 136. Jain, S. R., and Kar, A. (1971). Antibacterial activity of some essential oils and their combinations. Planta Medica 22, 118. Jatala, P. (1986). Biological control of plant parasitic nematodes. Ann. Rev. Phytopath. 24, 453– 489. Jennings, W. G., and Wrolstad, R. E. (1961). Volatile constituents of black pepper. J. Food Sci. 26, 499. Jeromi, P. D., and Ramanathan, A. (1983). World pepper market and India: An anlaysis of growth and instability. Ind. J. Agric. Econ. 48, 89 –97. Johri, R. K., and Zutshi, U. (1992). An Ayurvedic formulation of “Trikatu” and its constituents. J. Ethnopharmacol. 37, 85–91. Johri, R. K., Thusu, N., Khajuria, A., and Zutshi, U. (1992). Piperine mediated changes in permeability of rat intestinal epithelial cells. Biochem.Pharmacol. 43, 1401–1407. Jose, J., and Sharma, A. K. (1984). Chromosome studies in the genus Piper L. J. Indian Bot. Soc. 63, 313–319. Joseph, B., Joseph, D., and Philip, V. J. (1996). Plant regeneration from somatic embryos in black pepper. Plant Cell, Tissue and Organ Culture 47, 87–90. Kachru, K. P., Patil, R. T., and Srivastava, P. K. (1990). Post harvest technology of pepper. Spice India 3(8), 12–16. Kachru, K. P., and Gupta, R. K. (1993). Drying of spices-status and challenges. In “Proceedings of the National Seminar of Post Harvest Technology of Spices”, pp. 15 –27, Trivandrum, India. Kapoor, V. K., Chawla, A. S., Manojkumar, and Pradeepkumar (1993). Search for antiinﬂammatory agents. India Drugs 30, 481 –493. Kaul, I. B., and Kapil, A. (1993). Evaluation of liver protective potential of piperine-an active principle of black pepper. Planta Medica 59, 413– 417. Keating, M. (1993). The Earth Summit’s Agenda for Change. In “A Plain Language Version of Agenda 21 and the Other Rio Agreements”. Centre for our Common Future, Geneva, Switzerland. Keerthisinghe, G., und Mengel, K. (1979). Phosphatpufferung verschiedener Bo¨den und ihre Vera¨nderung infolge Phosphatalterung. Mitt. Dtsch. Bo¨denk, Ges. 29, 217–230. Kelker, S. M., and Krishnamoorthy, K. V. (1996). Control of bacterial contamination in in vitro cultures of pepper (Piper sp.) In “Proc National Seminar on Biotechnology of Spices and Aromatic Crops”, 24–25 April 1996. Calicut, Kerala State, India. Kerry, B. R. (1990). An assessment of progress towards microbial control of plant parasitic nematodes. J. Nematol. 22 (Suppl.), 621– 631. Kirtikar, K. R., and Basu, B. D. (1975). Indian Medicinal Plants III, 2133–2134 Bishan Singh Mahendrapal Singh. Dehra Dun, India. Koshy, P. K., and Sunderaraju, P. (1979). Response of seven black pepper cultivars to Meloidogyne incognita. Nematol. Medit. 7, 123–125. Kovar, J. L., and Barber, S. A. (1988). Phosphorus supply characteristics of 33 soils as inﬂuenced by seven rates of phosphorus addition. Soil Sci. Soc. Am. J. 52, 160 –165.

AGRONOMY AND ECONOMY OF BLACK PEPPER

379

Krishnamoorthy, B., and John Zachariah, T. (1992). Drying black pepper on polyethylene materials. Indian Cocoa, Arecanut and Spices J. 15, 75. Kueh, T. K. (1979). “Pests, Diseases and Disorders of Black Pepper in Sarawak Semongok Agricultural Research Centre”, p. 68. Department of Agriculture, Sarawak, Malaysia. Kueh, T. K. (1986). Pests and diseases of black pepper-a review. In “Pepper in Malaysia” (C. F. J. Bong and M. S. Saad, Eds.), pp. 115 –133. Universiti Pertanian Malaysia Cawangan, Sarawak, Kuching, Malaysia. Kueh, T. K. (1990). Major diseases of black pepper and their management. The Planter. 66, 59–69. Kueh, T. K., and Khew, K. L. (1982). Survival of Phytophthora palmvivora in soil and after passing through alimentary canals of snails. Plant Dis. 66, 897–899. Kueh, T. K., and Sim, S. L. (1992a). Etiology and control of Phytophthora foot rot of black pepper in Sarawak, Malaysia. In “Proceedings of the International Workshop on Black Pepper Diseases. Bander Lampung, Indonesia” (P. Wahid, D. Sitepu, S. Deciyanto, and U. Suparman, Eds.), pp. 155–162. Institute for Spice and Medicinal Crops, Bogor, Indonesia. Kueh, T. K., and Sim, S. L. (1992b). Important pepper diseases in Sarawak, Malaysia. In “Proceedings of the International Workshop on Black Pepper Diseases Bander Lampung, Indonesia” (P. Wahid, D. Sitepu, S. Deciyanto, and U. Suparman, Eds.), pp. 101 –117. Kueh, T. K., Fatimah, O., and Lim, J. L. (1993). Management of black berry disease of black pepper. In “The pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. B. Ipor, Eds.), pp. 162– 168. Universiti Pertanian Malaysia, Sarawak, Malaysia. Kurup, P. N. V., Ramdas, V. N. K., and Joshi, P. (1979). “Handbook of Medicinal Plants”. p. 143, New Delhi. Kurien, S., and Nair, P. C. S. (1988). Effect of pruning on yield in pepper (Piper nigrum L.). Agric. Res. J. Kerala 26, 137– 139. Lawrence, B. M. (1981). “Major tropical spices-pepper (Piper nigrum L.) Essential oils 1979–80”, pp. 141–228. Allured Publications Corporation, USA. Lee, E. B., Shin, K. H., and Woo, W. S. (1984). Pharmacological Study of Piperine. Arch. Pharm. Res. 7, 127–132. Leife, A. (1992). Sterilization of spices with pulsed steam. Livsmedelsteknik 34(1/2), 24 –27. Lewis, Y. S. (1984). “Spices and Herbs for Food Industry”, p. 69–77. Food Trade Press, England. Lewis, D. G., and Quirk, J. P. (1967). Phosphate diffusion in soil and uptake by plants III 31Pmovement and uptake by plants as indicated by 32P-autoradiography. Plant Soil 26, 445 –453. Lewis, Y. S., Nambudiri, E. S., and Krishnamurthy, N. (1969). Composition of pepper oil. Perfum. Essent. Oil Rec. 60, 259 –262. Lewis, Y. S., Krishnamurthy, N., Nambudiri, E. S., Sankarikutty Amma, B., Shivshankar, S., and Mathew, A. G. (1976). The need for growing pepper cultivars to suit pepper products. In “Proceedings of the International Seminar on Pepper”, pp. 4–9, Spices Export Promotion Council, Cochin, Kerala State, India. Lin, J. K. (1986). Food-borne amines and amides as potential precursors of endogenous carcinogens. In “Proceedings of the National Science Council of Republic of China”. (B), 10, pp. 20–34. Linderman, R. G. (1988). Mycorrhizal interactions with the rhizosphere microﬂora: The micorrhizosphere effect. Phytopathology 78, 366–370. Linnaeus, C. (1753). “Species Plantarum”. London. Lissamma, J., Nazeem, P. A., Mini, S. T., Shaji, P., and Mini, B. (1996). In vitro techniques for mass multiplication of black pepper (Piper Nigrum L.) and ex vitro performance of the plantlets. J. Plantation Crops 24 (Suppl.), 511–516. Liu, G., et al., (1984). Quoted from Ebenhoech and Spadaro (1992). Lokesh, M. S., and Gangadarappa, P. M. (1995). Management of Phytophthora foot rot and nematode diseases in black pepper (Piper nigrum L.). J. Spices and Aromatic Crops 4, 61 –63. Lu, S., and Miller, M. H. (1994). Prediction of phosphorus uptake by ﬁeld-grown maize with BarberCushman model. Soil Sci. Soc. Am. J. 58, 852 –857.

380

K. P. P. NAIR

Madasamy, M., and Gothandapani, L. (1995). Spiral separator for cleaning. Spice India 8(11), 7–8. Madhusudhanan, K., and Rahiman, B. A. (1996). In vitro response of Piper species on activated charcoal supplemented media. In “Biotechnology of Spices, Medicinal and Aromatic Plants” (S. Edison, K. V Ramana, B. Sasikumar, K. Nirmal Babu, and S. J. Eapen, Eds.), pp. 16 –19. Indian Soc. for Spices, Calicut, Kerala State, India. Mahindru, S. N. (1982). “Spices in Indian Life”. Sultan Chand & Sons, New Delhi, India. Majumdar, A. M., Dhuley, J. N., Deshmukh, V. K., Raman, P. H., Thorat, S. L., and Naik, S. R. (1990b). Effect of piperine on pentobarbitone induced hypnosis in rats. Indian J. Exp. Biol. 28, 486–487. Malebennur, N. S., Ganadarappa, P. M., and Hegde, H. G. (1991). Chemical control of foot rot of black pepper caused by Phytophthora capsici (Phytophthora palmivora MF 4). Indian Cocoa, Arecanut and Spices J. 14, 148–149. Mamootty, K. P. (1978). “Quick wilt disease of pepper (Piper nigrum Linn.)-1 Symtomatological studies on the quick wilt disease of pepper”, p. 87, M. Sc(Ag.) Thesis, Kerala Agricultural University, Thrissur, Kerala State, India. Manilal, V. B., and Gopinathan, K. M. (1995). “Annual Report, Regional Research Laboratory, Council of Scientiﬁc and Industrial Research”, pp. 9–10. Trivandrum, Kerala State, India. Manjunath, A., and Bagyaraj, D. J. (1982). Vesicular arbuscular mycorrhizas in three plantation crops and cultivars of ﬁeld bean. Curr. Sci. 51, 707 –709. Manohara, D., Kasim, R., and Sitepu, D. (1992). Current research status of foot rot disease in Indonesia. In “Proceedings of the International Workshop on Black Pepper Diseases Bander, Lampung, Indonesia” (P. Wahid, D. Sitepu, S. Deciyanto, and U. Suparman, Eds.), pp. 144 –154. Institute for Spice and Medicinal Crops, Bogor, Indonesia. Mapiliandri, I. (1989). The extraction of black pepper oleoresin. In “Thesis from Diplome. IV. Ministry of Industry, Indonesia, Ministry of Agriculture (1996)”. Statistik Perkebunan, Jakarta. Marschner, H. (1994). Micronutr. News Information. 14, 3–5. Masakazu, M., Isamu, N., and Hirosuki, Y. (1980). The Piperaceae amides Part III. Insecticidal joint action of pipercide and co-occuring compounds isolated from Piper nigrum. Agri. Biol. Chem. 44, 1071–1073. Mathew, P. M. (1958). Studies on Piperaceae. J. Indian Bot. Soc. 37, 155–171. Mathew, P. M. (1972). Karyomorphological studies on Piper nigrum. In “Proceedings of the National Symposium on Plantation Crops.” J. Plantation Crops, 1, pp. 15–18, (Suppl). Mathew, A. G. (1992). Chemical constituents of pepper. International Pepper News Bull. 16(2), 18–21. Mathews, V. H., and Rao, P. S. (1984). In vitro responses in black pepper (Piper nigrum). Curr. Sci. 53, 183 –186. Mayore, M., Smith, J. L., and Garhaur, D. Y. (1987). Effect of red pepper and black pepper on the stomach. Am J. Gastroenterol. 82, 211–214. McCown, H. B., and Amos, R. (1979). Initial trials of micropropagation with birch. Proc. Inter. Plant Prop. Soc. 29, 387–393. Mengel, K. (1985). Dynamics and availability of major nutrients in soils. Adv. Soil Sci. 2, 65– 131. Menon, K. K. (1949). The survey of pollu and root diseases of pepper. Indian J. Agric. Sci. 19, 89–136. Miquel, F. A. W. (1843). “Systema Piperacearum”. Rotterdam, The Netherlands. Miyakado, M., Nakayama, I., Yoshoka, H., and Nakatani, N. N. (1979). The Piperaceae amides: structure of pipercide. Agri. Biol. Chem. 43, 1609–1611. Mohanakumaran, B., and Abi, Cheeran. (1981). Nutritional requirement of pepper vine trailed on living and dead standards. Agric. Res. J. Kerala. 19, 3– 5. Mohandas, C., and Ramana, K. V. (1987). Slow wilt disease of back pepper and its control. Indian Cocoa, Arecanut and Spices J. 11, 10 –11.

AGRONOMY AND ECONOMY OF BLACK PEPPER

381

Mohandas, C., and Ramana, K. V. (1991). Pathogenicity of Meloidogyne incognita and Radopholus similis on black pepper (Piper nigrum L.) J. Plantation Crops. 19, 41 –43. Montaya, L. A., Sylvain, P. G., and Umana, R. (1961). Effect of light intensity and nitrogen fertilization upon growth differentiation balance in Coffea arabica L. Coffee (Turrialba) 3, 97 –115. Mos, A. M. (1995). Microencapsulation enhances ﬂavourings and ingredients. Voedingsmiddelentechnologie 28(23), 64–65. Mujumdar, A. M., Dhuley, J. N., Deshmukh, V. K., Raman, P. H., and Naik, S. R. (1990a). Antiinﬂammatory activity of piperine. Japn. J. Med. Sci. Biol. 43, 95–100. Muller, C. J., and Jennings, W. G. (1967). Constituents of black pepper. Some sesquiterpene hydrocarbons. J. Agric. Food Chem. 15, 762 –766. Muller, C. J., Creveling, R. K., and Jennings, W. G. (1968). Some minor sesquiterpene hydrocarbons of black pepper. J. Agric. Food Chem. 16, 113–117. Murashige, T., and Skoog, F. (1962). A revised medium for rapid growth and bioassays with tobacco tissue culture. Physiol. Plant 15, 473–497. Murni, A. M., and Faodji, R. (1990). The combination effect of KCl with two nitrogen sources on the growth of black pepper. Bull. Penelitian Tanaman Rempah dan Obat. 2(2), 79–84. Mustika, I. (1990). Studies on the interactions of Meloidogyne incognita, Radopholus similis and Fusarium solani on black pepper (Piper nigrum L.). p. 127, Ph. D. Thesis, Wageningen Agricultural University, Wageningen, The Netherlands. Mustika, I. (1991). Response of four black pepper cultivars to infection by Radopholus similis, Meloidogyne incognita and Fusarium solani. Industrial Crops Res. J. 4(1), 17–22. Mustika, I., and Zainuddin, N. (1978). Efﬁcacy tests of some nematicides for the control of nematodes on black pepper. Pemberitaan L.P.T.I. 30, 1–10. Nadkarni, K. M. (1976). “Indian Materia Medica”, pp. 971–972. Popular Prakashan, Bombay, India. Nair, K. P. P. (1984a). Towards a better approach to soil testing based on the buffer power concept. In “Proc 6th. Int. Colloquium for the Optimization of Plant Nutrition”, vol 4, 2 –8, September (Prevel Pierre-Martin, Ed.), pp. 1221–1228. Montpellier, France. Nair, K. P. P. (1984c). Zinc buffer power as an important criterion for a dependable assessment of plant uptake. Plant Soil 81, 209–215. Nair, K. P. P. (1989). Comments on Phosphorus supply characteristics of 33 soils as inﬂuenced by seven rates of phosphorus addition. Soil Sci. Soc. Am. J. 53, 984. Nair, K. P. P. (1992). Measuring P buffer power to improve routine soil testing for phosphate. Eur. J. Agron. 1(2), 79–84. Nair, P. M. (1993). Spice irradiation-Present and future scenario. In “Post Harvest Technology of Spices, Indian Soc. Spices” (C. S Narayanan, B. Sankarikutty, A. Nirmala Menon, P. N. Ravindran, and B. Sasikumar, Eds.), pp. 28 –33. Indian Institute of Spices Research, Calicut, Kerala State, India. Nair, K. P. P. (1996). The Buffering Power of Plant Nutrients and Effects on Availability. Adv. Agron. 57, 237–287. Nair, K. P. P. (2002). Sustaining crop production in the developing world through The Nutrient Buffer Power Concept. In “Proceedings of the 17th World Soil Science Congress”, 14–21 August, vol. 2, p. 652, Bangkok, Thailand. Nair, K. P. P. (2003). Measuring Zn buffer power to predict Zn availability. J. Plant Nutr. (in Press). Nair, M. K., and Gopalasundaram, P. (1993). Nutrient management in tree based cropping system. In “Potassium for Plantation Crops” (Mahatim Singh, and M. K. Mishra, Eds.), pp. 200–211. Potash Research Institute of India, Gurgaon, India. Nair, K. P. P., and Mengel, K. (1984b). Importance of phosphate buffer power for phosphate uptake by rye. Soil Sci. Soc. Am. J. 48, 92–95.

382

K. P. P. NAIR

Nair, P. K. U., and Sasikumaran, S. (1991). Effect of some fungicides on quick wilt (foot rot) disease of black pepper. Indian Cocoa, Arecanut and Spices J. 14, 95 –96. Nair, M. R. G. K., Das, N. M., and Menon, M. R. (1966). On the occurrence of the burrowing nematode, Radopholus similis (Cobb 1893) Thorne 1949, on banana in Kerala. Indian J. Ent. 28, 553–554. Nair, K. P. P., Nand, Ram., and Sharma, P. K. (1984d). Quantitative relationship between zinc transport and plant uptake. Plant Soil 81, 217–220. Nair, P. K. U., Mammootty, K. P., Sasikumaran, S., and Pillay, V. S. (1993). Phytophthora foot rot of black pepper (Piper nigrum L.). A management study with organic amendments. Indian Cocoa, Arecanut and Spices J. 17, 1–2. Nair, K. P. P., Sadanandan, A. K., Hamza, S., and Jose, Abraham. (1997). The Importance of Potassium Buffer Power in the Growth and Yield of Cardamom. J. Plant Nutr. 20(7 & 8), 987–997. Nakatani, N., Inatani, R., Ohta, H., and Nishioka, A. (1986). Chemical constituents of pepper and application to food preservation. Naturally occurring antioxidant Environ. Health Perspect. 67, 135–147. Nambiar, K. K. N. (1978). Diseases of pepper in India. In “Proceedings of the National Seminar on Pepper” (M. K Nair and M. Haridasan, Eds.), pp. 11–14. Central Plantation Crops Research Institute, Calicut, Kerala State, India. Nambiar, K. K. N., and Sarma, Y. R. (1977). Wilt diseases of black pepper. J. Plantation Crops 5, 92–103. Nambiar, K. K. N., and Sarma, Y. R. (1980). Factors associated with slow wilt of pepper. In “Proceedings of the PLACROSYM-II 1979.” pp. 348 –358, Indian Society for Plantation Crops, Kasaragod, Kerala State, India. Nandakumar, C., Ragunath, P., Visalakshi, A., and Mohandas, N. (1987). Management of pollu beetle on pepper by adjusting the spray schedules “Abst. Natn. Symp. Integrated Pest Control-Progress and Perspectives.” p. 50, Trivandrum, Kerala State, India. Nazeem, P. A., Joseph, L., Thampi, M. S., Sujatha, R., and Nair, G. S. (1993). In vitro culture system for indirect organogenesis for black pepper (Piper nigrum L.) In “Golden Jubilee Symposium on Horticultural Research-Changing Scenario,” 24–28 May, p. 250, Bangalore, India. Ng, S. C. (1993). Uses of pepper. In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. B. Ipor, Eds.), pp. 347–353. Universiti Pertanian Malaysia, Sarawak, Malaysia. Nigam, M. D., and Handa, K. L. (1964). Constituents of essential oils. Indian Perfumer. 8, 15–17. Nirmal Babu, K., Naik, V. G., and Ravidran, P. N. (1993a). Two new taxa of Piper (Piperaceae) from Kerala, India with a note on their origin and inter-relationships. J. Spices and Aromatic Crops 2, 26–33. Nirmal Babu, K., Rema, J., Lukose, R., Ravindran, P. N., Johnson George, K., Sasikumar, B., and Peter, K. V. (1993b). “Spices Biotechnology at National Research Centre for Spices”, Technical Report II, Calicut, Kerala State, India. Nirmal Babu, K., Geetha, S. P., Manjula, C., Sajina, A., Minoo, D., Samsudeen, K., Ravindran, P. N., and Peter, K. V. (1996b). Biotechnology-its role in conservation of genetic resources of spices. In “Biotechnology for Development.” (M. R. Das and Satish Mundayoor, Eds.), pp. 198 –212. State Committee on Science, Technology and Environment, Trivandrum, Kerala State, India. Nybe, E. V., and Nair, P. C. S. (1986). Nutrient deﬁciency in black pepper (Piper nigrum L.) I. Nitrogen, Phosphorus and Potassium. Agric. Res. J. Kerala 24, 132–150. Nye, P. H. (1979). Diffusion of ions and uncharged solutes in soils and clays. Adv. Agron. 31, 225–272. Pangburn, R. M., Jennings, W. G., and Noelting, C. F. (1970). Preliminary examination of odour quality of black pepper oil. Flavour Ind. 1, 763.

AGRONOMY AND ECONOMY OF BLACK PEPPER

383

Parmar, V. S., Jain, S. C., Bisht, K. S., Jain, R., Taneja, P., Jha, A., Tyagi, O. D., Prasad, A. K., Wengel, J., Olsen, C. E., and Boll, P. M. (1977). Phytochemistry of the genus Piper. Phytochemistry 46, 597–673. Parry, J. W. (1969). “Spices Vol 1. Chemical Publishing Co. Inc.” New York, USA. Paulus, A. D., Eng, L., Teo, C. H., and Sim, S. L. (1993). Screening black pepper genotypes and Piper spp. for resistance to root knot nematode. In “The Black Pepper Industry-Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. P. Ipor, Eds.), pp. 132–139. Universiti Pertanian Malaysia, Sarawak, Malaysia. Perez, C., and Anesini (1994). Antibacterial activity of elementary plants against Staphylococcus aureus growth. American J. Chinese Medicine 22, 169–174. Peter, K. V. (1996a). Emerging potential for global spices trade. The Hindu Businessline, 21st December 1996. Peter, K. V. (1996b). Spices research and development-an updated overview. Employment News 21, 1. Philip, V. J., Joseph, D., Trigs, G. S., and Dickinson, N. M. (1992). Micropropagation of black pepper (Piper nigrum L.) through shoot tip cultures. Plant Cell Rep 12, 41– 44. Pillai, V. S. (1992). Management of improved cultivars of black pepper. In “Black Pepper and Cardamon: Problems and Prospects” (Y. R. Sharma, S. Devasahayam, and M. Ananadaraj, Eds.), pp. 17–19. Indian Soc. Spices, Calicut, Kerala State, India. Pillai, V. S., and Sasikumaran, S. (1976). A note on the chemical composition of pepper plant (Piper nigrum L.). Arecanut and Spices Bull. 8, 13– 17. Pillai, V. S., Chandy, K. C., Sasikumaran, S., and Nambiar, P. K. V. (1979). Response of Panniyur-1 variety to nitrogen and lime application. Indian Cocoa, Arecanut and Spices J. 3(2), 35 –38. Pillai, V. S., Ali, A. B., and Chandy, K. C. (1982). Effect of 3-IBA on root initiation and development in stem cuttings of pepper (Piper nigrum L.). Indian Cocoa, Arecanut and Spices J. 6(1), 7–9. Pillai, V. S., Sasikumaran, S., and Nambiar, P. K. V. (1987). N, P and K requirement of black pepper. Agric. Res. J. Kerala 25(1), 74 –80. Prabhakaran, P. V. (1994). Assessment of losses due to pests, diseases and drought in black pepper. J. Indian Soc. Agric. Stat. 47, 59–60. Pradeep, K. U., and Geervani, P. (1994). Inﬂuence of spices on protein utilisation of winged bean and horsegram. Plant Foods Hum. Nut. 46, 187–193. Premkumar, T., and Nair, M. R. G. K. (1987a). Reduction in damage by Longitarsus nigripennis to berries of black pepper by insecticides. Madras Agric. J. 74, 326 –327. Premkumar, T., and Nair, M. R. G. K. (1988). Effect of some planting conditions on infestation of black pepper by Longitarsus nigripennis Mots. Indian Cocoa, Arecanut and Spices J. 10, 83 –84. Premkumar, T., Banerjee, S. K., Devasahayam, S., and Koya, K. M. A. (1986). Effect of different insecticides on the control of pollu beetle Longitarsus nigripennis Mots., a major pest of black pepper Piper nigrum L. Entomon. 11, 219– 221. Premkumar, T., Devasahayam, S., and Koya, K. M. A. (1994). Pests on spice crops. In “Advances in Horticulture-Plantation and Spice Crops Part 2” (K. L. Chadha and P. Rethinam, Eds.), vol. 10, pp. 787–823. Malhotra Publishing House, New Delhi, India. Pruthi, J. S. (1992). Advances in sun/solar drying and dehydration of pepper (Piper nigrum L.). International Pepper News Bull. 16(2), 6–17. Pruthi, J. S. (1993). “Major Spices of India: Crop Management and Post-Harvest Technology”, pp. 44–105. Indian Council of Agricultural Research, New Dehhi, India. Pruthi, J. S. (1997). Diversiﬁcation in pepper utilization. Part I & II. Green pepper products. International Pepper News Bull. 14(4), 5–9; 14(5), 6–9. Pruthi, J. S., Bhat, A. V., Kutty, S. K., Varkey, A. G., and Gopalakrishnan, M. (1976). Preservation of fresh green pepper by canning, bottling and other methods. In “International Seminar on Pepper”, pp. 64–72. Spices Export Promotion Council of India, Cochin, Kerala State, India.

384

K. P. P. NAIR

Purseglove, J. W., Brown, E. G., Green, C. L., and Robbins, S. R. J. (1981). “Spices”, vol. 1, pp. 1 –99. Longman, London. Rahiman, B. A. (1981a). Biosystematic studies in varieties and spices of Piper occurring in Karnataka region. Ph. D. Thesis, University of Mysore, Karnataka State, India. Raj, H. G. (1978). “A comparison of the systems of cultivation of Black pepper-Piper nigrum L. in Malaysia and Indonesia”, pp. 65–75. The Directorate of Extesion Education, Kerala Agricultural University, Trichur, Kerala State, India. Raj Mohan, K. (1985). “Standardization of tissue culture techniques in important horticultural crops”. Ph. D. Thesis, Kerala Agricultural University, Thrissur, Kerala State, India. Ramachandran, N. (1990). “Bioefﬁcacy of systemic fungicides against Phytophthora infections in black pepper (Piper nigrum L.)”, p. 138, Ph.D Thesis, University of Calicut, Calicut, Kerala State, India. Ramachandran, N., and Sarma, Y. R. (1985). Efﬁcacy of three systemic fungicides in controlling Phytophthora infections in black pepper. Indian Phytopath. 38, 160–162. Ramachandran, N., Sarma, Y. R., and Nambiar, K. K. N. (1986). Spatial distribution of Phytophthora palmivora MF4 in the root zones of Piper nigrum. Indian Phytopath. 39, 414 –417. Ramachandran, N., Sarma, Y. R., and Anandaraj, M. (1988b). Sensitivity of Phytophthora species affecting different plantation crops in Kerala to Metalaxyl. Indian Phytopath. 41, 438 –442. Ramachandran, N., Sarma, Y. R., Anandaraj, M., and Abraham, J. (1988c). Effect of climatic factors on Phytophthora leaf infection in black pepper grown in arecanut–black pepper mixed cropping system. J. Plantation Crops 16, 110–118. Ramachandran, N., Sarma, Y. R., and Anandaraj, M. (1990). Vertical progression and spread of Phytophthora leaf infection in black pepper in arecanut–black pepper mixed cropping system. Indian Phytopath. 43, 414 –419. Ramachandran, N., Sarma, Y. R., and Anandaraj, M. (1991). Management of Phytophthora infections in black pepper. In “Diseases of Black Pepper” (Y. R. Sarma and T. Premkumar, Eds.), pp. 158 –174. National Research Centre for Spices, Calicut, Kerala State, India. Ramadasan, A. (1987). Canopy development and yield of adult pepper vines in relation to light interception. Indian Cocoa, Arecanut and Spices J. 11, 43–44. Ramana, K. V. (1992). “Final Report of the project: Role of Nematodes in the incidence of Slow Decline (Slow Wilt Disease) of Black Pepper and Screening Pepper Germplasm against Nematodes”, p. 149. National Research Centre for Spices, Calicut, Kerala State, India. Ramana, K. V. (1994). Efﬁcacy of Paecilomyces lilacinus (Thom). Samson in suppressing nematode infestations in black pepper (Piper nigrum L.). J. Spices and Aromatic Crops 3, 130 –134. Ramana, K. V., and Eapen, S. J. (1998). Plant parasitic nematodes associated with spices and condiments. In “Nematode Diseases in Plants” (P. C. Trivedi, Ed.), pp. 217 –251. C. B. S. Publishers and Distributors, New Delhi, India. Ramana, K. V., and Mohandas, C. (1986). Reaction of black pepper germplasm to root knot nematode Meloidogyne incognita. Indian J. Nematol. 16, 138–139. Ramana, K. V., Mohandas, C., and Ravindran, P. N. (1987b). Reaction of black pepper germplasm to the burrowing nematode (Radopholus similis). J. Plantation Crops. 15, 65–66. Ramana, K. V., Mohandas, C., and Eapen, S. J. (1994). “Plant Parasitic Nematodes and Slow Decline Disease of Black Pepper”, p. 14. Tech. Bull. National Research Centre for Spices, Calicut, Kerala State, India. Ramanathan, P. K., and Rao, P. N. S. (1974). Equipment needs of spice industry. In “Proceedings of the Symposium on Development and Prospects of Spice Industry in India”, pp. 82–84. Central Food Technology Research Institute, Mysore, Karnataka State, India. Ramesh, C. R. (1982). Root infection and population density of VA mycorrhizal fungi in a coconut based multi-storied cropping. In “Proceedings of the Plantation Crops Symposium (PLACROSYM V)” (K. V. A. Bavappa, R. D. Iyer, P. K. Das, M. R. Sethuraj, V. Ranganathan, K. K. N. Nambiar, and M. J. Chacko, Eds.), pp. 548 –554. Central Plantation Crops Research Institute, Kasaragod, Kerala State, India.

AGRONOMY AND ECONOMY OF BLACK PEPPER

385

Rao, C. S. S., and Nigam, S. S. (1976). Antimicrobial acitivity of some Indian essential oils. Indian Drugs 14, 62. Ravindran, P. N. (1991). “Studies on Black Pepper and Some of its Wild Relatives”. Ph. D. Thesis University of Calicut, Calicut, Kerala State, India. Ravindran, P. N., and Nirmal Babu, K. (1994a). Genetic Resources of Black Pepper. In “Advances in Horticulture” (K. L. Chadha and P. Rethinam, Eds.), pp. 99–120. Malhotra Publications Co., New Delhi, India. Ravindran, P. N., and Peter, K. V. (1995). Biodiversity of major spices and their conservation in India. In “Proceedings of the South Asia Coordinators Meeting on Plant Genetic Resources. IPGRI” (R. K. Arora and V. R. Rao, Eds.), pp. 123–134. New Delhi, India. Ravindran, P. N., Nair, M. K., and Nair, R. A. (1987). Two taxa of Piper (Piperaceae) from Silent Valley Forests, Kerala. J. F. Eco. Tax. Bot. 10, 167 –169. Ravindran, P. N., Nair, M. K., Nair, R. A., Nirmal Babu, K., and Chandran, K. (1990). Ecological and taxonomical notes on Piper spp. from Silent Valley Forests, Kerala. J. Bombay Nat. His. Soc. 87, 421 –426. Reddy, B. N., Sadanandan, A. K., Sivaraman, K., and Abraham, J. (1992). Effects of planting density on black pepper on yield and nutrient availability in soil. J. Plantation Crops 20, 10–13 (Suppl.). Reen, R. K., Jamwal, D. S., Taneja, S. C., Koul, J. L., Dubey, P. K., Wiebel, F. J., and Singh, J. (1993). Impairment of UDP-GDH and glucuronidation activities in liver and small intestine of rat and guinea pig in vitro by piperine. Biochem. Pharmacy 46, 229–238. Rehiman, P. V. A., and Nambiar, P. K. V. (1967). Insecticidal control of pollu beetle (Longitarsus nigripennis) on pepper. Madras Agric. J. 54, 39 –40. Revankar, G. D., and Sen, D. P. (1974). Antioxidant effects of a spice mixture on sardine oil. J. Food Sci. Technol (India) 11, 31– 32. Rheede, H. Van (1678). “Hortus Indicus Malabaricus”, vol. 7, pp. 23–31. Amstelodami, pl. 12–16. Richard, H. M., and Jennings, W. G. (1971). Volatile composition of black pepper. J. Food Sci. 36, 584 –589. Richard, H. M., Russel, G. F., and Jennings, W. G. (1971). The volatile components of black pepper varieties. J. Chromatogr. Sci. 9, 560 –566. Risfaheri Hidayat, T. (1996). Study on decorticating of pepper berries by soaking in boiling water method. International Pepper News Bull 20(1), 17–20. Rosengarten, F. Jr. (1973). The Book of Spices. Pyramid Books, New York, U.S. A. Ruiz, Pavon. (1794). Cited by Trelease and Yuncker (1950). Russel, G. F., and Else, J. (1973). Volatile compositional differences between cultivars of black pepper (Piper nigrum). J. Assoc. off. Anal. Chem. 56, 344– 351. Sadanandan, A. K. (1990). “Response of pepper to applied potassium (Abs.)”, p. 149. National Seminar on Potassium on Plantation Crops, Bangalore, India. Sadanandan, A. K. (1994). Nutrition of black pepper. In “Advances in Horticulture. Vol. 9-Plantation and Spice Crops (Part I).” (K. L. Chadha and P. Rethinam, Eds.), pp. 423 –456. Sadanandan, A. K., and Hamza, S. (1996). Studies on nutritional requirement of bush pepper for yield and quality (Abs.). In “Proceedings of the PLACROSYM-XII, RRII”, p. 48, Kottayam, Kerala State, India. Sadanandan, A. K., Abraham, J., and Anandaraj, M. (1992). Impact of high production technology on productivity and disease incidence of black pepper. In “Proceedings of the PLACROSYM-IX” (K. K. N. Nambiar, R. D. Iyer, and E. V. V. B. Rao, Eds.), 20, pp. 384–387. Indian Soc. Plantation Crops, Kasaragod, Kerala State, India. Saito, Y., and Asari, T. (1976). Studies on the antioxidant properties of spices. I. Total tocopherol content in spices. J. Japanese. Soc. Food. Nutr. 29, 289–292. Sajina, A., Minoo, D., Geetha, S. P., Samsudeen, K., Rema, J., Nirmal Babu, K., Ravindran, P. N., and Peter, K. V. (1996). Production of synthetic seeds in spices. In “Biotechnology of Spices,

386

K. P. P. NAIR

Medicinal and Aromatic Plants” (S. Edison, K. V. Ramana, B. Sasikumar, K. Nirmal Babu, and S. J. Eapen, Eds.), pp. 65–69. Indian Soc. for Spices, Calicut, Kerala State, India. Salzer, U. J. (1975). Uber die Fettsaurezussamensetzung de lipoide einiger gewurze. Fette Stefen Anstrichm. 77, 446 –450. Salzer, U. J., Broker, U., Klie, H. P., and Liepe, H. U. (1977). Effect of pepper and pepper constituents on the microﬂora of sausage products. Fleischwirtschafe 57, 2011–2121. Samraj, J., and Jose, P. C. (1966). A Phytophthora wilt of Pepper (Piper nigrum). Sci Cult. 32, 90 –92. Samuel, R., Prabhu, V. K. K., and Narayanan, C. S. (1984). Inﬂuence of spice essential oils on the life history of Lasioderma seriorne. Entomon. 9, 209. Sanchez, P. A. (July 10–16, 1994). Tropical soil fertility research: Towards the second paradigm. In “Proceedings of the XV ISSS Symposium”, vol. 1, pp. 89–104. Acapulco, Mexico. Sankarikutty, B., Padmakumari, K. P., Howa, Umma., and Narayanan, C. S. (1994). pp. 7–8. Annual Report, Regional Research Laboratory, Council of Scientiﬁc and Industrial Research, Trivandrum, Kerala State, India. Santosh, P. (1985). “Cost of cultivation and marketing of pepper in Cannanore District”. M. Sc. (Ag) Thesis, Kerala Agricultural University, Thrissur, Kerala State, India. Sarkar, B. B., Choudhary, D. M., Nath, S. C., Bhattacharjee, N. R., and Barna, D. T. C. (1985). Wilt disease of pepper and its control in Tripura. Indian Cocoa, Arecanut and Spices J. 8, 63–66. Sarma, Y. R., Anandaraj, M., and Ramana, K. V. (1992c). Present status of black pepper diseases in India and their management. In “Proceedings of the International Workshop on Black Pepper Diseases Bander Lampung, Indonessia” (P. Wahid, D. Sitepu, S. Deciyanto, and U. Suparman, Eds.), pp. 67–68. Institute for Spice and Medicinal Crops, Bogor, Indonesia. Sarma, Y. R., Anandaraj, M., and Venugopal, M. N. (1996). Biological control of diseases in spices. In “Biological Control in Spices” (M. Anandaraj and K. V. Peter, Eds.), pp. 1– 19. Indian Institute of Spices Research, Calicut, Kerala State, India. Sasikumar, B., and Veluthambi, K. (1994). Kanamycin sensitivity of cultured tissues of Piper nigrum L. J. Spices and Aromatic Crops 3, 158–160. Satheesan, K. N., Rajagopalan, A., Sukumarapillai, V., Jagadeesh Kumar, T. N., Mammooty, K. P., and Unnikrishnan Nair, P. K. (1997a). Inﬂuence of irrigation levels on yield, quality and incidence of pests and diseases in black pepper (Piper nigrum L.) (Abs.). In “National Seminar on Water Management for Sustainable Production and Quality of Spices”, p. 21, Madikheri, Karnataka State, India. Schenk, R. U., and Hildebrandt, A. C. (1972). Medium and techniques for induction and growth of monocotyledonous and dicotyledonous plant cell cultures. Can. J. Bot. 50, 199– 204. Schneider, B. (1993). Steam sterilization of spices. Fleischwirtschaft 73(6), 646–649. Schu¨ller, H. (1969). Die CAL-Methode, eine neue Methode zur Bestimmung des pﬂanzen verfu¨gbaren Phosphates in Bo¨den. Z. Pﬂanzenerna¨hr Bodenkd. 123, 48 –63. Scott, R. (1992). Master spice: Extruded low count spices. In “Proceedings of the New Technologies Symposium Bristol”, pp. 6– 8, U. K. Scott, W. P., and McKibben, G. H. (1978). Toxicity of black pepper extract to ball weevil. J.Eco. Ent. 71, 343 –344. Selim, H. M. (1992). Modeling the transport and retention of inorganics in soils. Adv. Agron. 47, 331–384. Sen, A. R., and Roy, B. R. (1974). Adulteration in spices. In “Proceedings of the National Symposium on Development and Prospects of Spice Industry in India”. Central Food Technology Research Institute, Mysore, Karnataka State, India. Sethuram, P. (1995). Cochin Pepper market-a journey through history. Ind. Spices. 32, 3. Sewell, G. W. F. (1965). The effect of altered physical conditions of soil on biological control. In “Ecology and Soil Borne Plant Pathogens” (K. F. Baker and W. C. Snyder, Eds.), pp. 479– 494. University of California Press, Berkeley, U. S. A.

AGRONOMY AND ECONOMY OF BLACK PEPPER

387

Sharma, R. D., and Loof, P. A. A. (1974). Nematodes of cocoa region of Bahia, Brazil IV. Nematodes in the rhizosphere of pepper (Piper nigrum L.) and clove (Eugenia caryophylla Thumb). Revista Theobroma. 4, 26 –32. Sharma, N. L., Nigam, N. C., and Handa, K. L. (1962). Pepper husk in oil. Perfuem. Kosmet. 43, 505 –506. Sheela, M. S., Venkitesan, T. S., and Mohandas, N. (1993). Status of Bacillus spp. as biocontrol agents of root knot nematode (Meloidogyne incognita) infesting black pepper (Piper nigrum L.). J. Plantation Crops 21, 218– 222 (Suppl.). Sher, S. A., Chunram, C., and Pholcharoen, S. (1969). Pepper yellows disease and nematodes in Thailand. FAO Plant Prot. Bull. 17, 33. Shin, K. H., Yun, H. S., Won, W. S., and Lee, C. K. (1980). Pharmacological activity of Piper retrofractrum. Soul. Tacha. Sean. Young Opi. 18, 87 –89. (Chem. Abstr. 93, 215484s). Shuman, L. M. (1975). The effect of soil properties on zinc adsorption by soils. Soil Sci. Soc. Am. Proc. 39, 454– 458. Shwaireb, M. H., Waba, H., El Mofty, M. M., and Dutter, A. (1990). Carcinogenesis induced by black pepper and modulated by vitamin-A. Exp. Pathol. 40, 233–238. Sim, E. S. (1971). Dry matter production and major nutrient contents of black pepper (Piper nigrum L.). MalayAgric. J. 48, 73. Sim, S. L., Jafar, R., Grierson, D., Power, J. B., and Davery, M. R. (1995). Applications of molecular biology and genetic manipulation to pepper (Piper nigrum L.) breeding: present state and prospects In “Abstracts FAO/IAEA International Symposium on the Use of Induced Mutations and Molecular Techniques for Crop Improvement”. Vienna, Austria. Singh, A., and Rao, A. R. (1993). Evaluation of the modulatory inﬂuence of black pepper on the hepatic detoxiﬁcation system. Cancer Lett. 72, 5– 9. Sitepu, D., and Kasim, R. (1991). Black pepper diseases in Indonesia and their control strategy. In “Diseases of Black Pepper” (Y. R. Sarma and T. Premkumar, Eds.), pp. 13–28. National Research Centre for Spices, Calicut, Kerala State, India. Siti Hajijah, A. S. (1993). Observations of root knot infestation on pepper (Piper nigrum L.) in Sarawak. In “The Pepper Industry-Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. B. Ipor, Eds.), pp. 140–147. Universiti Pertanian Malaysia, Sarawak, Malaysia. Siti Hajijah, A. S., and Bong, C. F. J. (1993). Pepper fruit stalk as substrate for mushroom cultivation. In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. B. Ipor, Eds.), pp. 342– 346. Universiti, Pertanian Malaysia, Sarawak, Malaysia. Sivadasan, C. R. (1996). Import regulations in USA. Spice India 9, 2–10. Sivakumar, C., and Wahid, P. A. (1994). Effect of application of organic materials on growth and foliar nutrient contents of black pepper (Piper nigrum L.). J. Spices and Aromatic Crops 3, 135 –141. Soil Science Society of America, (1965). Glossary of soil science terms. Soil Sci. Soc. Am. Proc. 29, 330 –351 (Committee Report). Sosamma, V. K., and Koshy, P. K. (1995). Effect of Pasteuria penetrans and Paecilomyces lilacinus on population build up of root knot nematode, Meloidogyne incognita on black pepper. Indian J. Nematol. 25, 16–17 (Abst.). Sparks, D. L. (1987). Potassium dynamics in soils. Adv. Soil Sci. 6, 2–63. Sparks, D. L. (1989). “Kinetics of Soil Chemical Processes”. Academic Press, San Diego, CA. Sreeja, T. P., Eapen, S. J., and Ramana, K. V. (1996). Occurrence of Verticillium chlamydosporium Goddard in a black pepper (Piper nigrum L.) garden in Kerala, India. J. Spices and Aromatic Crops 5, 143–147. Sreekumar, M. M., Sankarikutty, B., Nirmala Menon, A., Padmakumari, K. P., and Narayanan, C. S. (1993). pp. 5–6. Annual Report. Regional Research Laboratory, Council of Scientiﬁc and Industrial Research, Trivandrum, Kerala State, India.

388

K. P. P. NAIR

Sridharan, K., Kalla, A. K., and Singh, J. (1978). Chemical and pharmacological screening of Piper nigrum L. leaves. J. Res. Indian Med. Yoga and Homeop. 13, 107. Stirling, G. R. (1991). In “Biological Control of Plant Parasitic Nematodes”, p. 282. C. A. B. International, Wallingford, U. K. Su, H. C. F. (1977). Insecticidal properties of black pepper on cotton weevil and cow-pea weevil. J. Eco. Ent. 70, 18 –21. Subramanyam, V., Sreenivasamoorthy, N., Krishnamoorthy, K., and Swaminathan, M. (1957). Studies on the antibacterial effects of spices. J. Scien. Ind. Res. 16, 240. Sundararaju, P., Koshy, P. K., and Sosamma, V. K. (1979). Plant parasitic nematodes associated with spices. J. Plantation Crops 7, 15–26. Sundararaju, P., Koshy, P. K., and Sosamma, V. K. (1980). Survey of nematodes associated with spices in Kerala and Karnataka. In “Proceedings of the PLACROSYM-II, 1979”, pp. 39 –44. Indian Society for Plantation Crops, Kasaragod, Kerala State, India. Suparman, U., and Zaubin, R. (1988). Effect of defoliation, IBA and Saccharose on root growth of black pepper cuttings. Industrial Crop Res. J. 1, 54 –58. Tantaoui, E. A., and Beraoud, L. (1994). Inhibition of growth and aﬂatoxin production in Aspergillus parasiticus by essential oils. J. Environ. Toxicol. Pathol. Oncol. 13, 67 –72. Thomas, K. M., and Menon, K. K. (1939). The present position of pollu disease of pepper in Malabar. Madras Agric. J. 17, 347 –356. Tien, Ju-K’ang (1981). China and the pepper trade. Hemisphere, pp. 220 –222. Trelease, W., and Yuncker, T. G. (1950). “The Piperaceae of Northern South America”. University of Illinois, U. S. A. Tripathi, S. N., Tiwari, C. N., Upadhyaya, B. N., and Singh, K. R. S. (1979). Screening of hypoglycaemic action in certain indigenous drugs. J. Res. Ind. Med. Yoga and Homeop. 14, 159. Uijil, C., and den., H. (1992). New continuous steaming method for herbs and spices. Voedingsmiddelentechnologie 25(6), 40–43. Uma Pradeep, K., Geervani, P., and Eggum, B. O. (1993). Common Indian Spices: nutrient composition, consumption and contribution to dietary values. Plant Foods Hum. Nutr. 44, 137–148. Ummer, C. (1989). Indian Spices-from the leaves of history. In “Spice Fair Commemorative Volume”, pp. 27–40. Van der Vecht, J. (1950). Plant parasitic nematodes. In “Diseases of Cultivated Plants in Indonesian Colonies” (L. G. E. Karshoven and J. van der Vecht, Eds.), pp. 16–45. I. S’ gravenhage, W. van Woeve. Variyar, P. S., Pendharkar, M. B., Banerjee, A., and Bandopadhyaya, C. (1988). Blackening in green pepper berries. Phytochemistry 27, 715–717. Varughese, J., and Anuar, M. A. (1992). Etiology and control of slow wilt disease in Johore, Malaysia. In “Proceedings of the International Workshop on Black Pepper Diseases” (P. Wahid, D. Sitepu, S. Deciyanto, and U. Suparman, Eds.), pp. 188–197. Research Institute for Spice and Medicinal Crops, Bogor, Indonesia. Velayudhan, K. C., and Amalraj, V. A. (1992). Piper pseudonigrum — a new species from Western Ghats. J. Eco. Tax. Bot. 16, 247 –250. Venkitesan, T. S., and Charles, J. S. (1980). A note on the chemical control of nematodes infesting pepper vines in Kerala. In “Proceedings of the PLACROSYM-II 1979”, pp. 27–30. Indian Society for Plantation Crops, Kasaragod, Kerala State, India. Venkitesan, T. S., and Setty, K. G. H. (1978). Reaction of 27 black pepper cultivars and wild forms to the burrowing nematode Radopholus similis(Cobb) Thorne. J. Plantation Crops 6, 81 –84. Venkitesan, T. S., and Setty, K. G. H. (1979). Control of the burrowing nematode, Radopholus similis on black pepper. Pesticides 13, 40–42. Vezyuez Olivincia, W., Shah, P., and Pitchumoni, C. S. (1992). The effect of red and black pepper on orocecal transit time. J. M. College Nutr. 11, 228–231.

AGRONOMY AND ECONOMY OF BLACK PEPPER

389

Vijayakumar, K. R., Unni, P. N., and Vamadevan, V. K. (1984). Prevention of photo-induced chlorophyll loss by the use of lime reﬂectant on the leaves of black pepper (Piper nigrum L.). Agric. For. Meteorol. 34, 17–20. Vinod, G. (1984). “Cost of cultivation and marketing of pepper in Idukki District”. M. Sc.(Ag) Thesis. Kerala Agricultural University, Thrissur, Kerala State, India. Wahid, P. (1976). Studies on yellows disease in black pepper on the island of Bangka. Pembr. LPTI 21, 64–79. Wealth of India, (1969). Piper nigrum Linn (Piperaceae). In “Wealth of India-Raw Materials”, vol. 8, pp. 99 –115. Publications and Information Directorate, Council of Scientiﬁc and Industrial Research, New Delhi. Wilkinson, H. F., Loneragan, J. F., and Quirk, J. P. (1968). The movement of zinc to plant roots. Soil Sci. Soc. Am. Proc. 32, 831–833. Winoto, S. R. (1972). Effect of Meloidogyne species on the growth of Piper nigrum L. Malaysian Agric. Res. 1, 86–90. Seoul Teahakkyo Saengyak Yonguso Opidukdip. Won, W. S., Lee, E. B., and Shin, K. H. (1979). CNS depressant activity of piperine. 18, 66–70. (Chem. Ab. 93, 215482q). Wrolstad, R. E., and Jennings, W. G. (1965). Volatile constituents of black pepper III. The monoterpene hydrocarbon fraction. J. Food Sci. 30, 274–279. Wuba, H., El-Mofty, M. M., Schweureb., M. H., and Dutter, A. (1992). Carcinogenicity testing of some constituents of black pepper. Exp. Pathol. 44, 61– 65. Yuncker, T. G. (1958). The Piperaceae: A family proﬁle. Brittonia 10, 1 –17. Zadocks, J. C., and Van den Bosch, F. (1994). On the spread of plant disease: A theory of foci. Ann. Rev. Phytopath. 32, 503 –521. Zaubin, R., and Robbert (1979). Pengaruh keasaman tanah terhadap pertumbuhan tanaman lada. Pebr. Littri. 33, 1987.

This page is intentionally left blank

SOIL MINERAL – ORGANIC MATTER – MICROORGANISM INTERACTIONS: FUNDAMENTALS AND IMPACTS P. M. Huang Department of Soil Science, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK, Canada S7N 5A8

I. Introduction II. Mechanisms of Binding of Nonhumic Organics and Humic Substances by Soil Mineral Colloids A. Nature of Mineral Colloid Surfaces B. Binding of Nonhumic Organics C. Binding of Humic Substances III. Inﬂuence of Organic Substances on the Formation, Transformation, and Surface Properties of Metal Oxides A. Aluminum Oxides B. Iron Oxides IV. Role of Soil Minerals in Abiotic Catalysis of the Formation of Humic Substances A. Polyphenol Pathway B. Maillard Reaction V. Interactions of Soil Mineral Colloids with Microorganisms A. Surface Interactions B. Inﬂuence on Microbial Activity and Survival VI. Interactions of Soil Colloids with Enzymes A. Mineral Colloid – and Humic– enzyme Complexes B. Effects on Enzymatic Activity VII. Microbial Mediation of Soil Mineral Weathering and Transformation A. Mineral Weathering B. Fine-grained Mineral Development VIII. Interactions of Soil Mineral Colloids with Organic Substances and Microorganisms in Relation to Soil Structure Stability A. Organo-mineral Complexes in Relation to Soil Structure B. Dynamics of Aggregate Turnover IX. Inﬂuence of Mineral Colloids on Biogeochemical Cycling of C, N, P, and S in Soil A. Role of Mineral Colloids in Soil Organic Matter Storage and Turnover B. Decomposition and Stabilization of C, N, P, and S in Relation to Primary and Secondary Particles 391 Advances in Agronomy, Volume 82 Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved. 0065-2113/03 $35.00

392

P. M. HUANG X. Effects of Interactions Between Microorganisms and Soil Colloids on the Transformation of Organic Pollutants A. Catalytic Transformations of Organic Pollutants B. Binding of Organic Pollutants, Enzymes, and Microorganisms on Mineral and Humic Surfaces and the Effects on Pollutant Bioavailability XI. Impact of Interactions of Physicochemical, Biochemical, and Biological Processes on Metal Transformation A. Redox Reactions of Metals B. Complexation Reactions of Metals C. Adsorption – desorption Reactions of Metals D. Precipitation – dissolution Reactions of Metals E. Metal Sorption and Uptake by Microorganisms and Biomineralization F. Metal Transformation by Microbial Excretions and Mycorrhizal Infection XII. Foreseeable Impacts of Soil Mineral –organic Matter – microorganism Interactions on the Ecosystem A. Global Ion Cycling B. Global Climate Change C. Biodiversity D. Biological Productivity and Human Nutrition E. Geomedicine F. Ecotoxicology and Human Health G. Biotechnology H. Risk Assessment I. Risk Management, Remediation, and Restoration XIII. Summary and Conclusions Acknowledgment References

Minerals, organic matter, and microorganisms are integral parts of the pedosphere and related environments. These three components are not separate entities but rather a united system constantly in association and interactions with each other in the terrestrial environment. Interactions of these components mediated by soil solution and atmosphere govern mechanisms of mineral weathering reactions, formation of short-range ordered (SRO) metal oxides, abiotic catalysis of the formation of humic substances, formation of organo-mineral complexes, microbial ecology, enzymatic activity, soil structure stability, dynamics of aggregate turnover, biogeochemical cycling of C, N, P, and S, and transformations and dynamics of metals and organic pollutants in the terrestrial environment. Foreseeable impacts of these interactions include global ion cycling and climate change, biodiversity, biological productivity and human nutrition, geomedicine, ecotoxicology and human health, biotechnology in relation to food production and security, risk assessment, and ecosystem restoration. Therefore, soil mineral – organic matter –microorganism interactions play a key role in inﬂuencing agricultural sustainability and ecosystem health. Fundamental understanding of these interactions at the molecular level is

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

393

essential for developing innovative management strategies for land and water q 2004 Academic Press. resources.

I. INTRODUCTION Soil is an integral compartment of the environment and the central organizer of the terrestrial ecosystem. Minerals, organic components, and microorganisms are three major solid components of the soil. They profoundly affect the physical, chemical, and biological properties and processes of terrestrial systems. To date, scientiﬁc accomplishments in individual disciplines of the chemistry of soil minerals, soil organic matter, and soil microbiology are commendable. However, minerals, organic matter, and microorganisms should not be considered as separate entities but rather a united system constantly in close association and interactions with each other in soil environments. Interactions of these components have enormous impact on terrestrial processes critical to environmental quality and ecosystem health. In view of the signiﬁcance of soil mineral – organic matter –microorganism interactions, the International Society of Soil Science (ISSS) established the Working Group MO “Interactions of Soil Minerals with Organic Components and Microorganisms” in 1990. The objective of this Working Group is to promote research and education on the interactions of these major solid components of soil and the impact on the production of foodstuffs and ﬁbers, the sustainability of the environment, and ecosystem health including human health on the global scale. Since its establishment, the Working Group has contributed substantially to advance the knowledge on physical/chemical/biological interfacial interactions in soil environments. The study of the interactions of these soil components has to be considered from the molecular level to ﬁeld/landscape systems and is, indeed, essential in stimulating further research to uncover the dynamics and mechanisms of soil processes. Major research thrusts of these interactive soil processes include: (1) mineral and biological catalysis and enzyme – mineral interactions leading to humus and organo-mineral complex formation, (2) surface reactions of microand macrobiota and biomolecules with soil particles, (3) the effect of soil interactive processes on the structure, dynamics, and activities of microbial communities, and (4) the impact of soil interactive processes on porosity formation by structure or organization development and on biogeochemical transformation and transport of chemical and biological components at different spatial scales, and on ecosystem integrity.

394

P. M. HUANG

This paper addresses a variety of issues on fundamentals of soil mineral – organic component – microorganism interactions at the molecular and microscopic levels and the impact on the terrestrial ecosystem.

II.

MECHANISMS OF BINDING OF NONHUMIC ORGANICS AND HUMIC SUBSTANCES BY SOIL MINERAL COLLOIDS A. NATURE

OF

MINERAL COLLOID SURFACES

The surface chemistry of soils is dependent on weathering processes of the parent materials that are initiated by geological, hydrological, and biological agents. Therefore, soil formation processes have tremendous impact on the formation and modiﬁcation of surface functional groups of soils. The siloxane surface of 2:1 expansible layer silicates is negatively charged. The ditrigonal cavities of the surface can thus adsorb positively charged ions and polymers to form hydroxy-interlayered minerals in soil weathering processes, as illustrated in Fig. 1. These hydroxy interlayered minerals occur in soils throughout the world and are present in soils of several orders of soil taxonomy systems (Rich, 1968; Barnhisel and Bertsch, 1989). In addition to the formation of hydroxy-interlayered minerals, positively charged hydroxy polymers can be adsorbed on the edges and external surfaces of phyllosilicates (Huang and Kozak 1970) as illustrated in Fig. 1. The surfaces of soil mineral colloids can also complex with organics (Mortland, 1986; Schnitzer, 1986). It is generally accepted that organics can form inner-sphere and/or outer-sphere complexes with surface functional groups of minerals. Stereochemistry has a much more important role in surface complexation with organics than in the formation of surface complexes with inorganic polymers (Sposito, 1984). The surface complexation reactions may result in the development of thick envelopes of colloidal organics on metal oxides and aluminosilicates. The complexity of the surface properties of soil mineral colloids is further manifested by the formation of iron oxyhydroxide or calcium carbonate coatings on the external surfaces of layer silicates, the growth of two-dimensional solid solutions of trace metal oxides on the periphery of Mn oxide nodules (Sposito, 1984), the coating of calcium carbonate on Mn oxide (Oscarson et al., 1983), and the coating of smectite by hydroxy-Al-organic complexes (Goh and Huang, 1986). Moreover, similar effects can occur on the surfaces of other minerals in soils (Johnston and Tombacz, 2002). Therefore, the surface chemistry of soil minerals is mainly that of the functional groups exposed on the coatings rather than that of the groups in the underlying matrix. These surface functional

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

395

Figure 1 Structural models depicting the hydroxy-Al coatings on the interlamellar and external planar surfaces and edges of mica-vermiculite. From Huang (1980).

groups have a very signiﬁcant inﬂuence on the transformation of organics in soil environments.

B. BINDING

OF

NONHUMIC ORGANICS

Nonhumic organic molecules interact with the surfaces of mineral colloids by a variety of mechanisms as discussed below. The reaction mechanisms mainly depend on the intrinsic nature and properties of the organic species, the kind of exchangeable cation on the surface of the mineral, the water content of the system, and the properties of the mineral colloids (Mortland, 1986; Dec et al., 2002).

1.

Ion Exchange

Organic cations such as protonated amines and quarternary ammonium cations will enter into exchange reactions with metal cations normally occupying exchange sites on the surfaces of mineral colloids. Organic cations are often preferred over metal cations because of size considerations. However, small organic cations

396

P. M. HUANG

can be replaced by metal cations. Organic cations will also exchange with other species of the organic cations on the exchange sites of mineral colloids.

2. Protonation Some organic molecules may become cationic after adsorption at the surfaces of mineral colloids by protonation. The relative basicity of the organic species and the Brønsted acidity of the mineral surface determine whether or not protonation happens. Organic molecules which possess basic amino groups have the possibility of accepting protons from the mineral surface to render the species cationic. The degree of protonation of the organic molecule is determined by its intrinsic basicity and the ability of the surfaces of the mineral colloids to donate protons. Although protons can occupy some exchange sites, many of the protons available may arise from hydrolysis of water associated with exchangeable metal cations. Therefore, the nature of exchangeable cation, the water content, and the nature of mineral colloids play an important role in inﬂuencing the degree of protonation of organic molecules.

3.

Coordination and Ion-dipole

Coordination or ion-dipole interaction is another kind of mineral colloid– organic interaction. Where exchangeable cations are of the transition metal type, unﬁlled d orbitals will allow the formation of coordinate covalent bonds wherein an electron donor such as N or O on functional groups of organic molecules are available as ligands. Where exchangeable cations are not able to form classical coordination complexes, they will interact with polar molecules through iondipole processes.

4.

Hydrogen Bonding

Hydrogen bonding is another type of interaction possible between organic molecules and mineral surfaces. Where a polar organic molecule cannot displace a water molecule solvating an exchangeable metal cation, it settles for a secondary role by H bonding with the directly coordinated water. This type of arrangement is referred to as water bridge where water acts as the linking entity between an exchangeable metal cation and a polar organic molecule. Another H-bonding interaction may take place between an organic cation on an exchange site and another organic molecule. For example, a protonated amine on an exchange site is H bonding with a carbonyl group.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

5.

397

Hydrophobic Bonding on Clay –Organic Complexes

Natural clay – organic complexes may have adsorption properties quite different from pure clays. Many organic molecules including aromatics and particularly the hologenated types such as chlorinated and brominated phenyls and biphenyls are adsorbed a little if any on clay surfaces. In natural environment, they are more likely to be adsorbed on organic components of clay –organic complexes. The surfaces of clays can be made hydrophobic after reaction with organic molecules that have some hydrophobic properties. When this happens, the surfaces of clay –organic complexes become hydrophobic and in turn organophilic. Therefore, hydrophobic portions of the adsorbate can react with the hydrophobic organic portions of the clay – organic complexes. Such adsorption interactions may take place since natural organic residues may have regions of hydrophobicity wherein hydrophobic molecules might be bound.

6. Anion Adsorption Organic anions may be adsorbed on the surfaces of mineral colloids. Although pure clay minerals are negatively charged, hydroxy-interlayered clay minerals (Fig. 1) have positively charged sites. Organic anions can interact with polyvalent cations such as Al and Fe on the surface coatings of clay minerals. Such an interaction has also been suggested for humic (HAs) and fulvic (FAs) acids in soils. At low water content, the dissolved organic species approach the surfaces of mineral colloids more closely than in a suspension environment, and, therefore, have a better opportunity for interaction with polyvalent metal cations on the surface coatings.

C. BINDING

OF

HUMIC SUBSTANCES

The mechanisms of binding of nonhumic substances by mineral colloids discussed above are fundamental in understanding the formation of mineral colloid –humic complexes. Comprehensive reviews of clay – humic interactions have been published (Theng, 1979; Schnitzer, 1995, 2000). The binding sites of mineral colloids for humic substances include external surfaces and interlayers of mineral colloids as discussed below.

1.

Adsorption on External Mineral Surfaces

The extent of adsorption of humic substances on external mineral surfaces depends on the physical and chemical characteristics of the surface, the pH of the system, and its water contents. The formation of a wide range of mineral –humic

398

P. M. HUANG

associations involve chemical bonding with widely differing strengths. One can visualize that complexation reactions of HA and FA with structural cations of edges and hydroxy Al (or Fe) coatings on mineral colloids are important binding mechanisms. Hydrogen bonding is also of considerable importance as clearly indicated in IR spectra of HA- and FA-mineral complexes (Schnitzer and Khan, 1972). These reactions apparently involve H and O of COOH and OH groups in HA and FA and O and H of external planar surfaces and edges. Cations with high hydration energies on mineral surfaces react via water bridges with HA- and FA- functional groups. Fulvic acid can also be adsorbed by sepiolite, apparently through displacement of bound and/or zeolitic water in the structural channels by undissociated FA (Kodama and Schnitzer, 1974).

2.

Adsorption in Clay Interlayers

Interlayer adsorption of FA by expansible layer silicates is pH dependent, being greatest at low pH, and no longer occurring at pH 5.0 (Fig. 2). An inﬂection occurs in the adsorption – pH curve near the pH corresponding to the pK of FA. In this type of reaction, the FA apparently penetrate the coordination shell of the dominant cation in the clay and displace water coordinated to the cation in the clay interlayer. The ease with which water can be displaced depends on the afﬁnity for water of the dominant interlayer cation and also on the degree of dissociation of the FA. Since the latter is very low at low pH, undissociated FA

Figure 2 Effect of pH on basal spacing (d001) and adsorption of FA. From Schnitzer and Kodama (1966).

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

399

can penetrate the clay interlayer. Therefore, interlayer adsorption of FA is greatest at low pHs. Wang and Huang (1986) reported that one of the well-deﬁned precursors for the formation of humic substances, hydroquinone, can be transformed in aqueous solution even at near-neutral pH (6.5) to humic macromolecules and adsorbed in the interlayers of nontronite (Fe(III)-bearing smectite) saturated with Ca (Fig. 3), which is the most common and most abundant exchangeable cation in soils and sediments. The decrease in the organic carbon contents of the complex on heating to between 150 and 5008C is accompanied by a decrease in basal d values. The heat treatment at 5008C destroys the humic macromolecule interlayers, as indicated by the disappearance of the slight skewness to low angles of the X-ray diffraction pattern. Most of the interlayer humic macromolecules are highly resistant to alkali extraction and are, thus, humin-type materials. Therefore, the formation of humic substance interlayers in 2:1 expansible layer silicates,

Figure 3 X-ray diffractograms of Ca-nontronite, inﬂuenced by deposition of humic macromolecules in the interlayers through polymerization of hydroquinone and the associated reactions. Diffractograms were recorded on a Philips X-ray diffractometer using Fe-Ka radiation with an Mn ﬁlter. (a) Ca-nontronite; (b, c) humic macromolecule-Ca-nontronite complex before (b) and after (c) NaOH extraction. From Wang and Huang (1986).

400

P. M. HUANG

through polymerization of phenol monomers and associated reactions in soils and sediments, merits close attention.

III. INFLUENCE OF ORGANIC SUBSTANCES ON THE FORMATION, TRANSFORMATION, AND SURFACE PROPERTIES OF METAL OXIDES Metal oxides are ubiquitous in soil. They play a very signiﬁcant role in inﬂuencing soil behavior and, thus, have great impact on the ecosystem. They may exist as crystalline minerals, as SRO minerals, or noncrystalline precipitates, which are partly present as coatings on clay minerals and humic substances. The SRO Al and Fe oxides are undoubtedly among the most reactive inorganic components of acidic and neutral soils (Ka¨mpf et al., 2000; Huang et al., 2002; Bigham et al., 2002). Organic compounds exert an important inﬂuence on the hydrolytic reactions of Al and Fe and on the formation, transformation, and surface properties of these metal oxides.

A. ALUMINUM OXIDES Numerous conceptual models have been proposed to explain reaction pathways of Al hydroxide formation from hydrolyzed Al species. The most recently proposed model (modiﬁed from Bertsch and Parker, 1996) includes concepts from the Al(OH)3 fragments of Hsu (1988) and more recent data on Al13 polynuclear species ðAlO4 Al12 ðOHÞ24 ðH2 OÞ7þ 12 Þ (Fig. 4). This model incorporates multiple reaction paths, many of which might occur simultaneously but at different reaction rates. Reaction Path I is based on the Al(OH)3 fragments model and is believed to be common when solutions are brought to supersaturation through simple dilution, or when systems are neutralized very rapidly, within seconds to hours. Under these conditions, little or no Al13 forms; the formation of Al(OH)3 nuclei is rapid, and gibbsite appears within days or weeks of aging. Reaction paths II and III both involve the formation of Al13. Path II is applicable when initial conditions favor formation of soluble Al13. The individual Al13 ions can remain in solution for months to years, but eventually they transfer by one of three pathways. If no Al13 nuclei are present, the soluble Al13 ions can slowly dissociate into Al3þ ions and deposit onto Al(OH)3 nuclei (Path IIa). These soluble Al13 may also aggregate and form Al13 nuclei (Path IIb). If some Al13 nuclei are already present, the remaining Al13 ions can be deposited onto them within weeks and months (Path IIc). Path III is applicable when Al13 forms under conditions that promote the rapid formation of Al13 aggregates via an anion bridging mechanism, i.e., outer-sphere associations.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

401

Figure 4 A proposed reaction scheme for Al hydroxide mineral formation from hydrolyzed aluminum solutions. From Huang et al. (2002).

These aggregates rapidly rearrange themselves in Al13 nuclei. The Al13 nuclei in Paths IIb, IIc, and III then transform into a poorly ordered phase, i.e., microcrystalline boehmite (pseudobehmite). This phase then transforms into gibbsite or sometimes bayerite. The Al(OH)3 fragments model and the Al13 polymer model are complementary. The best model to describe Al transformations in soils is still an open question particularly since the existence of Al13 in nature remains to be established. The relative importance of each pathway (Fig. 4) and thus the rate at which gibbsite forms in a given system depend greatly on reaction conditions. The hydrolysis and polymerization of Al and the subsequent transformation into more crystalline phases via the various pathways described above are strongly inﬂuenced by the nature and concentration of soluble inorganic and organic ions and solidstate ions such as clay minerals and humus (Huang and Violante, 1986; Huang, 1988; Bertsch and Parker, 1996; Krishnamurti et al., 1999; Huang et al., 2002). The inﬂuence of organic acids has been studied extensively, with most of the focus on the particular solid phases that form (Huang et al. 2002). The inﬂuence of a particular organic acid is generally related to the stability constant of the complex that the acid forms with Al (Table I). Therefore, p-hydroxybenzoic acid, which forms the least stable complex with Al, does not inhibit the crystallization of Al hydroxides, whereas aspartic, tannic, malic, and citric acids increasingly

402

P. M. HUANG

Table I Stability Constants of the Complexes Formed Between Al and Five Organic Acids at 258C Stability constants of the complexes Organic acids p-hydroxybenzoic acid Aspartic acid Tannic acid Malic acid Citric acid

log K1

log K2

1.66 2.60 3.78 5.14 7.37

– – – 8.52 13.90

From Kwong and Huang (1979a).

retard crystallization (Fig. 5). Besides the stability constant of the complex, the concentration of the organic acid is important. At certain low concentrations, the presence of some organic acids actually promotes the crystallization of particular Al(OH)3 polymorphs, but above a certain critical concentration, it disrupts crystallization (Huang and Violante, 1986). This is because organic acids replace water molecules that would otherwise coordinate with the Al3þ ion and the extent to which this occurs depends upon the chemical afﬁnity of the organic acid for the Al, that is its stability constant, and its concentration relative to Al.

Figure 5 The X-ray diffraction patterns of hydrolytic precipitation products of Al, showing how four different organic acids inﬂuence the transformation to more crystalline phases. The initial Al concentration was 1.1 £ 1023 M at an OH/Al molar ratio of 3 and the solution was aged for 40 d at room temperature in the presence of 1024 M organic acid. From Kwong and Huang (1979b).

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

403

Humic substances such as FA and HA also inﬂuence the transformation of Al by promoting the formation of microcrystalline boehmite and hampering the formation of more crystalline phases (Kodama and Schnitzer, 1980; Singer and Huang, 1990). Fulvic acids and HA resemble aliphatic acids, such as citric and malic acids, in that they contain COOH and aliphatic OH groups. They also resemble quercetin and tannic acid, because they contain phenolic hydroxyl and ketonic C ¼ O groups. Through these functional groups, FA and HA form stable complexes with Al and inhibit the crystallization of Al hydroxides. The inﬂuence of FA and HA also resembles that of salts in that they favor the formation of Al –O – Al (oxo) over that of Al – (OH) –Al (01) linkages. Organics can have an enormous impact on the surface properties of the Al transformation products. For instance, the presence of organic acids during aging of Al hydroxide gels for 40 d increases the speciﬁc surface of the products up to 30-fold over that of the control, and higher acid concentrations result in higher speciﬁc surface (Table II). The surface charge characteristics of the products also dramatically changed (Kwong and Huang, 1978, 1979c, 1981). The intermediate transformation products of Al, which include soluble mononuclear and polynuclear Al species and colloidal SRO Al hydroxides, are the most reactive Al species in inﬂuencing physical, chemical, and biological processes of soils and associated environments. These Al species inﬂuence environmental quality and ecosystem health by impacting acidiﬁcation of the terrestrial environment, formation of humic substances, development of soil structure, dynamics of organic C cycling and global warming, transformations of nutrients and pollutants, biological productivity and food chain contamination, microbial and enzymatic activity, aquatic ecosystem integrity, and human and animal health (Huang et al., 2002).

Table II Speciﬁc Surface Area of Al Hydroxide Precipitation Products Formed in the Presence of Tannic Acid and Selected Low Molecular Weight Organic Acids in Systems at Initial Al Concentrations of 1.1 £ 1023 M and OH/Al Molar Ratios of 3.0 and Aged for 40 d at Room Temperature Organic acid concentration (mol L21) Organic acid

1.0 £ 1026

1.0 £ 1024 (m2 g21)

None p-hydroxybenzoic Aspartic Tannic Malic Citric From Kwong and Huang (1981).

20 22 27 95 36 117

20 28 587 195 635 295

404

P. M. HUANG

B. IRON OXIDES The formation pathways of different Fe oxides in soils may be elucidated on the basis of observation of the pedogenic environment under which a certain phase or association of phases occur in a particular soil or soil sequence. Based on ﬁeld observations and laboratory synthesis under near-earth surface conditions, some general pathways of formation and transformation of Fe oxides in soil environments have been proposed (Schwertmann and Taylor, 1989; Cornell and Schwertmann, 1996). Soil environmental factors, which inﬂuence the modes of formation of Fe oxides, include the initial valence of the Fe source, its concentration in soil solution, the rate of oxidation of Fe (II), and parameters such as pH, Eh, temperature, soil moisture, soil organic matter, activity of Al, and other inorganic and organic ionic factors. Soil organic matter plays a vital role in the formation of pedogenic Fe oxides (Fig. 6). In soil environments where the amount of organic matter is low, for example in subsoils, the Fe supplied will form goethite and hematite depending on environmental factors (Schwertmann, 1985). As the organic matter content increases, more of the Fe will be complexed with organics leading to decreases in Fe activity. The activity of Fe(III) ions is so low that only the solubility product of goethite (10241 –10242), but not the solubility product of ferrihydrite (10237 – 10239) is exceeded. Consequently, goethite but not the ferrihydrite may form. Therefore, no hematite will form in an environment where the organic matter content is high, since ferrihydrite is deemed a necessary precursor for

Figure 6 Tentative schematic representation of the effect of organic matter content and rate of soluble Fe supply on the formation of various Fe forms in soils. From Schwertmann et al. (1986).

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

405

hematite. This trend is generally observed in the soils in the temperate and cool regions as well as in wet depressions and surface soils of the subtropical and tropical regions. At a higher content of organic matter, the rate of Fe supply is high, ferrihydrite will form and may survive for pedogenic times. If the content of organic matter is even higher such as in O horizons or in peaty environments, all of the Fe may be in the form of Fe –organic complexes. The interaction of organic matter with surface Fe may be the main mechanism for inhibiting the formation of crystalline Fe oxides in organic C-rich horizons. Laboratory studies verify this postulation and elucidate the mechanism of the inhibition (Cornell and Schwertmann, 1996). In an aqueous weathering environment, Fe oxides generally form via solution transformation. The oxidation products of Fe(II) solutions are, thus, important because it is in this valence state that Fe is commonly mobilized during weathering under the EH – pH region of natural soil environments. The kinetics of Fe(II) oxidation and the nature of Fe oxides formed as inﬂuenced by a series of organic ligands have been reported (Krishnamurti and Huang, 1990, 1991, 1993). The relative stability of the Fe(II) – ligand complex inﬂuences the rate of Fe(II) oxidation. The rate constant of Fe(II) oxidation at the same ligand/Fe molar ratio generally decreases with the increase in the stability constant of Fe(II) –ligand complex (Krishnamurti and Huang, 1990). In the absence of any complexing ligands, the precipitation product formed at pH 6.0 is dominantly goethite with small amounts of poorly crystalline lepidocrocite. In the presence of oxalate (log K ¼ 2.52, where K is the stability constant of the Fe(II) – ligand complex), the slower rate of formation of Fe(III) and its hydrolysis apparently modiﬁes the oxygen coordination in the edge-sharing Fe(O, OH)6 octahedron, resulting in the formation of lepidocrocite at the expense of goethite. The strong complexation of Fe(II) with tartrate (Log K ¼ 4.85) inﬂuences the kinetics of Fe(II) oxidation and Fe(III) hydrolysis and perturbs the crystallization of hydrolytic products of Fe(III), resulting in the formation of X-ray noncrystalline Fe oxides. The rate constant of Fe(II) oxidation also decreases with the increase in the perturbing ligand/Fe(II) molar ratio due to the increasing formation of the Fe(II) – ligand complex (Table III). The retardation of Fe(II) oxidation inhibits the nucleation of goethite and promotes the crystal growth of lepidocrocite especially at a citrate/Fe(II) molar ratio of 0.001. Further increase in the amount of citrate distorts the structural order, resulting in the increasing inhibition of crystal growth of lepidocrocite and the formation of noncrystalline Fe oxides. The surface of Fe oxides is the region of their interactions with the soil solution, organic and inorganic particles, plant root, microorganisms, and other soil biota. Therefore, surface properties of Fe oxides should have profound impacts on microaggregate formation, water ﬂux, nutrient and pollutant ﬂux, and the ability of soils to promote plant growth, to maintain reasonable soil biotic habitat, and to respond to management, and resist degradation. However, most studies on the surface chemistry of Fe oxides have been conducted on the two common Fe oxide

406

P. M. HUANG

Table III Rate Constants of Fe(II) Oxidation and Nature of Fe Oxides Formed in Absence and Presence of Citrate in Ferrous Perchlorate-NaOH System at pH 6.00 and 23.58C XRDa 020 peak of lepidocrocite Initial citrate/Fe molar ratio 0 0.0003 0.0005 0.0010 0.0050 0.0060 0.0080 0.0100 0.1000

Rate constant (min21) £ 104

Dominant Fe-minerals

WHH ^ 0.02 (82u)

Area (mm2)

41.3 ^ 2.0 28.6 ^ 3.5 26.9 ^ 1.8 24.4 ^ 3.0 20.2 ^ 1.8 19.1 ^ 1.5 16.8 ^ 0.6 15.2 ^ 0.7 7.6 ^ 0.7

G,L L,G L L L L L PC no ppt

1.40 1.00 1.00 0.80 1.10 1.20 1.20 n.a.b n.a.

40 ^ 5 97 ^ 5 124 ^ 12 263 ^ 3 82 ^ 4 25 ^ 3 10 ^ 3 n.a. n.a.

From Krishnamurti and Huang (1991). a X-ray powder diffraction data. WHH ¼ width at half height; L ¼ lepidocrocite; G ¼ goethite; PC ¼ poorly crystalline. b n.a. ¼ not applicable.

minerals, goethite and hematite. This is based on the assumption that all the Fe oxides could have similar surface properties. Xue and Huang (1995) reported that the presence of citric acid, one of the most common organic ligands in soils, during the formation of Fe oxides substantially modiﬁes the characteristics of surface properties of the oxide of Fe. Further, the ﬁne scale morphology, mean surface roughness, fractal dimension, speciﬁc surface, and microporosity of the Fe oxides depend on the citrate concentration in the solution in which the Fe oxides are formed (Liu and Huang, 1999). The surface properties of Fe oxides formed under the inﬂuence of organic ligands merit close attention in advancing our understanding of their surface chemistry pertaining to dynamics and transformations of nutrients and pollutants in terrestrial and aquatic environments.

IV. ROLE OF SOIL MINERALS IN ABIOTIC CATALYSIS OF THE FORMATION OF HUMIC SUBSTANCES A. POLYPHENOL PATHWAY The mechanisms of the formation of humic substances are very complicated processes in which a variety of organic components, such as phenolic compounds, carbohydrates, and nitrogenous substances, can participate as

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

407

starting materials. Humic substances can be synthesized through biotic and abiotic processes. Soil minerals play an important role in catalyzing the abiotic polymerization of phenolic compounds and the polycondensation of phenolic compounds and amino acids and the subsequent formation of humic substances. Kumada and Kato (1970) are among the pioneers in the study of browning of pyrogallol catalyzed by clay-size layer silicates. In a series of studies, Wang and his co-workers reported that layer silicates catalyze the abiotic formation of humic substances through oxidative polymerization of phenolic compounds common in soil environments (Wang and Li, 1977; Wang et al., 1986). Intracrystal surfaces of transition metal saturated layer silicates can catalyze the formation of aromatic cation (Pinnavaia et al., 1974). Aromatic molecules may donate electrons to metal ions, such as Cu((II) and Fe(III), on the cation-exchange complex of smectite (Mortland and Halloran, 1976). The ability of muscovite to catalyze the oxidation of some phenolic compounds to form HA has also been demonstrated (Filip et al., 1977). Since the early 1980s, Huang and co-workers have studied the sequence of catalytic power of layer silicates and their reaction sites in the polymerization of phenolic compounds and the subsequent formation of humic substances. The promoting effect of 2:1 layer silicates is higher than 1:1 layer silicates (Shindo and Huang, 1985a). This is attributed to the higher speciﬁc surface and higher lattice imperfection of the former than the latter. The edges of kaolinite are virtually the only catalytic sites for the formation of polyphenolderived humic substances. The edges of nontronite have a very important role as catalytic sites in the formation of hydroquinone-derived humic macromolecules (Wang and Huang, 1988). The Fe(III) in the octahedral sheet of nontronite also serves as a Lewis acid site to catalyze the oxidative polymerization of hydroquinone (Wang and Huang, 1986). Further, nontronite has the ability to cleave the ring of pyrogallol, catechol and hydroquinone; the extent of ring cleavage of the polyphenols to release CO2 varies with their structures (Wang and Huang, 1994). Iron oxides and oxyhydroxides can catalyze polymerization of hydroquinone (Scheffer et al., 1959). The importance of hydrous oxides of Fe in catalyzing the oxidative polymerization of phenolic compounds is inﬂuenced by the structure and functionality of phenolic compounds. Hydrous oxides of Al also catalyze the oxidative polymerization of phenolic compounds (Wang et al., 1983a; Shindo and Huang, 1984; Wang and Huang, 2000a). Aluminum and Fe oxides and oxyhydroxides may form surface coatings on layer silicates and humic substances. The ability of various Al and Fe oxides and oxyhydroxides and their surface coatings on soil particles to promote oxidative polymerization of phenolic compounds remains to be uncovered. Among metal oxides, hydroxides, and oxyhydroxides, Mn oxides are the most powerful catalysts in the transformation of phenolic compounds (Shindo and Huang, 1982, 1984). Manganese oxides (birnessite, cryptomelane, and pyrolusite), which are common in soils, act as

408

P. M. HUANG

Figure 7 Solid-state 13C NMR spectra of synthesized HA: (a) pyrogallol and (b) birnessite– pyrogallol systems. The dry, powdered HA formed in the reaction systems was analyzed with cross-polarization and magic-angle spinning (CPMAS), proton decoupling; contact time 2.00 ms, recycle time 1.00 s, spectral width 20,000.0 Hz, spinning speed 3700 Hz, number of scans 60,000 and spectrometer frequency 37.7 MHz. The band positions and assignments for the spectra are as follows: 196.3 ppm (carbons in aldehydes and ketones), 180.9– 170.7 ppm (carboxyl carbons

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

409

Lewis acids that accept electrons from phenolic compounds, leading to their formation of semiquinone and their oxidative polymerization and formation of humic substances. The solid state cross-polarization magic-angle spinning (CPMAS) 13C NMR spectrum of HA (MW . 1000) formed in the Mn (IV) pyrogallol system (Fig. 7b) is very similar to those of HAs extracted from natural soils (Hatcher et al., 1981; Schnitzer and Preston, 1983). The spectrum (Fig. 7b) and the spectra of natural humic substances in the cited references show clearly the presence of carbons in aldehydes and ketones (196.7 ppm), carbonyl carbons in carboxyl and ester groups (173.3 ppm), aromatic carbons (141.4, 132.5, 120.3 ppm), CO carbon– alcohols, esters, ethers, carbohydrates (79.4 ppm), and alkyl carbons (41.3, 34.4, and 19.6 ppm). However, the spectrum of synthesized HA (MW . 1000) formed in the pyrogallol system in the absence of Mn(IV) oxide (Fig. 7a) is very different from those of natural humic substances. Wang and Huang (2000a, b) reported that the catalytic power of metal oxides in the abiotic ring cleavage of pyrogallol and the polycondensation of the resultant fragments varies greatly with the nature of the oxides. The abiotic ring cleavage of polyphenols catalyzed by SRO oxides of Al and especially Fe and Mn (Wang and Huang, 2000a, b) may account, in part, for the carboxylic group contents and the origin of the aliphaticity of humic substances in soils (Schnitzer, 1977; Preston et al., 1982). Therefore, the catalytic power of these oxides in affecting the C turnover and humic substance formation via abiotic processes in soils and related environments deserve attention. Poorly crystalline aluminosilicates are common in soils (Wada, 1989). Allophane has the ability to catalyze the polymerization of polyphenols (Kyuma and Kawaguchi, 1964). Humiﬁed polyphenols resemble natural humic substances in their infrared spectra (Wang et al., 1983b). There are a series of poorly crystalline aluminosilicates in soil environments (Huang, 1991). Their ability to catalyze the transformation of phenolic compounds to humic substances remains obscure. Primary minerals differ in their ability to catalyze the abiotic polymerization of hydroquinone. The sequence of the catalytic power of primary minerals is tephroite . actinolite . hornblende . fayalite . augite . biotite . muscovite < albite < orthoclase < microcline < quartz (Shindo and Huang, 1985b). Except for the tephroite, which is a Mn-bearing olivine, the hydroquinone-derived products are largely low-molecular weight humic polymers.

in carboxyl and ester groups), 152.5–111.4 ppm (aromatic carbons and oleﬁnic carbons), 82.9 ppm (CO carbons—alcohols, esters, ethers, and carbohydrates), and 43.0 ppm (alkyl carbons). The formation of alkyl carbons, CO carbons, and carboxyl carbons in (b) is due to the ring cleavage and fragmentation of pyrogallol catalyzed by birnessite. From Wang and Huang (1992).

410

P. M. HUANG

B. MAILLARD REACTION The Maillard reaction (Maillard, 1913) is thought to be a major pathway in humiﬁcation because of signiﬁcant similarities between humic substances and melanoidins formed through this pathway involving sugar-amino acid condensations (Ikan et al., 1996). Evershed et al. (1997) detected the presence of characteristic products of the Maillard reaction (alkyl pyrazines) in archaeological plant remains up to 1500 years from Egypt. Despite the importance of the Maillard reaction, the mechanisms and rates of polycondensations of sugars and amino compounds in nature remain obscure (Ikan et al., 1996). Jokic et al. (2001a) reported that birnessite (d-MnO2) signiﬁcantly increases the extent of humiﬁcation of the glucose –glycine system over the pH range of 6– 8 (Fig. 8). They veriﬁed the formation of humic substances in the glucose – glycine – birnessite system by measuring the CPMAS 13C NMR spectrum of the isolated FA. The chemical shifts of FA formed in the Maillard reaction system (Jokic et al., 2001a) resemble those of natural humic substances such as soil and stream FAs (Malcolm, 1989; Schnitzer, 2000). Their data clearly show that birnessite greatly accelerates the Maillard reaction between glucose and glycine forming polycondensation products resembling humic substances under ambient conditions. Furthermore, light intensity of 168 m E S21m22 exerts an increase in the browning of the glucose –glycine –birnessite system compared with the same system, which is kept in the dark (Jokic et al., 2001b). Equally important is that even in complete darkness, birnessite catalyzes the Maillard reaction between glucose and glycine. Therefore, birnessite catalysis can occur in soil or sediment environments at any depth, but the presence of sunlight should strongly accelerate the reaction.

V. INTERACTIONS OF SOIL MINERAL COLLOIDS WITH MICROORGANISMS A. SURFACE INTERACTIONS Soil is a habitat for myriads of microorganisms. Although the microbial biomass constitutes only a very small proportion (, 3%) of the total organic carbon in soil, it is the most active and dynamic fraction of the living organic pool. Among the inorganic components of soil, mineral colloids are the most reactive because of their large speciﬁc surface and reactive charge characteristics. Being enriched in ions, water, and organic matter relative to the bulk soil, the surface of mineral colloids serves as a preferred habitat for soil microorganisms. That complex formation between mineral colloids and microorganisms can and does occur, when the components are brought together in the laboratory, has been amply documented (Marshall, 1980).

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

411

Figure 8 Absorbance versus wavelength plots in the Maillard reaction between glucose and glycine as inﬂuenced by birnessite catalysis. (a), (b) and (c): 30 d reaction period. (d), (e) and (f): 15 d reaction period. From Jokic et al. (2001a).

Although the surface of bacterial cells and crystalline clay minerals are both negatively charged, bacteria have a propensity for producing extracellular polysaccharides (EPS) which bind simultaneously to cell and clay surfaces through cation bridging involving polyvalent cations (Fig. 9). Cation bridging may be direct [Scheme (1)] or effected by H-bonding to water molecules in the primary hydration shell of polyvalent cation [Scheme (2)]: ½Clay2 · · · Mnþ · · · 2 ½EPS – B

ð1Þ

or 2 ½Clay21 · · · Mnþ · · · H OH · · · ½EPS – B

ð2Þ

412

P. M. HUANG

Figure 9 Diagram illustrating the interaction of bacteria and fungi with mineral particles in a soil aggregate. Bacterial cells with a coat of extracellular polysaccharides (EPS) are enveloped by clay particles. The pore space where clays and bacteria interact, bounded by silt- and sand-size particles, is relatively enriched in organic matter including EPS residues. Fungal hyphae are attached to the outside surface of an aggregate. Inset shows an enlarged view of a bacterial cell with its complement of EPS. At normal soil pH conditions, the cell has a net negative surface charge. Most clay particles adhere to the cell surface by bridging through polyvalent cations, represented by Mnþ (a) although some may be attached directly by electrostatic interactions, either in face-to-face (b), or edge-to-face (c) association. From Theng and Orchard (1995).

where Mnþ denotes a polyvalent cation of valency n, and B is a bacterial cell (Theng and Orchard, 1995). The predominant mechanism would depend on the nature of the polyvalent cation but, more so, on the hydration status of the soil. Dehydrating conditions would clearly favor mechanism (1). In addition, shortrange interactions (van der Waals, H-bonding) may come into play because the EPS chains are induced to make close contact with the surface, increasing the net energy of interactions. Mineral colloids in soil are generally coated with hydroxy Al (or Fe) polymers in soil (Fig. 1). These coated minerals thus behave as a positively charged species

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

413

or display amphoteric characteristics. Therefore, mineral colloids can strongly interact with negatively charged microbial cells in soil environments. This type of bonding, which is much stronger than cation bridging, is also expected to occur with Al and Fe oxides over the pH range of soils. The attachment of microorganisms to SRO minerals and the crystal edges of layer silicates through electrostatic interactions would also be predicted to occur when the soil pH falls below 6 because all of these surfaces would then be positively charged. Furthermore, in the majority of cases, minerals in topsoils are partially covered with organic materials, especially humic substances (HS), which are microbially resistant and the most prevalent. The most common mode of mineral colloid-organic material-microorganism interactions may be represented by Scheme (3). ½Mineral colloid – HS2 · · · Mnþ · · · 2 ½EPS – B

ð3Þ

In humic-rich calcareous Mollisols, Ca would be the predominant bridging cation. In Andisols, Oxisols, Ultisols, and the B horizon of Spodosals, HS largely occur as complexes with Al and Fe, or their respective poorly crystalline oxides (Theng et al., 1989; Oades et al., 1989). In soils with little organic matter and in subsoils, mineral colloid– microorganism interactions are largely inﬂuenced by the mineralogical composition and pH of the system. To date, the research data on the interactions of microorganisms with crystalline layer silicates with different structural conﬁgurations and surface properties are limited and even less information is available with respect to SRO minerals and Al and Fe oxides. Much of the evidence on the mechanisms for direct in situ-as opposed to in vitro-surface interactions of mineral colloids with microorganisms is circumstantial and equivocal (Stotzky, 1986; Theng and Orchard, 1995).

B. INFLUENCE

ON MICROBIAL AND SURVIVAL

ACTIVITY

Mineral colloids, in general, can directly or indirectly inﬂuence the activity of microorganisms in their immediate vicinity. The effect of mineral colloids may be positive, negative, or sometimes so small as to escape detection. Mineral colloids have a stimulating effect on the activity of adhering bacteria by keeping the pH of microhabitats within the optimum physiological range for growth. The content and type of mineral colloids are inﬂuential in determining the balance between different microbial

414

P. M. HUANG

populations in soil. A well-known example is the failure of some fungi to thrive and spread in certain soils (Stotzky, 1986). This is largely attributed to the presence of montmorillonite in the suppressive soils. Montmorillonite can serve as a proton sink and is thus able to promote the growth of acidsensitive bacteria in these soils. This gives bacteria a selective advantage over fungi in competing for available nutrients. Therefore, fungal growth and proliferation are effectively suppressed. Mineral colloids can also stimulate microbial activity by sorbing metabolites that would otherwise have an adverse effect on microbial growth (Filip et al., 1972; Filip and Hattori, 1984). Montmorillonite is more effective than kaolinite and ﬁnely ground quartz. Furthermore, surfaces of mineral colloids can adsorb other toxic substances to microorganisms such as antibiotics and pesticides (Theng and Orchard, 1995). Adhering microorganisms may beneﬁt from being close to nutrients adsorbed on mineral colloids. However, the potential substrates may not be readily available or physically accessible (Fletcher, 1991). Furthermore, beyond a certain concentration, the addition of mineral colloids to the system may result in a reduction in microbial activity (Marshall, 1971). This is attributable to restricted diffusion of oxygen and nutrients to microbial cell as it is progressively enveloped by mineral colloids. However, by forming an envelope around bacteria cells, mineral colloids may provide protection from extreme ﬂuctuations in physicochemical environments, and thus enhance bacterial survival (Stotzky, 1986; Theng and Orchard, 1995). The protective effect of mineral colloids, especially montmorillonite, is manifested in the ability of soil microorganisms to withstand exposure to hypertonic osmotic pressure, dessication, and ultraviolet radiation. Besides serving as a substratum for bacteria to adhere to and as a protective envelope of bacterial cells, mineral colloids can act as a cementing agent of soil particles. Therefore, the addition of clay to soils, especially those of light texture, would modify the spatial arrangement of particles and especially the pore –size distribution within soil aggregates. Such a modiﬁcation of aggregate structure often beneﬁts the bacterial population by increasing the proportion of pores of a certain size range (, 6 mm) from which bacteria could freely enter to colonize pores but bacterial predators notably protozoa are effectively barred from entering such pores due to steric hindrance. Another survival mechanism is adhesion of bacteria to mineral colloids, since these surfaces are relatively enriched in substrates and water. Even if they lack available nutrients, solid surfaces may trigger cells to diminish in size, resulting in an increase in the surface-tovolume ratio and packing density (Kjelleberg et al., 1982). Such bacterial cells can respond rapidly to conditions of high nutrient and water supply and regain their normal size and growth pattern.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

415

VI. INTERACTIONS OF SOIL COLLOIDS WITH ENZYMES A. MINERAL COLLOID –

AND

HUMIC – ENZYME COMPLEXES

Extracellular enzymes are rapidly sorbed at mineral colloids and humic substances in soil environments. Mineral colloids have a high afﬁnity for enzymes although this is not always synonymous with the retention of catalytic ability. Enzymes associated with soil humic substances retain their activity for long periods. Mineral colloid –enzyme interactions have been documented (Theng, 1979; Burns, 1986; Stotzky, 1986; Naidja et al., 2000; Violante and Gianﬁeda, 2000). The mechanisms whereby enzymes are retained by surfaces of mineral colloids have usually been assumed to include cation – exchange reactions but not all observations have supported this view. If this were true, one would expect the adsorption to increase as pH values decrease below the isoelectric point of the enzyme. Under these conditions, the protonation of amino groups should give rise to an increase in positive charges, resulting in more adsorption to mineral colloids. However, the adsorption of polysaccharases is highest at their isoelectric points when the enzymes have no net charge. Therefore, besides cation-exchange reactions, adsorption of enzymes by mineral colloids may proceed through ionic, covalent, hydrophobic, and hydrogen bonding, and van der Waals forces. Although individual van der Waal’s forces are regarded as weak, retention may be cumulative if the enzyme is in close contact with the adsorbing surface. Naidja et al. (2002) reported that the tyrosinase protein have a very high afﬁnity to the surfaces of hydrous manganese oxide birnessite. Analysis by atomic force microscopy (AFM) conﬁrms the globular structure and spheroid shape of the protein molecules and clearly shows the unfolding in an ellipsoidal shape of the protein molecules after adsorption and immobilization on the mineral surfaces, especially the frayed edges of the phyllomanganate layers (Fig. 10). AFM pictures along with X-ray diffraction and infrared absorption data of the protein– birnessite complex show the coating of the mineral surfaces by the protein molecules as well as the changes in the protein conformation, resulting in the unfolding and ﬂattening of the protein molecules after binding to mineral surfaces (Naidja et al., 2002). The information obtained is of fundamental signiﬁcance for understanding the nature of enzyme –mineral complexes. Relatively few studies have been made of enzyme adsorption to humic substances, probably because of the biochemical heterogeneity, dynamic nature, and physicochemical complexity of soil organic matter. Nevertheless, it is well

416

P. M. HUANG

Figure 10 Atomic force micrographs of wet samples of free protein and protein–birnessite complexes after immobilication of the protein at an initial protein concentration of 3 g/g, pH 6.0, and room temperature (238C) without air-drying: (a) Free tyrosinase sample, (b) protein – birnessite complex formed from the tyrosinase sample, (c) free bovine serum albumin (BSA), and (d) protein–birnessite complex formed from the BSA. The same scale was used from (a) to (d). From Naidja et al. (2002).

known that enzymes are stabilized by association with soil organic matter (Conrad, 1942; Burns, 1986; Nannipieri and Gianfreda, 1999). Many mechanisms have been proposed to account for the stability of enzymes which are complexed with soil humic substances. These include ion exchange, H-bonding, entrapment within three-dimensional micelles, lipophylic reactions, and covalent bonding during organic matter genesis. Although it is not inconceivable that some or even all of these reactions and processes are involved, the intrinsic mechanisms of interactions of enzymes with humic substances are still obscure. Functional groups of enzymes implicated in covalent bonding to humic polymers include terminal and basic amino, carboxyl, sulfhydryl, phenolic, and imidazole groups. All these may be involved in the stabilization process provided they do not form part of the active sites of the enzyme and not crucial to the retention of its tertiary structure. Enzyme– humic complexes can also be attached to mineral colloids, and in some instances, this further enhances enzyme stability. Recently, Violante and Gianfreda (2000) reported the stability of enzymes bound to organomineral complexes. Their ﬁndings are basic to understanding enzymatic activities in soils and related environments.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

B. EFFECTS

ON

417

ENZYMATIC ACTIVITY

Enzymes adsorbed or immobilized on mineral colloids operate in a structurally restricted system. The behavior of such enzymes toward their respective substrate is, thus, different from that occurring in a homogeneous, aqueous solution. When enzymes are adsorbed on mineral colloids, changes in the tertiary structure (i.e., the folding of the helix or coil into a compact, globular molecule stabilized by interfold hydrogen bonding, van der Waals and hydrophobic interactions) of the enzymes and their active sites decreases activity or eliminate it altogether (Burns, 1986). However, there are notable exceptions to the adsorption-decline in activity rule. Gainfreda and Bollag (1994) investigated the behavior of a laccase and a peroxidase in the presence of a montmorillonite, a kaolinite, and a silt loam soil. The various supports show different enzyme immobilization capabilities (Table IV). There is considerable variation in the retained activities of the two enzymes on the four supports, as well as the same enzyme on the four supports. The residual speciﬁc activities (calculated as percentages of the speciﬁc activity of the free enzyme) of laccase and peroxidase

Table IV Immobilization of a Laccase (from Trametes versicolor) and a Peroxidase (from Horseradish) on Different Supports Enzymatic activity Enzyme and support Laccase Glass beads Montmorillonite Kaolinite Soil Peroxidase Glass beads Montmorillonite Kaolinite Soil

Protein Adsorbeda (mg/%)

Units adsorbedb

Speciﬁc activityc

0.452/56 0.622/71 0.566/64 0.644/73

28.8 19.8 13.1 15.7

63.7 31.8 23.1 24.4

236 118 85.5 90.4

0.092/17 0.224/43 0.120/23 0.162/31

8.4 23 9.5 15

91.6 102.8 78.9 92.6

93.8 105.2 80.7 94.8

Residual speciﬁc activityd(%)

From Gianfreda and Bollag (1994). a Difference between proteins initially added to 200 mg of support (0.88 mg laccase and 0.52 mg of peroxidase) and those recovered in the supernatant and washings. b Expressed as mmol O2 consumed min21 for laccase and mmol guaiacol transformed min21 for peroxidase. c Units adsorbed/protein adsorbed. d Calculated as percentage of the speciﬁc activity (sa) of the free enzyme (laccase, sa ¼ 27 mmol min21 mg21; peroxidase, sa ¼ 97.7 mmol min21 mg21).

418

P. M. HUANG

immobilized on all supports are high. Further, laccase immobilized on montmorillonite shows speciﬁc activities higher than that of the free enzymes. This may be attributable to steric modiﬁcation of the immobilized enzyme. The data clearly indicate that the performance of enzymes in soil environments is signiﬁcantly affected by soil mineral colloids. The study of interactions of enzymes with soil inorganic constituents was mainly concerned with crystalline clay minerals. The role of SRO oxides, hydroxides, and oxyhydroxides of Al, Fe, and Mn in inﬂuencing enzyme activity involved in the transformations of natural organics and xenobiotics (Huang, 1990) has now been recognized (Naidja et al., 1997, 2000; Huang et al., 1999; Violante and Gianfreda, 2000) and merits further attention. A signiﬁcant portion of immobilized enzymes in soils is associated with the humic fraction, not by adsorption or occlusion, but by the formation of enzyme – phenolic copolymers during the genesis of humic substances (Burns, 1986; Huang, 1990). These immobilized enzymes are active and extremely stable. They can be adsorbed on mineral colloids. The catalytic role of soil minerals with different structural conﬁguration and surface properties in the formation of enzyme– phenolic co-polymers and the subsequent impact on the activity and stability of enzymes in the environment deserve close attention.

VII. MICROBIAL MEDIATION OF SOIL MINERAL WEATHERING AND TRANSFORMATION A. MINERAL WEATHERING Chemical weathering of minerals, as a part of the soil formation process, can be enhanced by microbial activity by a factor as high as 106 (Kurek, 2002). Microorganisms can dissolve minerals by direct and indirect action under aerobic and anaerobic conditions (Robert and Berthelin, 1986; Ehrlich, 2002; Kurek, 2002). The modes of attack of minerals by microorganisms include: (1) direct enzymatic oxidation or reduction of a reduced or oxidized mineral component, (2) indirect attack with a metabolically produced redox agent or inorganic and organic acid, (3) indirect attack by metabolically produced alkali, usually in the form of ammonia, (4) indirect attack with a metabolically produced ligand that form a highly soluble product with a mineral component, and (5) indirect attack by biopolymer. The mode of microbial attack of a mineral may involve a combination of some of these mechanisms. In some cases of attack, the microorganisms may be dispersed in the soil solution; in others, they may grow in bioﬁlms on the surface of the susceptible minerals. When oxidized metal compounds such as Fe(III), Mn(IV), and As(V) act as electron acceptors, anaerobic respiration becomes an example of

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

419

direct dissolving action under anaerobic conditions (Ahmann et al., 1994; Ehrlich, 2002). Oxidation of sulfur entities of metal sulﬁdes to obtain energy is an example of direct dissolving action under aerobic conditions (Kurek, 2002). Volatization of metals and metalloids or biomethylated metal and metalloid compounds from the soil into the atmosphere can be a mechanism of detoxiﬁcation for toxic elements such as Hg, As, and Se (Gadd, 1993). Uptake and concentration of elements by absorption and/or adsorption were observed with bacteria and fungi for P, Ca, Fe, and K. Weed et al. (1969) have shown that fungi can adsorb K from solution and thus shift K equilibrium in suspensions of trioctahedral and dioctahedral micas and transfer them to vermiculites. Such process can also occur for many major and trace elements (Robert and Berthelin, 1986). Some microorganisms can promote the transformation of one mineral into another by a process called diagenesis. Mineral diagenesis can be an indirect effect of aerobic and anaerobic microbial metabolism (Ehrlich, 2002; Kurek, 2002). The formation of a new mineral can be resulted from a chemical reaction between a product of microbial dissolution of a mineral and appropriate cations present in the environment.

B. FINE-GRAINED MINERAL DEVELOPMENT Geochemical modeling of metal speciation and transport is only beginning to include bacteria as geochemically active surfaces. The physical and chemical characteristics of bacteria, such as their large surface area-to-volume ratio, serve to increase the metal-binding capacity of their charged surfaces leading to precipitation and formation of mineral phases on their cell walls or other surface polymers (McLean et al., 2002). The mechanisms by which bacteria initiate the formation of minerals in bulk solution vary widely between species. There may be a combination of biochemical and surface-mediated reactions during the process. Bacteria surface layers may passively adsorb and indirectly serve as a nucleation template. Bacteria can also more directly initiate mineral precipitation by producing reactive compounds (e.g., enzymes, metallothioneins, siderophores), which bind metals or catalyze metal transformations. Further, bacteria can instigate the spontaneous precipitation of metals by altering the chemistry of their microenvironments (Beveridge et al., 1997; Douglas and Beveridge, 1998). Metal complexing ligands such as metallothioneins and siderophores serve to sequester metals from the environment for incorporation into cellular components (McLean et al., 2002). Reactive inorganic ligands such as sulﬁde

420

P. M. HUANG

and phosphate may also be produced as cellular metabolic by-products. Sulﬁde reacts with metal ions to form metal sulﬁdes, a common reaction in anoxic environments with sulfate-reducing bacteria. Phosphate pumped out of cells can also react with metal ions to form cell-associated and extracellular precipitates. Another example of microbially mediated ﬁne-grained mineral development is the formation of Mn oxides. Microbial oxidation of Mn(II) is a major process that can produce Mn oxide coatings on soil particles 105 times faster than abiotic oxidation (Tebo et al., 1997). This microbially mediated formation is illustrated in Fig. 11. Manganese oxides are highly reactive minerals and help restrict the mobility of metals in soils and sediments through adsorption on their surfaces. Biogenic Mn oxides formed by Leptothrix discophora SS-1 have signiﬁcantly higher Pb adsorption capacity and larger speciﬁc surface than abiotically precipitated Mn oxides (Nelson et al., 1999). Therefore, bioformation of minerals should have a signiﬁcant bearing on remediation of metal contamination.

Figure 11 Thin section of Leptothrix sp., which is precipitating manganese oxide on its outermost structure called a sheath. The arrows point to the manganese mineral phase identiﬁed by EDS. Scale bar ¼ 150 nm. From McLean et al. (2002).

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

421

VIII. INTERACTIONS OF SOIL MINERAL COLLOIDS WITH ORGANIC SUBSTANCES AND MICROORGANISMS IN RELATION TO SOIL STRUCTURE STABILITY A. ORGANO-MINERAL COMPLEXES TO SOIL STRUCTURE

IN

RELATION

The processes by which soil organic matter strengthens the bonds between soil mineral particles are complicated. Root exudation and microbial action produce organic compounds with a range of composition and molecular weights; these compounds interact with the mineral particles, which also vary in size, shape, crystallinity, and electric charge (Emerson et al., 1986). Interactions between soil mineral particles, organic substances, and microorganisms can occur at many different size scale, as these materials have a large size range in soils (Fig. 12). Therefore, it is important to indicate the size scale being considered when discussing soil structure and the mechanisms of its stabilization, because the potential mechanisms for stabilization vary with aggregate size. The adsorption of organic molecules such as microbially derived polysaccharides and other unaltered and altered biomolecules onto mineral surfaces can enhance the stability of individual clay microstructures. Adsorption is also important in binding together the clay microstructures and silt particles into small microaggregates with 2 – 50 mm diameters and densities . 2.0 Mg m23 (Baldock, 2002). The high stability of these small microaggregates is demonstrated by their resistance to ultrasoniﬁcation. Many microaggregates exist as pieces of fungal hyphae, bacteria or bacterial colonies coated with EPS and clay minerals (Oades and Waters, 1991). The polysaccharides are present throughout the matrix but concentrated in pores between clay microstructures. Particulate organic matter (POM) are, important stabilizing agents at larger size scales: large microaggregates and small macroaggregates (Jastrow and Miller, 1997). Soil structure can be stabilized by POM through two mechanisms related to its physical properties and its susceptibility to biological decomposition (Baldock, 2002). POM can bridge the failure zones that exist between adjacent stable aggregates. The bridging can result from a combination of binding to aggregate surfaces, penetration through aggregates, and the formation of a network capable of holding groups of aggregates together. The ability of POM to stabilize soil structure by this bridging mechanism is controlled by its physical size and morphology. POM can also enhance the stability of soil structure by providing a substrate for microorganisms to enhance the production of fungal hyphae and microbial metabolites such as polysaccharides. The two mechanisms of structural stabilization by POM can operate independently or synergistically. POM is often found at the core of microaggregates , 250 mm (Oades and Waters, 1991). The size of the microaggregate formed depends on the size of

422

P. M. HUANG

Figure 12 Size scales associated with soil mineral particles, organic components, pores and aggregations of mineral and organic components. The deﬁnitions of pore size have used those developed by IUPAC (micropores , 2 nm, mesopores 2 – 50 nm and macropores . 50 nm). Alternatively, the pore sizes corresponding to the lower (Cm ¼ 2 1500 kPa) and upper (Cm ¼ 2100 kPa) limits of water availability to plants may be used to deﬁne the boundaries between the different classes of pore size. Cm is soil water metric potential. From Baldock (2002).

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

423

Figure 13 A conceptual model of aggregate hierarchy in soils where organic materials play an important role in the stabilization of aggregates. From Jastrow and Miller (1997).

the POM and the nature and amount of binding materials secreted by microorganisms as decomposition proceeds. While POM continues to provide a substrate for microorganisms, the production of biochemical aggregating agents continues, and structural stability is maintained. Mechanisms of stabilization of soil structure can operate over larger distances to bind microaggregates together to form macroaggregates as illustrated in Fig. 12. In view of the distances involved, the stabilization of macroaggregates is related to the presence of nonliving POM capable of spanning distances . 100 mm or the existence of a network of fungal hyphae and plant roots that physically enmeshes microaggregates (Fig. 13). The death of roots and hyphae growing within and through macroaggregates results in the formation of biochemical binding agents capable of stabilizing the structure of macroaggregates.

B. DYNAMICS

OF

AGGREGATE TURNOVER

Except for the biologically unavailable char/charcoal fraction, all components of soil organic matter contribute to the stabilization of soil structure in mineral soils. The diverse range of living and nonliving components varying in size, chemical composition, and extent of decomposition provides soil organic matter with the capacity to stabilize aggregations of soil particles at size scales varying

424

P. M. HUANG

over nine orders of magnitude (Fig. 12). Therefore, it is essential to maintain the proper balance of components of soil organic matter to ensure the stability of the entire soil matrix. Since organic matter responsible for the stabilization of soil structure is not inert and thus subject to decomposition, aggregation is a dynamic process in soils (Baldbok, 2002). Biological activity has the potential not only to stabilize soil structure through the production of organic substances capable of binding soil particles, but also to destabilize soil structure by decomposing organic binding agents. The balance between these two processes dictates the level of soil structural stability. A continual addition of organic matter is essential to ensure that the contents of each type of aggregating agent are adequate to maintain structural stability of soils. Therefore, management that provides an ongoing addition of organic materials has the greatest potential to maintain or enhance soil structural stability. The model proposed by Golchin et al. (1997) can be used to relate the physical and chemical properties of soil organic materials and their distribution and cycling to the stabilization of soil structural form (Fig. 14). The major organic materials contributing to the structural stability of soils are perceived to be free and occluded POM fractions and the biomolecules synthesized and exuded by roots, mycorrhizal fungi, and other microorganisms in soils. This model depicts the dynamics of soil aggregation and is most applicable to the aggregation of soils with appreciable contents of clays that protect aggregating agents from rapid decomposition (Baldock and Skjemstad, 2000). In sandy soils, the protection mechanisms that stabilize organic binding agents from microbial attack are less effective. Therefore, aggregation in sandy soils is a much more dynamic process. The large size of the mineral particles and pores in sandy soils reduces the effectiveness of polysaccharides and other unaltered and altered biomolecules to stabilize soil structure. These biomolecules are important to the adherence of soil particles to plant roots and fungal hyphae and soil microorganisms to mineral particles. However, in sandy soils, they do not have the capacity to span the distances between the large primary particles and stabilize soil structure. In the model depicting the dynamics of soil aggregation (Fig. 14), freshly deposited POM is typically present as the free POM fraction of soil organic matter. The binding of soil particles to free POM is limited. Therefore, POM has little direct inﬂuence on soil structural stability (Stage 1, Fig. 14). Once free POM is colonized by soil microorganisms and decomposition is initiated, mineral particles adhere to the extracellular mucilage released by decomposer organisms (Stage 2, Fig. 14). The remaining free POM becomes occluded within a new macroaggregate. Further decomposition of occluded POM results in a decrease in the size of POM through preferential utilization of exposed POM that has not been encapsulated with mineral particles. The ability of pieces of POM to maintain the stability of the original macroaggregate decreases and the macroaggregate deteriorates into smaller microaggregates on exposure to

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM 425

Figure 14 A model depicting the dynamics of soil aggregation and the roles of particulate organic matter (POM) and microbial metabolites in the stabilization of soil aggregates. From Baldock (2002).

426

P. M. HUANG

disruptive forces (Stage 3, Fig. 14). Once carbohydrates are removed from the microaggregate POM, the more recalcitron cores are no longer capable of maintaining the microaggregate stability. The microaggregate then deteriorates into aggregations of soil particles bound together with remnant microbial mucilage and other unaltered and altered biomolecules (Stages 4 and 5, Fig. 14). Although mechanisms exist that can contribute to some protection against biological attack, all components of soil organic matter involved in aggregation decompose with time. A continuous input of organic materials, mainly from plant production is, thus, essential to maintain or enhance the stability of soil structure.

IX. INFLUENCE OF MINERAL COLLOIDS ON BIOGEOCHEMICAL CYCLING OF C, N, P, AND S IN SOIL A. ROLE

OF

MINERAL COLLOIDS IN SOIL ORGANIC MATTER STORAGE AND TURNOVER

Most of the organic C in soils is degraded to inorganic forms slowly, on timescales from centuries to millennia. Soil minerals are known to play a stabilizing role in soil organic matter. The interaction of Al and Fe with humus is of primary importance in the determination of the content of organic matter in tropical and temperate soils (Wada, 1995). The Al and Fe that complex and stabilize humus against microbial decomposition are released from soil minerals during soil formation and present as polymeric hydroxy cations. Their supply rates apparently control the content of soil organic matter to a great extent, as demonstrated by the relationship between pyrophosphate-extractable C and pyrophosphate-extractable Al plus Fe (Fig. 15). Torn et al. (1997) used radiocarbon analyses to explore interactions between soil mineralogy and soil organic C along two natural gradients — of soil-age and of climate — in volcanic soil environments. The total stock of organic C in soil increases with substrate age up to 150 kyr, to 60 kg m22, and then decreases to 31 kg m22 at the oldest site (Fig. 16a). Carbon inventory in the surface horizons (O and A horizons) reaches its maximum at 20 kyr, and varies less with age than did deep soil C. Most of the decrease in soil organic C stored in older substrates is attributed to faster turnover of soil C, rather than to a decrease in plant productivity. The loss of net primary productivity is too small to explain the loss of soil C. The K 14C (0/00), which is the turnover time of soil organic matter (the reciprocal of the decomposition rate), shows that the surface horizons are dominated by fast-cycling organic matter (Fig. 16b). The K 14C (0/00) values also show the turnover of soil organic matter decreases with substrate time up to 150 kyr and then increases at the oldest site. During the ﬁrst 150 kyr of soil development, the volcanic parent material weathers to

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

427

Figure 15 Relationship between pyrophosphate-extractable C and pyrophosphate-extractable (Al þ Fe) in Japanese Andisols of different ages. From Wada and Higashi. (1976).

metastable noncrystalline minerals. Thereafter, the amount of noncrystalline minerals increases up to 150 kyr and then declines with greater age (Fig. 16c). In contrast, the amount of more stable crystalline minerals remains low until 150 kyr, then increases steeply (Fig. 16d). Soil organic carbon content follows a similar trend, accumulating to a maximum after 150 kgr, and then decreasing by 50% over the next four million years (Fig. 16a). The abundance of noncrystalline minerals accounts for . 40% of the variation in organic C content across all the mineral horizons, substrate age, and soil orders (excluding the O and A horizons dominated by fast-cycling plant litter). Noncrystalline minerals also strongly inﬂuence turnover of soil organic matter. Organic matter D 14C is highly and negatively correlated with abundance of noncrystalline minerals (R 2 ¼ 0.62) except in the oldest site, which has , 10% noncrystalline minerals. In contrast, there is no discernible correlation between the abundance of crystalline minerals and C content or turnover time across sites. The amount of C stabilized per gram is much greater for noncrystalline than for crystalline minerals. A positive relationship between noncrystalline minerals and organic C also exists in soils through the climate gradient. This indicates that the accumulation and subsequent loss of organic matter are largely driven by changes in the millennial scale cycling of mineral-stabilized C, rather than by changes in the amount of fast-cycling organic matter or in net primary productivity. Kennedy et al. (2002) showed that 85% of variation in total organic C could be explained by mineral surface area in a black shale deposit from two locations in

428

P. M. HUANG

Figure 16 Soil inventory of carbon in soil organic matter (SOM) (a), K14C of SOM (b), noncrystalline minerals (c), and crystalline minerals (d) versus age of soil substrate. Filled circles, total proﬁle; ﬁlled triangles, surface (O and A) horizons. From Torn et al. (1997).

the late Cretaceous Western Interior Seaway, United States. This relation suggests that adsorption of organic C compounds on mineral colloids plays a fundamental role in the preservation of organic C. Their data also provided evidence for organic matter within smectite interlayer, which is in accord with the ﬁnding of Wang and Huang (1986).

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

429

Therefore, soil mineral –organic interactions are important in determining the quantity of organic matter stored in soil, its turnover time, and atmosphere – ecosystem carbon ﬂuxes during long-term soil development. An understanding of how minerals inﬂuence the dynamics of soil organic matter should yield signiﬁcant improvements to our understanding of the role of soils in the global ion cycling.

B. DECOMPOSITION AND STABILIZATION OF C, N, P, AND S IN RELATION TO PRIMARY AND SECONDARY PARTICLES Generally, more than 95% of the N and S and between 20 and 90% of the P in surface soils are present in soil organic matter (Guggenberger and Haider, 2002). The close relationship between the organic forms of C, N, P, and S is well established. The turnover of organic C is closely associated with the dynamics of N, P, and S in soils. Soil mineral colloids exert a profound inﬂuence on the stabilization and degradation of soil organic matter and its associated nutrients. Chemical and physical interactions of minerals with soil organic matter result predominantly in the stabilization of the organic substances. However, minerals can also contribute to a more rapid cycling of associated nutrients of soil organic matter. Some soil minerals such as birnessite may enhance the cycling of the organically bound nutrients by catalytic degradation of the organic substances (Wang and Huang, 1987, 1992, 2000b; Huang, 2000a). The release of the stored organic nutrients depends on the mean residence time of soil organic matter associated with the mineral phase. The mean residence time of soil organic matter varies widely with the type of the organomineral associations and the spatial location within the aggregate structure of soil (Table V). Some of soil organic matter of rapid turnover is probably related to microbial residues, which are often loosely bound to layer silicates. In contrast, soil organic matter of slow turnover is tightly bound to Al and Fe oxides, hydroxides, and/or oxyhydroxides by formation of inner-sphere complexes. In soil environments, primary particles (layer silicate clay minerals and Al and Fe oxides, hydroxides, and/or oxyhydroxides) are usually clustered to larger entities — the secondary soil particles or aggregates. The arrangement of primary and secondary particles results in the formation of a series of pores, which range in size from nanometers to centimeters and is deﬁned as soil structure. Aggregation and soil structure determine the pore-size distribution and, hence, the accessibility of substrates to exoenzymes, microorganisms, and faunal grazers. Pools I and II are generally related to plant fragments, divided into easily available cell constituents and lignocelluloses at various degrees of degradation, respectively. As these fragments are more degraded and reduced in size,

430

P. M. HUANG

Table V Comparison of Estimated Mean Residence Times of Soil Organic Matter in Soil Physical Fractions Mean residence time (years)

Pool I

II III IV

V

Jenkinson and Rayner (1977) Decomposable plant material, 0.24 Resistant plant material, 3.33 Soil biomass, 2.44 Physically stabilized OM, 72 Chemically stabilized OM, 2857

Parton et al. (1987) Buyanovsky et al. (1994) Metabolic plant residues, 0.1– 1 Structural plant residues, 1– 5 Active SOM pool, 1–5 Slow SOM pool, 25–50

Carter (1996)

Vegetative fragments 2– 0.2 mm, 0.5–1

Litter, 1– 3

Vegetative fragments 0.05–0.025 mm, 1–3 OM in aggregates 2– 1 mm, 1–4 OM in aggregates 1– 0.1 mm, 2–10

Free POM (light fraction), 1–15 Microbial biomass, 0.1 –0.4 Intermicroaggregate OMa, 5–50

Passive SOM OM in ﬁne silt, ~400 pool, 1000– 1500 OM in ﬁne clay, ~1000

Intramicroaggregate OMb physically sequestered, 50– 1000 Chemically sequestered, 1000–3000

From Guggenberger and Haider (2002). Organic matter stored within macroaggregates but external to microaggregates; includes coarse occluded POM and microbial organic matter. b Organic matter stored within microaggregates; includes ﬁne occluded POM and microbial derived organic matter. a

the microbially available components are exhausted and the residues are more resistant to degradation. Pool III consists of soil biomass and readily available organic matter within large aggregates. These materials have a relatively short mean residence time. On average, the Pools I–III account for about 20 – 30% of the total C in soil organic matter. These pools must be renewed continuously by fresh plant residues to maintain a relatively constant nutrient level and release by mineralization. The balance between decay and renewal processes, which is sensitive to management, controls the availability of N, P, and S. Pool IV is physically protected soil organic matter and is also affected by cultivation, because physical disturbance such as ploughing destroys macroaggregates and large microaggregates. Pool V is chemically stabilized soil organic matter, which has the longest mean residence time. This pool represents 50 – 70% of the total C in soil organic matter and is seldom affected by management practices. Organic substances may weather soil minerals (Robert and Berthelin, 1986; Tan, 1986) and impede crystallization of secondary minerals

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

431

(Schwertmann, 1966; Kwong and Huang, 1975; Inoue and Huang, 1984; Huang and Violante, 1986; Schwertmann et al., 1986; Huang et al., 2002). Noncrystalline minerals have the ability to stabilize soil organic matter and, thus, reduce the turnover of C, N, P, and S (Torn et al., 1997; Guggenberger and Haider, 2002). Organic substances also act as binding agents to promote aggregation, which in turn, reduce the turnover of these nutrient elements through occlusion by minerals. Hence, in mutual interactions, noncrystalline minerals can retard the organic matter from being biodegraded, which in turn, inhibits the transformation of noncrystalline mineral phases into more stable crystalline minerals. Therefore, there are distinct interactive mechanisms between soil minerals and organic matter, which have direct effects on the cycling of C, N, P, and S in the environment.

X. EFFECTS OF INTERACTIONS BETWEEN MICROORGANISMS AND SOIL COLLOIDS ON THE TRANSFORMATION OF ORGANIC POLLUTANTS A. CATALYTIC TRANSFORMATIONS

OF

ORGANIC POLLUTANTS

Both enzymes and mineral colloids are involved in catalytic transformations of organic pollutants in soil environments. Enzymes are biotic catalysts and mineral colloids are abiotic catalysts. Frequently it is difﬁcult to determine whether an organic pollutant is transformed abiotically or biotically (Huang, 1990). In many cases, signiﬁcant abiotic and biotic catalytic reactions take place simultaneously (Huang and Bollag, 1999).

1.

Biotic Catalysis

Microorganisms are the major source of extracellular cellular enzymes in soils. However, considerable amounts of soil enzymes can also be derived from plant debris, plant root exudates, and soil fauna. The degradation of organic pollutants by extracellular enzymes is well documented (Dec et al., 2002). Malathion, for instance, is hydrolyzed to a monocarboxylic acid (via loss of an ethyl group) by a carboxyl esterase isolated from soil (Getzin and Roseﬁeld, 1971). The enzyme catalysis exhibits typical Michaelis-Menton kinetics. Enzymes commonly occurring in soil, such as esterases, amidases, phosphatases, and proteases catalyze the hydrolysis of the respective chemical bonds in xenobiotic molecules. (Dec et al., 2002).

432

P. M. HUANG

Extracellular phenoloxidases have the ability to catalyze the transformation of phenolic or anilinic compounds to their polymerized products (Bollag, 1992). For example, 2,4-dichlorophenol, an intermediate of 2,4-D (dichlorophenoxy acetic acid), is coupled to itself by a laccase of Rhizoctonia praticola to yield dimeric to pentameric products. In the presence of this laccase, 2,4-dichlorophenol reacts with various humic constituents (e.g., syringic acid, oricinol, or vanillin) to form a variety of hybrid oligomers. Various phenols and anilines are also enzymatically incorporated to natural HAs (Bollag et al., 1998). These oxidative coupling reactions were carried out only in vitro, there are indications they may also occur in soils. Phenoloxidases are believed to participate in humus formation by mediating the polymerization of the monomer products of microbial metabolism. They are also implicated in binding of xenobiotics to soil by catalysis of the oxidation of pollutants to free radicals, followed by chemical coupling of oxidation products to humic substances. Further, there are indications that participation of enzymes in humiﬁcation and binding may be triggered by the presence of toxic substrates that induce the production of extracellular phenoloxidases (Dec and Bollag, 2000).

2.

Abiotic Catalytic Transformation

Most of organic chemicals, including xenobiotics, exhibit a strong afﬁnity to humic substances. However, transformation of xenobiotics in terrestrial systems is greatly inﬂuenced by mineral components of soil (Huang, 1995). Mineral colloids, their high concentration in soil, large speciﬁc surface, and relatively high charge density, contribute to the overall xenobiotic transformation at least as much as does the organic matter. Organic matter can induce surface-catalyzed reactions of adsorbed pesticides, but theoretically it could also hinder the degradation of some pesticides by decreasing both their availability to microbial attack and their concentration in the soil solution (Huang and Bollag, 1999). The signiﬁcance of soil mineral-catalyzed abiotic transformation of xenobiotics in the environments has become widely recognized only recently (Cheng, 1990; Huang, 1990, 2000a). The processes of adsorption and abiotic degradation of xenobiotics through the action of the surfaces of soil minerals vary with the structural and surface properties of the minerals, saturating cations and hydration status, molecular structures of xenobiotics, and associated environmental factors. Besides clay minerals with well-ordered crystalline structures, soils contain mineral colloids such as metal oxides, aluminosilicates, and carbonates that exist in poorly ordered forms (Huang, 1995). These SRO mineral colloids

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

433

may exist as discrete phases in soils. They may also form coatings on crystalline minerals and organic matter and, thus, alter the surface chemistry of soil components. These SRO mineral colloids appear to play a dual role in humiﬁcation and binding interactions between xenobiotics and soil. First of all, they may be directly involved in the catalytic oxidation of organic chemicals with the formation of organic residues (Chorover et al., 1999; Karthikeyan et al., 1999). Secondary, they may inﬂuence the efﬁciency of enzymes that contribute to the same reactions (Claus and Filip, 1990; Gianfreda and Bollag, 1994; Huang et al., 1999). The contribution of these SRO mineral colloids to the transformation of organic chemicals merits increasing attention. (a) Oxidative transformation of organic compounds by soil minerals. Soil minerals play an important role in catalyzing the abiotic transformation of phenolic compounds (Huang, 1990, 2000a). Abiotic catalysts include primary minerals, layer silicates, metal oxides, hydroxides, oxyhydroxides, and poorly crystalline aluminosilicates. They promote the transformation of phenolic compounds through oxidative polymerization, ring cleavage, decarboxylation, and/or dealkylation. The ability of soil inorganic constituents to catalyze the transformation substantially varies with their structural conﬁguration and surface chemistry and the structure and functionality of the organics involved. Manganese(IV) oxides are most powerful soil minerals in catalyzing abiotic oxidation of phenolic compounds (Shindo and Huang, 1982, 1984; Huang, 2000a). Organic compounds such as 2,4-D and ethyl ether can be degraded by the catalysis of birnessite (d-MnO2) (Chenney et al., 1996). These organic compounds can be adsorbed on birnessite and rapidly oxidized, both producing CO2 as a major product (Fig. 17), but by somewhat different mechanisms. In the case of 2,4-D only, methanol-extractable Mn2þ is detected. Therefore, the solid degradation of organochlorine herbicides can occur by catalysis of birnessite, which is a common component in the environment. (b) Brønsted and Lewis acidity of mineral surfaces and hydrolysis of organic compounds. Mineral surfaces may detoxify adsorbed organic pollutants by catalysis through their ability to behave as Brønsted acids and donate protons or to act as Lewis acids and accept electron pairs. Brønsted acidity derives essentially from the dissociation of water molecules coordinated to surface-bound cations. Therefore, this acidity is strongly inﬂuenced by the hydration status and polarizing power of surface-bound and structural cations on mineral colloids (Mortland, 1986). Certain organic pollutants such as organophosphate and s-triazine pesticides can be degraded by catalysis of mineral colloids through their surface acidity (Brown and White, 1969; Mingelgrin et al., 1977). This type of reaction deserves attention in a wide range of other organic compounds in soil environments (McBride, 1994; Huang, 2000a). Besides the Brønsted acidity, the Lewis acidity of metals

434

P. M. HUANG

Figure 17 Gas chromatographic measurements of CO2 and oxygen in the headspace of reaction vials and 2,4-D at 1, 2, 3, and 4 h, respectively. The values indicate the mole percent of CO2 and oxygen in the headspace gas. †, CO2 from air control; B, CO2 from ethyl ether control; O, CO2 from 0.3 mmol 2,4-D plus ethyl ether sample; W, oxygen from air control; A, oxygen from the ethyl ether control; K, oxygen from 0.3 mmol 2,4-D plus ethyl ether. The error bars indicate the standard error for two measurements. From Cheney et al. (1996).

such as Al and Fe exposed at the edges of minerals is important in mineralcatalyzed hydrolysis reactions. Aluminum and Fe oxides in water and especially in the dry state have the ability to catalyze organic hydrolysis reactions, at least for those that are hydroxyl ion-catalyzed (Hoffmann, 1990). Dissolved metals and metal-containing surfaces play a signiﬁcant role in the catalysis of the hydrolysis of organic pollutants. The ability of a metal ion to catalyze the hydrolysis varies with its ability to complex with reactant molecules and shift electron density and conformation in ways favorable to the reaction (Hoffmann, 1980). Metals vary in their complex formation constants, which reﬂect differences in metal – ligand bond strengths and solvation forces (Stone and Torrents, 1995). The reactivity of hydrolyzable organic pollutants arises from the presence of electrophilic (electron deﬁcient) sites within the molecules. The SN2 mechanism (neucleophilic substitution — bimolecular) involves attack of electrophilic sites by OH2 or H2O, generation of a higher number intermediate, subsequent elimination of the leaving group, and the formation of a hydrolysis product (X and Y in this mechanism are elements or functional groups attached to C (Stone and Torrent, 1995):

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

435

In the case of the SN1 mechanism (nucleophilic substitution — monomolecular), the reaction proceeds with the loss of the leaving groups to generate a lower coordination number intermediate and then the generation of the hydrolysis product by neucleophilic addition (X, Y, and Z in this mechanism are elements and functional groups attached to P (Stone and Torrent, 1995):

Metal ions can catalyze hydrolysis in a way similar to acid catalysis. Metal ions and protons coordinate to organic pollutants so that the electron density is shifted away from the site of nucleophilic attack to facilitate the reaction. Protons have an extremely high charge density and great polarizing power; metal catalysis is, thus, insigniﬁcant in acidic conditions. By contrast, metal ions can readily coordinate two or more ligand donor sites on a molecule and can greatly outnumber protons in neutral and alkaline conditions. Many organic compounds are susceptible to catalysis by metals exposed on the surfaces of metal oxides/hydroxides and aluminosilicates. These include carboxylic acid ester, amides, anilides, phosphatecontaining esters, and other hydrolyzable compounds (Stone and Torrent, 1995; Huang, 2000a). Surface catalysis by metals on minerals is observed when all participating reactants are adsorbed to a signiﬁcant extent and when rate constants for the reactions at the mineral – water interface are comparable with or exceed rate constants for the reactions in homogeneous solution. (c) Catalytic effects of humic substances on the transformation of organic pollutants. Soil organic matter and especially humic substances exert catalytic or inhibiting effects on the abiotic hydrolysis of organic pollutants (Senesi and Miano, 1995). In aqueous systems, HAs and FAs have the ability to enhance the acid hydrolysis of phenoxy acetic acids and esters (Struif et al., 1975) and chloros-triazines (Li and Felbeck, 1972; Khan, 1978) and to retard the alkaline hydrolysis of the n-octyl ester of 2,4-D (2,4 DOE) (Perdue and Wolfe, 1982). The mechanisms proposed to explain the catalytic effects of HAs and FAs on the dechlorohydroxylation of the chloro-s-triazines, simazine, atrazine, and propazine consists of the interaction through H-bonding between the surface carboxyl groups of HAs and FAs and the heterocyclic nitrogen atoms of the triazine ring (Fig. 18). This interaction would strengthen the electron-withdrawing effect of the electrondeﬁcient carbon bearing the chlorine atom and, thus, reduce the activation energy barrier for the hydrolytic cleavage of the C –Cl bond and facilitate the replacement of chlorine by the weakly nucleophilic water (Fig. 18). This mechanism would also explain the correlation between the amount of chloro-s-triazines adsorbed by soil organic matter and their chemical hydrolysis by soil catalysis.

436

P. M. HUANG

Figure 18 The mechanism of the catalytic effects of humic substances on the dechlorohydroxylation of the chloro-s-triazines. From Senesi and Miano (1995).

B. BINDING OF ORGANIC POLLUTANTS, ENZYMES, AND MICROORGANISMS ON MINERAL AND HUMIC SURFACES AND THE EFFECTS ON POLLUTANT BIOAVAILABILITY 1.

Biodegradation of Bound Xenobiotics

The degradation of organic pollutants may be considerably reduced when they are retained by soil colloids. The major reason for reduced biodegradation rates is the diminished bioavailability of chemicals involved in binding processes (Alexander, 1995). The availability of sorbed xenobiotics to microorganisms varies with the chemical properties of the pollutant, the nature of the sorbent, the mechanism of sorption, and the properties of the degradative organisms (Guerin and Boyd, 1992). For instance, phenol (a nonionic compound at neutral pH) adsorbed to stream sediments is available to microorganisms, whereas cationic surfactants are unavailable (Shimp and Young, 1988). The herbicide diquat sorbed to the external exchange sites of nonexpansible clay mineral, kaolinite, is readily accessible to microorganisms; however, diquat molecules entrapped in the interlayer space of the expansible clay mineral, montmorillonite, escape biodegradation (Weber and Coble, 1968). The mechanism of sorption may change with residence time of xenobiotics present in soil, leading to changes in their bioavailability (Alexander, 1995). For example, a strain of pseudomonas mineralized as much as 45% of phenanthrene during a 5-day incubation with the compound that was in soil for less than 1 day (Hatzinger and Alexander, 1995). When phenanthrene was aged for 315 days, only 5% of the compound was subject to microbial mineralization. The initial fast

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

437

adsorption of phenanthrene molecules on soil colloids apparently did not constitute as big a barrier to biodegradation as the slow diffusion of the chemical across organic matter and its sequestration in inaccessible microsites within the soil matrix. Contaminants sorbed by soil particles may be available to certain microorganisms and unavailable to others. For instance, soil-sorbed naphthalene was degraded by pseudomonas putida strain 17484, but it did not undergo degradation by a gram-negative soil isolate, designated NP-Alk (Guerin and Boyd, 1992; Crocker et al., 1995). Apparently, by directly mineralizing surfacelocalized, labile sorbed naphthalene, the pseudomonas strain established concentration gradients that promoted desorptive diffusion and mineralization of nonlabile naphthalene portioned into soil organic matter. On the other hand, naphthalene degradation of NP-Alk appeared to rely mainly on the passive desorption of the surface-sorbed compound (Guerin and Boyd, 1992). Some investigations have provided evidence contradictory to the general principle that binding interactions should reduce the rates of the transformation of xenobiotics (Dec et al., 2002). The enhanced microbial degradation of xenobiotics in the presence of soil colloids is attributed to their increased concentration in the vicinity of the microbial cells attached to organic and/or mineral surfaces.

2.

Adsorption of Microorganisms and Biodegradation of Xenobiotics

The attachment of microorganisms to solid surfaces is a common phenomenon in soil environments. Microorganisms may use many different mechanisms to achieve and maintain the attachment (Dec et al., 2002). Adsorption is critical to the initial adhesion of microbial cells to solid surfaces. The next step may involve special attachment structures such as pili or holdfasts. In most cases, microorganisms manage to strengthen the attachment by means of extracellular polymeric adhesives such as polysaccharides and proteins, which are frequently their own metabolic products. Both electrostatic and hydrophobic interactions are involved in microbial attachment (Fletcher, 1996). Microorganisms are closely associated with mineral colloids in surface soils (Huang and Bollag, 1999). Sorption of microorganisms may be extremely extensive, especially in soils with high contents of clay and organic matter. Some sorbed microorganisms may be released back to the solution, but some may be sorbed irreversibly (Dec et al., 2002). The effect of adsorption of microorganisms on their ability to degrade organic compounds is difﬁcult to predict. The microbial degradation may be enhanced or decreased, or may remain unchanged as the cells undergo the adsorption to solid surfaces. This appears to depend on the nature and properties of solid surfaces and substrates and the kinds of

438

P. M. HUANG

microorganisms. However, the utilization of sorbed organic compounds by microorganisms that adhere to the same surfaces is a common phenomenon (Alexander, 1999). For instance, naphthalene is metabolized in soil mainly by sorbed degraders whose population is twice as great as that of the free degraders (Di Grazia et al., 1990). Knowledge on the effects of mineral colloids and other soil components on microbial activity under ﬁeld conditions is relatively limited (Haider, 1995). These effects may vary with the type of soil components and microorganisms. Microbial activity (growth, metabolism, and uptake of substrates) may be stimulated by clays to a greater extent in bacteria than in fungi (Stotzky, 1986). The impact of structural conﬁguration and surface properties of soil mineral and organic components on the activity of microorganisms and their ability to degrade organic pollutants with different structure and functionality deserves increasing attention for years to come.

3. Immobilization of Extracellular Enzymes and Xenobiotic Transformation Soil is a living system in which enzymes are present either free in solution or bound onto mineral colloids, humic substances, and mineral colloid-humic complexes (Section VI). This immobilization is the major factor that determines their performance in soil environments. Therefore, immobilization of enzymes may have a considerable impact on the rate of xenobiotic degradation. After immobilization, enzymes show generally increased stability toward physical, chemical, and biological denaturation (Ruggiero et al., 1996; Naidja et al., 2000). Immobilization may have a considerable effect on the activity and kinetic behavior of enzymes in xenobiotic transformation due to steric and diffusional restrictions, direct involvement of the active sites in binding to the support, and modiﬁed conformation of the immobilized enzymes (Matinek and Mozhaev, 1985; Quiquampoix, 1987; Naidja et al., 2000, 2002). Conformational alteration of enzyme molecules may result in the decrease in accessibility of the active sites to the substrate, causing a setback in substrate transformation (Goldstein, 1976). In general, the intimate association of enzymes with soil colloids should have a negative effect on xenobiotic degradation (Ruggiero et al., 1996). However, there are notable exceptions to this trend (Burns, 1986; Claus and Filip, 1990; Wada et al., 1992; Gianfreda and Bollag, 1994; Naidja et al., 2000). Transformation and degradation of anthropogenic organic compounds, including chlorinated phenols, detergents, pesticides, and other organic pollutants can be mediated by various enzymes (Munnecke et al., 1982; Bollag, 1992; Naidja et al., 2000). Enzymes have great potential for in situ treatment of environmental pollutants in soils and waters. The lifetime and stability of

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

439

enzymes can be increased by immobilization on soil colloids (Burns, 1986; Dec et al., 2002). Enzymes immobilized by mineral colloids can be used more efﬁciently, with a possible longer lifetime, higher stability, and possible reusability (Ruggiero et al., 1989; Naidja et al., 2000). Further, many mineral colloids are efﬁcient abiotic catalysts in the transformation of organic pollutants (Huang, 1990; Naidja et al., 2000; Dec et al., 2002). It has even been suggested that bioremediation efforts should include the possibility of site-speciﬁc abiotic reactions (Morra, 1996). However, information on the involvement of immobilized enzymes in the transformation of organic pollutants in soil is sparse (Dec et al., 2002). To cope with the problems caused by increasing amounts of agrochemical and industrial efﬂuents disposed of in soil and water environments, a fundamental understanding of enzyme-soil colloid interactions and their role in remediation of anthropogenic organic compounds in the environment is essential for the development of effective, durable, and practical treatment methods. A thorough understanding of the use of immobilized enzymes in remediation of contaminated soil environments is needed. However, the mechanisms of immobilization of enzymes on mineral and organic colloids and the resultant changes of protein conformation of enzymes remain obscure. A wide range of soil minerals have yet to be studied that may have a high capability to immobilize enzymes, adsorb substrates and reaction products and be catalytically active in the transformation and degradation of organic pollutants. Further, the role of abiotic catalysis of the mineral supports (Naidja and Huang, 1996), the hidden half in the enzymemineral colloid complexes, is still at an early stage and deserves an intensive investigation. Although coating of the clay surfaces with Al hydroxides may improve their ability to immobilize enzymes and in some cases enhance their activity (Naidja et al., 1997), the effect of the nature of coatings of Al and Fe hydro(xides) and oxyhydroxides on the activity of immobilized enzymes is still unknown. In bioremediation processes, incorporation of xenobiotics into humus through oxidative coupling reactions mediated by enzymes, mineral colloids, or mineral colloid-immobilized enzymes decreases the mobility and toxicity of phenolic pollutants (Naidja et al., 2000). Researchers should continue to develop new methods for maximizing the binding process of xenobiotics to humus, for instance, through the use of immobilized enzymes or abiotic catalysis or the addition of copolymerizing agents (Bollag et al., 1992). The effect of xenobiotics on the activity of mineral colloid-immobilized enzymes also warrants in-depth studies. To date, the knowledge on detailed surface structures of the enzyme –mineral colloid and enzyme – humus complexes at the molecular level still needs to be advanced. Atomic force and scanning tunneling microscopies and synchrotronbased spectroscopic methods such as extended X-ray absorption ﬁne structure (EXAFS) and X-ray absorption near edge ﬁne structure (XANES) can be used to

440

P. M. HUANG

better elucidate the mechanisms of the formation of enzyme– mineral colloid and enzyme – humus complexes and their catalysis in the transformation of anthropogenic organic compounds. More research is needed to uncover new enzymes and new mineral composite supports to be used for catalytic degradation of a wide range of industrial and agricultural pollutants in soil and water environments.

XI. IMPACT OF INTERACTIONS OF PHYSICOCHEMICAL, BIOCHEMICAL, AND BIOLOGICAL PROCESSES ON METAL TRANSFORMATION The transformation of metals in the environment is inﬂuenced by interactions of physicochemical, biochemical, and biological processes. The impacts of these interactive processes on metal transformation are especially important in the soil rhizosphere where the kinds and concentrations of substrates are different from those of the bulk soil because of root exudation (McLaughlin et al., 1998; Huang and Germida, 2002). This leads to colonization by different populations of bacteria, fungi, protozoa, and nematodes. Plant – microbe interactions result in intense biological processes in the rhizosphere. These interactions, in turn, affect physicochemical reactions in the rhizosphere. Physicochemical properties that can be different in the rhizosphere include acidity, concentration of complexing biomolecules, redox potential, ionic strength, moisture, and nutrient status. The total rhizosphere environment is governed by an interacting trinity of the soil, the plant, and the organisms associated with the root (Lynch, 1990a,b). Therefore, reactions and processes in the soil rhizosphere, which is the bottleneck of metal contamination of the terrestrial food chain, can only be interpreted satisfactorily with interdisciplinary approaches. Much of the research on physicochemical reactions of metal transformation in soil has used well-deﬁned model systems that simulate bulk soil characteristics. The impacts of interactions of physicochemical, biochemical, and biological processes in soils on metal transformation and bioavailability, food chain contamination, and ecosystem health merit increasing attention.

A. REDOX REACTIONS

OF

METALS

Redox reactions are important in controlling the chemical speciation of a number of contaminant metals, notably As, Se, Cr, Pu, Co, Pb, Ni, and Cu (Oscarson et al., 1981; Bartlett and James 1993; Alloway, 1995; Myneni et al., 1997; Huang, 2000a; Sparks, 2002). Redox reactions also are important in controlling the transformation and reactivity of Mn and Fe oxides in soils, which

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

441

have enormous capacities to adsorb metal pollutants and are major sinks of these pollutants (Huang and Germida, 2002). Further, reduction of sulfate to sulﬁde in anaerobic environments also affects the transformation, solubility, and availability of metal pollutants through the formation of highly insoluble metal sulﬁdes (Alloway, 1995). The transformation, mobility, bioavailability, and toxicity of metals, which exist in more than one oxidation state, are signiﬁcantly inﬂuenced by redox reactions. Massechelyn and Patrick (1994) have summarized the critical redox potentials for the transformation of some metal contaminants in soils. There has been little study of how changes in soil redox potential in the rhizosphere could affect the chemistry of metal contamination. However, the creation of an oxidized zone adjacent to the plant root in wetland soils has been identiﬁed as one process affecting the chemistry of Zn, Cu, and As in soils. In wetland soils, it has been well established that steep gradients in redox potentials develop around plant roots as a result of CO2 release from the roots. This process is reﬂected in precipitation of FeOOH (iron plaque) on the roots (Otte et al., 1989; Kirk and Bajita, 1995). Compared with the surrounding soil, these Fe-rich plaques on the roots of the salt-marsh plant, Aster tripolium, are enriched in Zn and Cu (Otte et al., 1989). Zinc also accumulates in the rhizosphere of rice (Oryza sativa L.), which is the result of the formation of a zone of oxidation of Fe2þ to Fe3þ adjacent to the roots (Kirk and Bajita, 1995). Zinc concentration in red roots (with iron plaque) is higher than in white roots; a positive effect of the Fe concentration on the root surface, up to a certain level, on Zn uptake into the xylem ﬂuid has been demonstrated (Otte et al., 1989). Above this level of Fe coating, Zn uptake by the plant is reduced, which is attributed to complete coating of the root surface by FeOOH and blocking of absorption sites. In reduced conditions, As is mobilized as a result of reduction of Fe and Mn oxides and reduction of As(V) to As(III). However, in the rhizosphere in wetlands, As is immobilized because of the oxidation to As(V) and adsorption to FeOOH (Otte et al., 1991). Therefore, As has been found to accumulate in the rhizosphere of many plants, but most of the As is likely to be retained on the root surface (Otte et al., 1991, 1995).

B. COMPLEXATION REACTIONS

OF

METALS

A series of complexation reactions in the soil solution affect metal transformation in the rhizosphere. Complexation reactions of metals with ligands in the soil solution are signiﬁcant in determining the chemical behavior, availability, and toxicity of metals in the rhizosphere. In view of the occurrence of organic ligands in the rhizosphere due to root exudates and microbial metabolites (McLaughlin et al., 1998; Marschner, 1998) and the stability constants of the complexes of metals with these ligands (NIST, 1997), a large

442

P. M. HUANG

fraction of the soluble metal ions in the soil solution may actually be complexed with a series of organic ligands commonly present in the rhizosphere. The study of metal speciation in the soil solution has been encouraged by the free metal ion hypothesis in environmental toxicology (Lund, 1990). This hypothesis states that the toxicity or bioavailability of a metal is related to the activity of the free aquo ion. Although this hypothesis is gaining popularity in studies of soil – plant relations (Parker et al., 1995), some evidence is now emerging that free metal ion hypothesis may not be valid in all situations (Tessier and Turner, 1995). At the same free metal activity, plant uptake of metals varies with the kinds of chelators present in solution. Further, given the same chelate, total metal concentration in solution affects plant uptake of metal. Either kinetic limitations to dissociation of the chelate or uptake of the intact chelate could explain these observations (Laurie et al., 1991). Seasonal changes of the concentrations of such metals as Cu, Mn, Zn, and Co in the rhizosphere are related to the presence of complexing agents of biological origin (Nielsen, 1976; Linehan et al., 1989). Krishnamurti et al. (1996) reported variation in pH and the cadmium availability index (CAI) of the bulk and rhizosphere soils after 2 weeks of crop growth. The pH of the rhizosphere soil was lower than that of the corresponding bulk soil, and the CAI values were higher, indicating that more Cd was complexed with the low-molecular-weight organic acids (LMWOAs) at the soil-root interface. The plant and proliﬁc microbial activity is expected to result in increased amounts of LMWOAs at the soil –root interface. Therefore, a larger fraction of the metal contaminant will be in a complexed and usually soluble form in the rhizosphere than in the bulk soil.

C. ADSORPTION – DESORPTION REACTIONS OF METALS The transformation of metals is signiﬁcantly inﬂuenced by adsorption – desorption reactions in soil environments. The reactions are affected by biochemical and biological processes, which should be especially important in the rhizosphere. Few studies have investigated the adsorption –desorption reactions in the rhizosphere. Most studies examined these reactions in bulk soils and under modifying conditions to simulate rhizosphere conditions. Root exudates and microbial metabolites include a series of LMWOAs, which are capable of forming complexes with metal ions (Robert and Berthelin, 1986; Stevenson, 1994). There is increasing evidence to show that LMWOAs may modify the transformation and dynamics of metals in soils, especially in the rhizosphere (McLaughlin et al., 1998). Krishnamurti et al. (1997) reported that the increase in Cd release in the presence of LMWOAs can be explained by the surface complexation of the particulate-bound Cd in soil with LMWOAs, which is reﬂected in the increase in the release of Cd from the soils with the increase in

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

443

the stability constant of Cd –LMWOA complexes (Fig. 19). Further, the rate coefﬁcients of the Cd release from the soils, calculated from the parabolic diffusion equation are substantially inﬂuenced by LMWOAs (Table VI). The rate coefﬁcients of Cd release within each ligand vary from soil to soil. The complexibility of soil Cd should vary with the nature of the particulate-bound Cd of the soil. Therefore, the rate of Cd release by each ligand should vary with the nature of the particulate-bound Cd of the soils. The activity of Cd species in the soil solution of the soil – root interface governs the amount of labile soil Cd. The importance of the metal –organic complex-bound particulate Cd species in determining the bioavailability of soil Cd has been shown by Krishnamurti et al. (1995a). The rate coefﬁcients of Cd release from the soils by LMWOAs (Table VI), which is a measure of the rate of the release of soil Cd to soil solution through complexation of soil Cd with LMWOAs, follow the same order as that of the CAI values of the soils (Krishnamurti et al., 1995b). Furthermore, the amounts of the Cd released from the soils by renewal of LMWOAs (Krishnamurti et al., 1997), which is an indication of Cd sustaining power of the soils, also follow the same trend

Figure 19 Relationship between Cd released from the soils by selected low molecular weight organic acids (LMWOAs) during the reaction period of 0.25 h and the logarithm of the stability constant of Cd-LMWOA complexes. From Krishnamurti et al. (1997).

444

P. M. HUANG

Table VI Overall Diffusion Coefﬁcients of Cd Release from the Soils by 1022 M Low Molecular Weight Organic Acids (LMWOAs) During the 0.25 to 1 h Reaction Period Overall diffusion coefﬁcient

Soil site

Acetic acid (mmol kg21 h20.5)

Citric acid (mmol kg21 h20.5)

Fumaric acid (mmol kg21 h20.5)

Oxalic acid (mmol kg21 h20.5)

Succinic acid (mmol kg21 h20.5)

Luseland Waitville Jedbergh

0.112 ^ 0.010 0.046 ^ 0.004 0.036 ^ 0.005

0.200 ^ 0.015 0.049 ^ 0.003 0.196 ^ 0.009

0.199 ^ 0.012 0.050 ^ 0.005 0.041 ^ 0.003

0.079 ^ 0.006 0.036 ^ 0.004 0.026 ^ 0.004

0.090 ^ 0.005 0.019 ^ 0.003 0.009 ^ 0.003

From Krishnamurti et al. (1997).

as the CAI values of the soils. The data indicate the importance of the kinetics of Cd release from soils by LMWOAs found in the rhizosphere in understanding Cd availability. More research is needed to understand the dynamics of the adsorption – desorption of metals as inﬂuenced by biochemical and biological processes. Such information is fundamental to understanding the pathways of the contamination of metals to the terrestrial food chain and to developing innovative management strategies to protect ecosystem health.

D. PRECIPITATION – DISSOLUTION REACTIONS OF METALS For many of the most abundant elements such as Al, Fe, and Mn, precipitation of mineral forms is common and may control their solubility. For most of the trace metals, direct precipitation from solution through homogeneous nucleation appears to be less likely than adsorption –desorption by virtue of the low concentrations of these metals in soil solutions in well aerated dryland soils (McBride, 1989; Tiller, 1996; Christensen and Huang, 1999). In reduced environments where the sulﬁde concentration is sufﬁciently high, precipitation of trace metals as sulﬁdes may have a signiﬁcant role in metal transformation (Robert and Berthelin, 1986). In aerobic soils, although precipitation of trace metals through homogeneous nucleation is not likely, heterogeneous nucleation may play a signiﬁcant role in metal transformation because of the presence of mineral, organic, and microbial surfaces that catalyze the nucleation set of crystallization (Huang and Germida, 2002). The energy barrier to nucleation is reduced or removed by surfaces. This is especially true in cases where there are crystallographic similarities between

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

445

the surface and the precipitating phase. This catalytic process reduces the extent of supersaturation necessary for precipitation to occur. However, precipitation reactions are often slower than adsorption – desorption reactions in soil environments. Besides physicochemical reactions, metals have easy access to bacterial surfaces through diffusion. Metal sorption and precipitation on bacterial surfaces are interfacial effect. Surface metal concentrations frequently exceed the stoichiometry expected per reactive chemical sites within the cell walls (Beveridge, 1989; McLean et al., 2002). The sorption of metals can be so great that precipitates can be formed and distinct metallic minerals are eventually formed through biomineralization as discussed in section XI E in this article. Activities of free metal ions in the rhizosphere may be decreased through the uptake by plants and microorganisms. Because concentrations of complexing organic ligands in the rhizosphere are higher than in bulk soils, metal contaminants are substantially complexed with organic ligands. And activities of free metal ions should, thus, be decreased further. Therefore, compared with bulk soils, the activities of trace metal ions in the soil solution of the rhizosphere in aerobic dryland soils appear to be even less controlled by precipitation through homogeneous nucleation. However, in the rhizosphere, precipitation of metals through heterogeneous nucleation on microbial surfaces on one hand and metal mobilization by biomolecules on the other hand as a result of intense biological activity warrant in-depth research.

E. METAL SORPTION AND UPTAKE BY MICROORGANISMS AND BIOMINERALIZATION All microorganisms contain biopolymers, such as proteins, nucleic acids, and polysaccharides, which provide sites where metal ions can bind (Hughes and Poole, 1989). These binding sites include negatively charged groups such as carboxylate, thiolate, phosphate, and groups such as amines, which coordinate to the metal center through lone pairs of electrons. Because of the ability of these biopolymers to bind metals, large concentrations of metals are frequently associated, not only with living microbial biomass, but also with dead cells (Berthelin et al., 1995). Many metals bind with various degrees of tenacity to the largely anionic outer surface layers of microbial cells. Binding of metals by microbial cells alters cell wall composition and induces morphological, ultrastructural, and surface charge changes (Venkateswerlu et al., 1989; Collins and Stotzky, 1996). Some metals are bound by cell walls to a greater extent than by clay minerals (Table VII), indicating that bacterial cell walls and membranes may act as foci for accumulation of metals in soils.

446

P. M. HUANG Table VII Metals Bound by Native B. subtilis Walls, E. coli Envelopes, Kaolinite, and Smectite Amount of metal bound (mmol g21) (oven-day weight)a

Metal Ag Cu Ni Cd Pb Zn Cr

Walls

Envelopes

Kaolinite

Smectite

423 ^ 15 530 ^ 13 654 ^ 25 683 ^ 19 543 ^ 11 973 ^ 13 435 ^ 37

176 ^ 3 172 ^ 9 190 ^ 3 221 ^ 6 254 ^ 5 529 ^ 32 102 ^ 2

0.46 ^ 0.02 5 ^ 0.03 4 ^ 0.2 6 ^ 0.2 3 ^ 0.2 37 ^ 1 8 ^ 0.5

43 ^ 0.3 197 ^ 4 173 ^ 10 1 ^ 0.02 118 ^ 6 65 ^ 2 39 ^ 5

From Walker et al. (1989). a The data represent the average of three to ﬁve determinations for each sample from duplicate experiments and the standard error.

Microbial biomineralization (the formation of minerals by microorganisms) is another important activity of microorganisms that is now being deﬁned (Beveridge, 1989). Its scope is much larger than initially thought, as it involves metal transformation and the development of ﬁne-grain minerals of tremendous range and kind (cf. Section VIIB). Research on biomineralization indicates that speciﬁc molecular interactions at inorganic– organic interfaces can result in the controlled nucleation and growth of inorganic crystals (Mann et al., 1993). A central tenet of biomineralization is that the nucleation, growth, morphology, and aggregation (assembly) of the inorganic crystals are regulated by organized assemblies of organic macromolecules, ‘the organic matrix’. Control over the crystallochemical properties of the biominerals is achieved by speciﬁc processes involving molecular recognition at inorganic-organic interfaces. Electrostatic binding or association, geometric matching (epitaxis), and stereochemical correspondence are important in these recognitions. The subtle differences in the kinetics of these recognition processes on different crystal faces lead to speciﬁc changes in crystal morphology. Biomineralization has global consequences in dynamics, toxicity, and fate of metal pollutants. Compared with any other life form, bacteria may have a greater capacity to sorb and precipitate metals from solution resulting in mineral formation as they have the highest surface area to volume ratio (Beveridge, 1989). Although in most environments, soluble metal ions are present in low concentrations, bacterial cells have a remarkable ability to concentrate metal ions from solutions. The research on the pH-dependent binding of metal ions to bacterial cell walls and the effect on metal bioavailability is currently emerging (Huang and Bollag, 1999).

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

447

F. METAL TRANSFORMATION BY MICROBIAL EXCRETIONS AND MYCORRHIZAL INFECTION Microorganisms have a range of metal transport systems, which are often highly speciﬁc for certain metals and capable of accumulating metals against large concentration gradients (Kurek, 2002; cf. Section VIIA). Certain microorganisms synthesize compounds that bind speciﬁc metals with high afﬁnity (Lynch, 1990a,b). For instance, some microorganisms make Fe-binding siderophores, which are biomolecules usually a phenolate or hydroxamate ligand (Neilands, 1981). Other microorganisms produce compounds called metallothioneins that are small cystine rich protein that strongly bind Cd, Cu, and Zn. Ligands of this type and related biomolecules are of signiﬁcance, because they inﬂuence the transformation, transport, bioavailability, and toxicity of metals in soils especially in the rhizosphere. Most plants in natural habitats form association with mycorrhizae. Arbuscular mycorrhizal fungi (AMF) are obligate symbionts, and infection of plant roots exerts a metabolic load on the host plant (Reid, 1990). During infection and colonization of host plant roots, mycorrhizal fungi produce mycelium inside root cortical cells. They may form storage structures termed vesicles, and they also may form other structures which are referred to arbuscules and serve as the site of ion exchange between the host plant and the mycorrhizal fungus. The fungi also form extra cellular hyphae that penetrate out of the root and explore the soil in search of nutrients including metals. Therefore, mycorrhizal infection may expand the volume of soil the root can penetrate. Uptake of metals may, thus, be facilitated by mycorrhizae because these elements are diffusion limited in soils. Therefore, the effect of mycorrhizal infection on metal uptake depends on the ability of the fungal symbiont to absorb metals and to transfer them to the symbiotic roots through extensive vegetative mycelium. Further, mycorrhizal fungi release LMWOAs (e.g., oxalic acid) into soils, which enhance the solubilization of particulate-bound metals and might facilitate uptake of metals. However, there are considerable gaps in knowledge on the mechanisms fungi inﬂuence the transformation and uptake of metals.

XII. FORESEEABLE IMPACTS OF SOIL MINERAL– ORGANIC MATTER– MICROORGANISM INTERACTIONS ON THE ECOSYSTEM Soil is a focal point of the ecosystem (Odum, 1989; Coleman, et al., 1998). Physical, chemical, and biological processes are interacting in soil environments, and are governed by soil mineral – organic matter – microorganism interactions.

448

P. M. HUANG

Therefore, interactions of these three major solid components of soils have enormous impacts on reactions and processes critical to ecosystem health (Huang, 2002).

A. GLOBAL ION CYCLING Major biogeochemical transformations of elements include the cycling of C, N, P, S, and metals (Butcher et al., 1992), which should be very much inﬂuenced by interactions of soil minerals with organic matter and microorganisms. Transformation of soil organic matter are closely related to the mobilization of C, N, P, and S into the soil solution or their immobilization from the solution and to the release or ﬁxation of trace gases such as CH4, CO2, OCS (carboxide sulﬁde), H2 NO2, and NO. These microbially mediated processes are the basis for the interrelationship of the C, N, P, and S cycling within the soil-plant system (McGill and Cole, 1981). Interactions of soil minerals with microorganisms and organic components have an important role in inﬂuencing the stability and degradation of soil organic matter and its associated nutrients (Guggenburger and Haider, 2002) directly affecting the global cycling of C, N, P, and S. Metals are part of natural biogeochemical cycles. One of the characteristics of the cycle of metal mobilization and deposition is that the form of the metal is changed. This change in speciation of a metal has a profound effect on its fate and impact on ecosystem health (Benjamin and Honeyman, 1992; Hayes and Traina, 1998). Metals are found in the environment in solid, solution, and gaseous phases, associated with thousands of different compounds. These associations often reﬂect the afﬁnity of metal ions for other atoms with free electron pairs, particularly O, N, and S. The critical processes controlling global metal cycling are adsorption – desorption, precipitation –dissolution, complexation, and volatilization (Benjamin and Honeyman, 1992; Hayes and Traina, 1998; Kurek, 2002; McLean et al., 2002). Transport in solution or aqueous suspension is a major mechanism for metal movement in the ecosystem. This transport process is greatly inﬂuenced by adsorption – desorption on surfaces of minerals and organic matter, precipitation –dissolution especially in reduced environments where sulﬁde concentration is sufﬁciently high, and a series of inorganic and organic complexation reactions in dissolved and particulate phases. Through their effect on the chemical environment in soils and sediments, microorganisms can help dissolve, complex, or precipitate metals and can also directly mediate reactions involving metals, such as initiation of ﬁne-grain mineral formation. Transport of particles suspended in the air is an important process for transporting many metals to regions far from their sources. A few metals, most

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

449

notably Hg, can exist as gases at ambient temperatures. In the case of Hg, reduction of Hg2þ to Hg8 and alkylation to form methyl- or dimethylmercury can result in the loss of Hg from the aqueous phase. Microorganisms can also convert the methylated forms to Hg8, which is more volatile and less toxic. Several other metals such as As and Se also form organometallic compounds, which can be mediated by microorganisms (Gadd, 1993). These volatile organometallic compounds can dominate transport of the metal in local environments. However, mediation of alkylation of metals such as Hg by bacteria is substantially inﬂuenced by Hg speciation on surfaces of mineral colloids (Table VIII). Soil minerals, organic matter, and microorganisms have their respective roles in inﬂuencing metal speciation, toxicity, and cycling.

B. GLOBAL CLIMATE CHANGE Many biogeochemical and physical processes are involved in determining the climate of the Earth (Charlson et al., 1992). The biogeochemical cycles of C, N, and S are central to the radiative properties of the atmosphere. Carbon and N form radiatively important gases. Sulfur is a crucial component of clouds and most aerosols. The cycles of C and N contribute to the long-wave radiative properties,

Table VIII Biomethylation of Hg(II) Adsorbed on Mineral Colloids Common in Freshwater Sediments, by P. ﬂuorescens Isolate BPL85-48 during a 25 h Incubation Period Optical density Sample ID Blankd Controle KGa-1 STx-1 MnO2

Hg(II)sourcea

Absorbance at 530 nm

RGIb

CH3Hgþ (ng L21)c

– Hg(NO3)2 Kaolinite Montmorillonite Birnessite

0.551 0.430 0.423 0.451 0.478

1.28 a 1.00 d 0.98 d 1.05 c 1.11 b

– 32.86 ^ 0.67 a 30.53 ^ 1.64 ab 25.96 ^ 4.17 b ND f c

From Farrell et al. (1998). a Total concentration of Hg(II), added as Hg(NO3)2 or in adsorbed form, was 6 mmol/100 ml. b Relative growth index ¼ optical density of colloid-amended medium/optical density of the control. Values followed by the same letter are not signiﬁcantly different (P # 0.05; least signiﬁcant difference test, LSD ¼ 0.04). c Values followed by the same letter are not signiﬁcantly different (P # 0.05; least signiﬁcant difference test, LSD ¼ 5.30 ng CH3Hgþ L21). d Isolate grown in the M-IIY medium in the absence of Hg(II). e Isolate grown in M-IIY medium supplemented with Hg(NO3)2; total Hg(II) concentration ¼ 60 mM. f ND, not detectable.

450

P. M. HUANG

whereas the S cycle inﬂuences the short-wave radiative properties; all of the cycles are severely perturbed by human activity (Table IX). Transformation of C, N, and S in soils as inﬂuenced by land management and the impact on their ion cycling and global climate change should not be overlooked (Lal, 1998; Guggenburger and Haider, 2002). Jenkinson et al. (1991) estimated the additional degradative effects on soil organic matter if the global annual mean temperature rises during the next 60 years by 38C. According to their estimate, about 100 Gt C(1 Gt ¼ 109 t) should be additionally evolved as CO2 from soil organic matter (1600 Gt C). This will increase the present CO2 concentration in the atmosphere by 14%, whereas the combustion of fossil fuels (5.4 Gt C yr21) should add during this 60 year period 330 Gt C to the atmosphere. Microbial by-products and resistant plant residues adsorbed on soil particles have turnover times in terms of years. Fulvic acids have turnover times in terms of hundreds of years, whereas HAs and humins usually approach a few thousand years in their turnover time (Paul and Van Vee, 1978). Although the HAs and humins constitute by far the majority of the organic C in a system, they contribute only a small proportion to the annual cycling of C within the soil because of their very slow turnover rate. The undecomposed litter also includes the soil biomass and microbial metabolites. These, together with the plant residues, constitute the active fraction of organic matter that has a prominent role in the cycling of elements such as C, N, and S annually. The inﬂuence of crystalline and noncrystalline mineral colloids, which differ in structural conﬁguration and surface properties, on the biodegradation, turnover, and stabilization of organic components, the cycling of C, N, and S, and the impact on global climate changes merits close attention (Huang and Schnitzer, 1986; Berthelin et al., 1999; Violante et al., 2002a, b).

Table IX Radiatively Important Trace Species in the Atmosphere: Percent Change in Flux Measured Relative to the Pre-industrial Age Cycle change Long-wave absorbers Water Carbon Nitrogen Halogens Short-wave reﬂectors Sulfur From Charlson et al. (1992).

Species

% Change

H2O (vapor) CO2 CH4 N2 O Chloroﬂuorocarbons

Not known þ 50% . þ 65% þ 25% þ1

SO22 4

þ 230%

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

451

C. BIODIVERSITY The functioning and stability of terrestrial ecosystems are determined by plant biodiversity and species composition (Tilman et al., 1996; Hooper and Vitousek, 1997). However, the ecological mechanisms by which they are regulated and maintained are not well understood. These mechanisms need to be identiﬁed to ensure successful management for conservation and restoration of diverse natural ecosystems. Van der Heijden et al. (1998) recently reported that below-ground diversity of AMF is a major factor contributing to the maintenance of plant biodiversity and to ecosystem functioning. These results emphasize the need to protect AMF and to consider these fungi in future management practices in order to maintain diverse ecosystems. Research provides support for the view that ﬂouristically rich systems are more productive (Tilman et al., 1996), show greater stability under stress (Tilman and Downing, 1994), and are more likely to provide alleviation of global problems posed by atmospheric CO2 enrichment (Naeem et al., 1994). A recognition of these properties, coupled with an increasing awareness that the diversity of terrestrial vegetation systems is everywhere under stress, has encouraged the research on mechanisms that determine and affect species composition in plant communities. The scientiﬁc evidence indicates that below-ground microbial diversity substantially inﬂuences plant biodiversity and ecosystem variability. Further, microbial events are signiﬁcantly affected by surface-reactive particles (Stotzky, 1986, 2002; Burns, 2002). However, our knowledge on the effect of mineral – organic component – microorganism interactions on below-ground microbial diversity and the impact on above-ground biodiversity remains to be advanced.

D. BIOLOGICAL PRODUCTIVITY

AND

HUMAN NUTRITION

Soil is the life-sustaining material, which is the structurally porous and biologically active medium that has developed on the continental land surface on our planet. This material is created and continues to evolve through weathering processes driven by biological, climatic, geological, topographic, and chronological inﬂuences. Soil is the cradle of agriculture including crop and animal production and a fountain for sustaining human nutrition. However, it is the fragile epidermis of the planet Earth that can sustain human nutrition or cause starvation for humans depending on our management of soil resources. Interactions of soil minerals with organic components and microorganisms exert enormous inﬂuences on the transformation and dynamics of soil organic matter (cf. Sections II, IV, IX), nutrient cycling and bioavailability (cf. Section IX; Barber, 1995), efﬁcacy and toxicity of pesticides (cf. Section X; Cheng, 1990), microbial metabolic processes, growth, adhesion, and ecology (cf. Section V),

452

P. M. HUANG

enzymatic activity (cf. Section VI), and soil physical properties (cf. Section VIII). Therefore, interactions of soil minerals with organic components and microorganisms should have great impacts on plant nutrition and biological productivity of soils. In the rhizosphere, the narrow zone of soil surrounding a living plant that is subject to inﬂuence by the root and its exudates, more intense microbial activity and larger microbial populations occur than in the bulk soil (Waisel et al., 1996). Up to 18% of the C assimilated through photosynthesis can be released from roots. Since the rhizosphere is rich in root exudates, microbial population can be 10 – 100 times larger than the population in bulk soil (Sposito and Reginato, 1992). The rhizosphere typically extends away from the root for up to 2 mm, but some organisms (e.g., fungi) may be stimulated up to 5 mm away. The rhizosphere is bathed in root exudates and microbial metabolites. Both the amounts and proportion of organic compounds of root exudates vary substantially with plant species and cultivars. Further, the same plant cultivar grown in different soils varies in the kind and amount of LMWOAs present in the rhizosphere (Szmigielska et al., 1997). The chemistry and biology at the soil – root interface, thus, differs signiﬁcantly from soil to soil. The soil rhizosphere is the bottleneck of the nutrient-feeding zone in soils. Therefore, the dynamics, transformation, and bioavailability of nutrients are bound to be inﬂuenced greatly by the chemistry and biology at the soil-root interface. The intense soil mineral – organic component-microorganism interactions in the rhizosphere, thus, deserve close attention in the development of innovative management strategies for land resources to increase biological productivity and sustain human nutrition.

E. GEOMEDICINE Geomedicine may be deﬁned as the science dealing with the environmental factors, which inﬂuence the geographical distribution of pathological and nutritional problems relating to human and animal health (La˚g, 1980). Geomedicine is a young science with very old roots. Knowledge on soil science is needed for solving many geomedical problems (La˚g, 1994). Hunger and malnutrition are serious issues for large groups of populations, especially in developing countries. In addition to prevention of starvation, promotion of better nourishment quality of food and feed is important. Pollution of the environment and related health problems have increased rapidly in many industrialized countries. Effects of many soil chemical, physical and biological factors and processes on geomedical problems should be studied in depth as they impact on the quality of vegetation and the food and feed produced. Biogeochemistry of trace elements is greatly inﬂuenced by soil – plant – microbe interactions (Chang et al., 2002). Therefore, soil mineral – organic

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

453

component – microorganism interactions deserve close attention in the transformation, dynamics, and bioavailability of trace elements, many of which are of concern to animal nutrition and health and well-being of humankind (National Academy of Sciences, 1974, 1977). They include Se, Fe, I, Zn, Cu, Mn, Mo, Cr, F, Co, Si, V, Ni, As, and Mg. One trace element may serve in one, several or dozens of different metalloenzymes or tissue constituents. All the trace elements appear to have vital metabolic roles in critical steps in health risks built into human genetics. To facilitate fundamental understanding of the linkage of trace elements in the soil-plant-environment-animal-human systems, and to provide practical solutions to their deﬁciency and toxicity problems, we need to advance the knowledge on soil mineral-organic component-microorganism interactions affecting the transformation, dynamics, and bioavailability of trace elements in soils and associated environments.

F. ECOTOXICOLOGY

AND

HUMAN HEALTH

Soil plays the central role as the organizer of terrestrial ecosystem (Coleman et al., 1998). Furthermore, it may be perceived as the center of the ecosystem, which evolves because of interactions of the lithosphere, hydrosphere, atmosphere, and biosphere. A factor of central importance of soil to ecological studies is that soils on a global scale have a range of characteristics, which enable an enormous array of microorganisms, plants, animals, and humans to co-exist and thrive. Humans have exploited the ability of soils to provide massive amounts of food. About . 40% of the net primary production of the world is exploited by humans (Vitrousek et al., 1986). The exploitation is increasing with the addition of 87.6 million people to the global population every year, with the rate addition steadily increasing (Brown and Flavin, 1996). The impact of population and the accompanying intensiﬁcation of agriculture and industrialization on ecotoxicology and human health is, thus, of increasing concern. Among the environmental compartments, about 90% of environmental pollutants are bound with soil particles and 9% of the pollutants are bound with aquatic sediments (Table X). These soil- and sediment-bound pollutants are in dynamic equilibrium with the hydrosphere, atmosphere, and biosphere. Soil mineral – organic component – microorganism interactions may enhance the release of environmental pollutants from soils and sediments and, thus, pose a threat to ecosystem health, including human health. Land resource management to sustain the capacity of a soil to function within ecosystem boundaries to improve biological productivity, maintain environmental quality, and promote plant and animal growth and human health is the central role of soil scientists in contributing to human welfare. Interactions of soil minerals with organic components and microorganisms have vital roles in governing transformation, speciation, transport, bioavailability, and toxicity of

454

P. M. HUANG Table X Theoretical Pollutant Distribution in the Environment at Equilibrium

Compartment Air Water Sediment Soil Aquatic biota

Concentrationa (mol m23) 4 £ 10210 1027 1022 1022 1027

% Distribution in compartmenta 0.35 0.01 9.10 90.50 0.01

From Crosby (1982). a Assumes approximately 100 kg of pollutant (MW 100) introduced into 10 km3 of the environment.

organics, metals, and other inorganics of agricultural and ecological concerns (Fig. 20). These interactions are fundamental to understanding and regulating the ecosystem at the molecular level.

G. BIOTECHNOLOGY The probability and frequency of the transfer of genetic information, both intra-and interspeciﬁcally, between microorganisms, especially bacteria, in soil and other natural habitats have become important concerns (Stotzky, 1986, 2002). Although genetic transfer has been extensively studied in pure culture systems, few studies have been conducted in natural microhabitats. More studies are necessary to clarify the role of mineral colloids, organic matter, and organomineral complexes in not only the survival of genetically engineered microorganisms but also the transfer of genetic information by conjugation in soil. Further, because of the potential hazards associated with the transfer of genes — whether introduced accidentally or deliberately — in soil and contiguous habitats (e.g., groundwater, surface water, atmosphere, plants, animals, human beings), the importance of mineral – organic matter –microbe interactions in inﬂuencing the fate of genetically engineered microbes in soil and related environments should receive increasing attention. Surface-active particles, primarily mineral colloids and humic substances, in soil and other natural habitats are important in the persistence of biomolecules which in the absence of such particles are rapidly degraded or inactivated by the indigenous microbiota. Many biomolecules are important in the ecology, biodiversity, evolution of microorganisms, environmental protection, and biological control of pests (Stotzky, 2002). Insecticidal proteins produced by Bacillus thuringiensis and by transgenic plants containing toxin genes, enzymes, bacterial transforming DNA, and viruses are examples of the persistence to biodegradation and the retention of biological activity of biomolecules when

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

455

Figure 20 Impact of environmental soil chemistry on agricultural sustainability and the ecosystem health. Environmental soil chemistry is governed by mineral – organic matter – microorganism interactions. From Huang (2000b).

bound on such particles. Therefore, these surface-active particles affect the transfer of genetic information among bacteria by conjugation, transformation, and transduction in soil. All these aspects affect ecosystem health. The gains in food production provided by the Green Revolution have reached their ceiling while world population continues to rise. The application of

456

P. M. HUANG

advances in plant biotechnology is going to be essential especially in developing countries if farmers’ yields and yield ceilings are to be raised, excessive pesticide use reduced, the nutrient value of basic foods increased, and farmers on less favored lands provided with varieties better able to tolerate drought, salinity and lack of soil nutrients (Conway and Toenniessen, 1999). However, when biotechnology is applied at the ﬁeld scale, the impact of soil mineral – organic matter –microorganism interactions on the effectiveness of growing genetically modiﬁed crops in different soil environments so as to achieve the objectives of being more productive, yet less damaging to the ecosystem and consequently more beneﬁcial to human health remains to be uncovered. We need to assure that biotechnology is applied to agriculture only when this can be done safely and effectively in helping to achieve future food security and protect ecosystem health for the world.

H. RISK ASSESSMENT One of the great ecological challenges today is to determine if soil contamination will lead to adverse effects on beneﬁcial organisms. The mere presence of a chemical in air, water, and soil, biota, or food can raise public fears and a call for regulation of that chemical or its removal from the environment in which it was found. Environmental monitoring may be able to show that the chemical is a contaminant because concentrations are above some established background. Contaminants have background levels in soils and biological tissues as a result of their persistence and global dispersion. What is not known is if the measured concentration of a contaminant can have an adverse effect on an organism. To answer this question, we must use the tool of risk assessment, which combines exposure assessment (how is the organism exposed to the chemical and how much exposure is there?) and effect assessment (how does the chemical harm the organism? how much exposure to the chemical is required to cause harm?). Assessing exposure to contaminants in soil environments includes the determination of the pathways to exposure (inhalation, drinking water, food or direct soil ingestion) and the extent of contaminant transfer from soil to the various pathway components (Logan, 1998). Possible soil pathways of human contaminant exposure are presented in Table XI. In environmental exposure assessment, a transfer coefﬁcient (K) must be determined for each component of the pathway, namely, the fraction of the component that is transferred from one component to another (e.g., from soil to plant, soil to air, soil to surface water). Such an approach is used in Mackay’s general fugacity model (Mackay and Paterson, 1991). The transfer coefﬁcient can be viewed as an availability index, although other terms (extractability, solubility, accessibility, volatility) can also be used. This will depend on the mechanism of transfer, which should be greatly inﬂuenced by interactions of soil minerals with organic components and

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

457

Table XI Pathways for Human Exposure to Soil-Borne Contaminants. The Most Exposed Individual (MEI) of the Population is also Identiﬁed. Based on U.S. Environmental Protection Agency Exposure Assessment for Land Applied Contaminants in Sewage Sludge (U.S. Environmental Protection Agency, 1995) Pathway

MEI

Soil ! plant ! human Soil ! plant ! human Soil ! human Soil ! plant ! animal ! human Soil ! animal ! human Soil ! dust ! human Soil ! surface water ! human Soil ! ground water ! human Soil ! air ! human

Human lifetime plant ingestion; general population Human lifetime plant ingestion; home gardener Child Human lifetime ingestion of animal products; animals raised on forage Human lifetime ingestion of animal products; animals ingest soil Human lifetime dust inhalation Human lifetime ingestion of surface water and ﬁsh Human lifetime ingestion of groundwater Human lifetime inhalation of volatilized contaminants

From Logan (1998).

microorganisms. These interactions should profoundly inﬂuence the transformation, dynamics, bioavailability, and toxicity of contaminants in soil and related environments (Huang, 2002). Because of these interactions, some pollutants that would be otherwise at an acceptable level may be released and enter the pathway affecting health. Therefore, soil mineral –organic component –microorganism interactions merit attention in soil pathways of human contaminant exposure and risk assessment.

I. RISK MANAGEMENT, REMEDIATION, AND RESTORATION Land use, past and present, critically inﬂuences the extent and intensity of soil contamination. Contamination of soils from anthropogenic chemicals and their subsequent degradation has become a major concern because of the critical role of soil resources in sustaining agricultural production, economic development, and ecosystem health. Both inorganic and organic anthropogenic compounds in soils may not only adversely affect their production potential but also deteriorate the quality of the food chain, the air, the surface water, and the underlying groundwater. Therefore, it is essential to manage the risk through use of remediation technology to restore ecosystem health Selected ameliorants have been used to remediate metal contaminated soils (Table XII). These ameliorants are rather inexpensive and readily available on a

458

P. M. HUANG Table XII Selected Ameliorants that are Adapted to Metal Contaminated Soils

Technique

Target contaminants

Soil processes involved

Limestone

Metals, radionuclides

Precipitation, sorption

Zeolite Apatite

Metals, radionuclides Metals

Clay mineral

Metals, radionuclides

Ion exchange, sorption, ﬁxation Sorption, precipitation, complexation Ion exchange, sorption, ﬁxation

Constraints Ineffective for oxyanions; certain crops (lettuce, spinach, tuber, and others); short term Insufﬁcient data; short term Selective; insufﬁcient data

Type of clay; short term

From Adriano et al. (1998).

global basis. However, they may provide only interim solutions in stabilizing contaminants. Soil processes involved in remediation by these ameliorants include precipitation, sorption, ion exchange, ﬁxation, and complexation. These processes may be substantially inﬂuenced by soil mineral – organic component – microorganism interactions especially in the rhizosphere soils where these interactions are intense. Plants have been described in engineering terms as solar driven pumps and ﬁltering systems that extract and concentrate elements from the environment (Cunningham and Berti, 1993; Salts et al., 1995). Some of the metals that may accumulate in plants are those that have nutritional value. Certain plants also have the ability to accumulate metals that have no known biological function. The value of metal-accumulating plants has recently been realized (Cunningham and Berti, 1993; Pierzynski et al., 2000). Further, plant-assisted bioremediation are used to remediate soil contaminated with organic compounds (Walton and Anderson, 1992; Anderson et al., 1993). In phytoremediation, a wide range of processes are involved in the soil rhizosphere where rhizospheric microorganisms interact intensely with soil minerals and root exudates. These interactions should have a profound inﬂuence on transformations of metals and anthropogenic organic compounds and the efﬁcacy of phytoremediation. Xenobiotics can also be degraded by oxidative coupling reactions that are catalyzed either biologically by polyphenol oxidases and peroxides (Bollag, 1992) or abiotically by metal oxides and clay minerals (Huang, 2000a). These processes may result in the detoxiﬁcation of xenobiotics and have been suggested as a means of decontamination. Abiotic and biotic catalysts co-exist in soil environments. Abiotic catalysts can inﬂuence microbial formation of enzymes

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

459

and enzymatic activity (cf. Section VI). Many soil abiotic catalysts can also immobilize enzymes and alter their performance. Further, many soil abiotic catalysts inﬂuence the activity of the desorbed enzymes. Therefore, the inﬂuence of interactions of abiotic and biotic catalysts on the transformation and toxicity of xenobiotics and the impact on risk management, remediation, and restoration of ecosystem health is an issue that needs to be addressed in the years to come.

XIII.

SUMMARY AND CONCLUSIONS

Soils are among the major compartments of the ecosystem. Their colloidal or particulate constituents, be minerals, organic matter, or microorganisms, play key roles in affecting physical, chemical, and biological processes in the pedosphere. Minerals, organic matter, and microorganisms are constantly in close association and in interactions with each other in soil environments. Interactions of these three solid components mediated by soil solution and atmosphere govern mineral weathering transformations, formation of organomineral complexes, microbial and enzymatic activities, soil structural stability, dynamics of aggregate turnover, biogeochemical cycling of C, N, P, and S, and transformation and dynamics of metals and organic pollutants in the terrestrial system. Therefore, foreseeable impacts of interactions of minerals with organic matter and microorganisms in soil environments include: (1) global ion cycling and climate change, (2) biodiversity, (3) biological productivity and human nutrition, (4) geomedicine, (5) industrialization, waste disposal, ecotoxicology, and human health with respect to the bioavailability and toxicity of metals, metalloids, other inorganics, xenobiotics, biomolecules, and pathogens, (6) biotechnology development in relation to agricultural sustainability and ecosystem integrity, (7) ecosystem risk assessment, and (8) ecosystem risk management and restoration. Fundamental understanding of soil mineral– organic matter –microorganism interactions at the molecular level is essential to understanding and regulating the behavior of the terrestrial ecosystem on the global scale. Future research on this extremely important and exciting area of science should be stimulated to uncover the dynamics and mechanisms of environmental processes in nature and the impact on human welfare.

ACKNOWLEDGMENT This study was supported by discovery grant GP 2383 — Huang of the Natural Sciences and Engineering Research Council of Canada.

460

P. M. HUANG

REFERENCES Adriano, D. C., Chlopecka, A., and Kaplan, D. I. (1998). Role of soil chemistry in soil remediation and ecosystem conservation. In “Soil Chemistry and Ecosystem Health” (P. M. Huang, D. C. Adriano, T. J. Logan, and R. T. Checkai, Eds.), pp. 361 –386. Soil Sci. Soc. Am., Madison, WI, SSSA Spec. Publ. 52. Ahmann, D., Robert, A. L., Krumholz, L. R., and Morel, F. M. M. (1994). Microbe grows by reducing arsenic. Nature (London) 371, 750. Alexander, M. (1995). How toxic are toxic chemicals in soil. Environ. Sci. Technol. 29, 2713– 2717. Alexander, M. (1999). “Biodegradation and bioremediation”, 2nd ed. Academic Press, San Diego, CA. Alloway, B. J. (1995). Soil processes and the behavior of metals. In “Heavy Metals in Soils” (B. J. Alloway, Ed.), pp. 11–37. Blackie Academic & Professional, London. Anderson, T. A., Guthrie, E. A., and Walton, B. T. (1993). Bioremediation. Environ. Sci. Technol. 27, 2630–2636. Baldock, J. A. (2002). Interactions of organic materials and microorganisms with minerals in the stabilization of structure. In “Interactions Between Soil Particles and Microorganisms. Impact on the Terrestrial Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 85–131. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK. Baldock, J. A., and Skjemstad, J. O. (2000). The role of soil mineral matrix in protecting natural organic materials against biological attack. Org. Geochem. 31, 697– 710. Barber, S. A. (1995). “Soil Nutrient Bioavailability: A Mechanistic Approach”, 2nd ed. Wiley, New York. Barnhisel, R. I., and Bertsch, P. M. (1989). Chlorites and hydroxy-interlayered vermiculite and smectite. In “Minerals in Soil Environments” (J. B. Dixon and S. B. Weed, Eds.), pp. 729– 788. Soil Sci. Soc. Am., Madison, WI. Bartlett, R. J., and James, B. R. (1993). Redox chemistry of soils. Adv. Agron. 50, 151– 208. Benjamin, M. M., and Honeyman, B. D. (1992). Trace metals. In “Global Biogeochemical Cycles” (S. S. Butcher, R. J. Charlson, G. H. Orians, and G. V. Wolfe, Eds.), pp. 317 –352. Academic Press, London. Berthelin, J., Huang, P. M., Bollag, J.-M., and Andreu, F. (1999). “Effect of Mineral–Organic– Microorganism Interactions on Soil and Freshwater Environments”. Plenum, New York. Berthelin, J., Munier-Lamy, C., and Leyval, C. (1995). Effect of microorganisms on mobility of heavy metals in soils. In “Environmental Impact of Soil Component Interactions. Volume II Metals, Other Inorganics and Microbial Activities” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L. Page, Eds.), pp. 3– 17. CRC Press/Lewis Publishers, Boca Raton, FL. Bertsch, P. M., and Parker, D. R. (1996). Aqueous polynuclear aluminum species. In “The Environmental Chemistry of Aluminum” (G. Sposito, Ed.), pp. 117–168. CRC Press/Lewis Publishers, Boca Raton, FL. Beverdge, T. J. (1989). Role of cellular design in bacterial metal accumulation and mineralization. Ann. Rev. Microbiol. 43, 147– 171. Beveridge, T. J., Makin, S. A., Kadurugamuwa, J. L., and Li, Z. (1997). Interactions between bioﬁlms and the environment. FEMS Microbial Rev. 20, 291–303. Bigham, J. M., Fritzpatrick, R. W., and Schulze, D. G. (2002). Iron oxides. In “Soil Mineralogy with Environmental Applications” (J. B. Dixon and D. G. Schulze, Eds.), pp. 323–366. Soil Sci. Soc. Am., Madison, WI, SSSA Book Ser. 7. Bollag, J.-M. (1992). Decontaminating soil with enzymes: an in situ method using phenolic and anilinic compounds. Environ. Sci. Technol. 26, 1876–1881. Bollag, J.-M., Dec, J., and Huang, P. M. (1998). Formation mechanisms of complex organic structures in soil habitats. Adv. Agron. 63, 237 –266.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

461

Brown, L. R., and Flavin, C. (1996). “State of the World”. W.W. Norton Co., New York. Brown, C. B., and White, J. L. (1969). Reactions of 12 s-triazines with soil clays. Soil Sci. Soc. Am. Proc. 33, 863– 867. Burns, R. G. (1986). Interactions of enzymes with soil mineral and organic colloids. In “Interactions of Soil Minerals with Natural Organics and Microbes. SSSA Spec. Publ. 17” (P. M. Huang and M. Schnitzer, Eds.), pp. 429– 451. Soil Sci. Soc. Am., Madison, WI. Burns, R. G., and Stach, J. E. M. (2002). Microbial ecology of soil bioﬁlms: Substrate bioavailability, bioremediation, and complexity. In “Soil Mineral–Organic Matter–Microorganism Interactions and Ecosystem Health. Vol. 28B. Ecological Signiﬁcance of the Interactions Among Clay Minerals, Organic Matter and Soil Biota” (A. Violante, P. M. Huang, J.-M. Bollag, and L. Gianfreda, Eds.), pp. 17–42. Elsevier, Amsterdam. Butcher, S. S., Charlson, R. J., Orions, G. H., and Wolfe, G. V. (1992). “Global Biogeochemical Cycles”. Academic Press, London. Buyanovsky, G. A., Aslam, M., and Wagner, G. H. (1994). Carbon turnover in soil physical fractions. Soil Sci. Soc. Am. J. 58, 1167–1173. Carter, M. R. (1996). Analysis of soil organic matter storage in agroecosystems. In “Structure and Organic Matter Storage in Agricultural Soils” (M. R. Carter and B. S. Stewart, Eds.), pp. 3–11. Adv. Soil Sci., CRC Press, Boca Raton, FL. Chang, A. C., Page, A. L., and Koo, B.-J. (2002). Biogeochemistry of phosphorus, iron, and trace elements in soils as inﬂuenced by soil — plant–microbial interactions. In “Soil Mineral– Organic Matter –Microorganism Interactions and Ecosystem Health. Vol. 28B. Ecological Signiﬁcance of the Interactions Among Clay Minerals, Organic Matter and Soil Biota” (A. Violante, P. M. Huang, J.-M. Bollag, and L. Gianfreda, Eds.), pp. 43–57. Elsevier, Amsterdam. Charlson, R. J., Orians, G. H., Wolfe, G. V., and Butcher, S. S. (1992). Human modiﬁcation of global biogeochemical cycles. In “Global Biogeochemical Cycles” (S. S. Butcher, R. J. Charlson, G. H. Orians, and G. V. Wolfe, Eds.), pp. 353 –361. Academic Press, London. Cheney, M. A., Sposito, G., McGrath, A. E., and Criddle, R. S. (1996.). Abiotic degradation of 2,4-D (dichlorophenoxyacetic acid) on synthetic birnessite: calorespirometric method. Colloids Surf. A. Phys. Eng. Asp 107, 131–140. Cheng, H. H. (1990). “Pesticides in the Soil Environment: Processes, Impacts, and Modeling”. Soil Sci. Soc. Am., Madison, WI. Chorover, J., Amistadi, M. K., Burgos, W. D., and Hatcher, P. G. (1999). Quinoline sorption on kaolinite–humic acid complexes. Soil Sci. Soc. Am. J. 63, 850–857. Christensen, T. H., and Huang, P. M. (1999). Solid phase cadmium and the reactions of aqueous cadmium with soil surfaces. In “Cadmium in Soils and Plants” (M. J. McLaughlin and B. R. Singh, Eds.), pp. 65 –96. Kluwer Academic Publishers, Dordrecht, The Netherlands. Claus, H., and Filip, Z. K. (1990). Effects of clays and other solids on the activity of phenoloxidases produced by some fungi and actinomycetos. Soil Biol. Biochem. 22, 483–488. Coleman, D. C., Hendrix, P. F., and Odum, E. P. (1998). Ecosystem health: an overview. In “Soil Chemistry and Ecosystem Health” (P. M. Huang, D. C. Adriano, T. J. Logan, and R. T. Checkai, Eds.), pp. 1–20. Soil Sci. Soc. Am., Madison WI, SSSA Spec. Publ. 52. Collins, Y. E., and Stotzky, G. (1996). Changes in the surface charge of bacteria caused by heavy metals does not affect survival. Can. J. Microbiol. 42, 621– 627. Conrad, J. P. (1942). The occurrence and origin of urease-like activities in soils. Soil Sci. 54, 367–380. Conway, G., and Toenniessen, G. (1999). Feeding the twenty-ﬁrst century. Nature (London) 402, C55–C58. Cornell, R. M., and Schwertmann, U. (1996). “The Iron Oxides”. VCH Publ., Weinheim, Germany. Crocker, F. H., Guerin, W. F., and Boyd, S. A. (1995). Bioavailability of naphthalene sorbed to cationic surfactant-modiﬁed smectite clay. Environ. Sci. Technol. 29, 2953–2958. Crosby, D. G. (1982). Environmental chemistry: an overview. Environ. Toxicol. Chem. 1, 1–8.

462

P. M. HUANG

Cunningham, S. D., and Berti, W. R. (1993). Remediation of contaminated soils with green plants: an overview. In Vitro. Cell. Dev. Bio. 29P, 207 –212. Dec, J., and Bollag, J.-M. (2000). Phenoloxidase-mediated interactions of phenols and anilines with humic materials. J. Environ. Qual. 29, 665–676. Dec, J., Bollag, J.-M., Huang, P. M., and Senesi, N. (2002). Impact of interactions between microorganisms and soil colloids on transformation of organic pollutants. In “Interactions Between Soil Particles and Microorganisms. Impact on Terrestrial Ecosystem”. (P. M. Huang, J.M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 323–378. IUPAC Series of Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, England. Di Grazia, P. M., Blackburn, J. W., Bienkowski, P. R., Hilton, B., Reed, G. D., King, J. M. H., and Sayler, G. S. (1990). Development of a system analysis approach for resolving the structure of biodegrading systems. Appl. Biochem. Biotechnol. 24/25, 237–252. Douglas, S., and Beveridge, T. J. (1998). Mineral formation by bacteria in natural microbial communities. FEMS Microbiol. Ecol. 26, 79–88. Ehrlich, H. L. (2002). Interactions between microorganisms and minerals under anaerobic conditions. In “Interactions Between Soil Particles and Microorganisms. Impact on the Terrestrial Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 459 –494. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK. Emerson, W. W., Foster, R. C., and Oades, J. M. (1986). Organo-mineral complexes in relation to soil aggregation and structure. In “Interactions of Soil Minerals with Natural Organics and Microbes, SSSA Spec. Publ. 17” (P. M. Huang and M. Schnitzer, Eds.), pp. 521–548. Soil Sci. Soc. Am., Madison, WI. Evershed, R. P., Bland, H. A., van Bergen, P. F., Carter, J. F., Horton, M. C., and Rowley-Conwy, P. A. (1997). Volatile compounds in archaeological plant remains and the Maillard reaction during decay of organic matter. Science 278, 432– 433. Farrell, R. E., Huang, P. M., and Germida, J. J. (1998). Biomethylation of mercury (II) adsorbed on mineral colloids common in freshwater sediments. Appl. Organomet. Chem. 12, 613–620. Filip, Z., and Hattori, T. (1984). Utilization of substrates and transformation of solid substrata. In “Microbial Adhesion and Aggregation” (K. C. Marshall, Ed.), pp. 251– 282. Dahlem Konferenzen, Springer-Verlag, Berlin. Filip, Z., Haider, K., and Martin, J. P. (1972). Inﬂuence of clay minerals on growth and metabolic activity of Epicoccum nigrum Stachybotrys chartarum. Soil Biol. Biochem. 4, 135–145. Filip, Z., Flaig, W., and Rietz, E. (1977). Oxidation of some phenolic substances as inﬂuenced by clay minerals. In “Isot. Radiat. Soil Org. Matter Stud., II”, pp. 91– 96. IAEA, Bull., Vienna. Fletcher, M. (1991). The physiological activity of bacteria attached to solid surfaces. Adv. Microb. Phyiol. 32, 53 –85. Fletcher, M. (1996). “Bacterial Adhesion: Molecular and Ecological Diversity”. Wiley-Liss, New York. Gadd, G. M. (1993). Microbial formation and transformation of organometallic and organometalloid compounds. FEMS Microbiol. Rev. 11, 297– 310. Getzin, L. W., and Roseﬁeld, I. (1971). Partial puriﬁcation and properties of a soil enzyme that degrades the insecticide malathion. Biochim. Biophys. Acta 235, 442–453. Gianfreda, L., and Bollag, J.-M. (1994). Effect of soils on the behavior of immobilized enzymes. Soil Sci. Soc. Am. J. 58, 1672–1681. Goh, T. B., and Huang, P. M. (1986). Inﬂuence of citric and tannic acids on hydroxy-Al interlayering in montmorillonite. Clays Clay Miner. 34, 37–44. Golchin, A., Baldock, J. A., and Oades, J. M. (1997). A model linking organic matter decomposition, chemistry and aggregate dynamics. In “Soil Processes and the Carbon Cycle” (R. Lal, et al., Ed.) pp. 245 –266. CRC Press, Boca Raton, FL. Goldstein, L. (1976). Kinetic behavior of immobilized enzyme systems. In “Methods in Enzymology, Volume XLIV” (K. Mosbach, Ed.), pp. 397–443. Academic Press, Orlando, FL.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

463

Guerin, W. F., and Boyd, S. A. (1992). Differential bioavailability of soil-sorbed naphthalene to two bacterial species. Appl. Environ. Microbiol. 58, 1142–1152. Guggenberger, G., and Haider, K. M. (2002). Effect of mineral colloids on biogeochemical cycling of C, N, P and S in soil. In “Interactions Between Soil Particles and Microorganisms. Impact on the Terrestrial Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 267–322. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK. Haider, K. (1995). Sorption phenomena between inorganic and organic compounds in soils: Impacts on transformation processes. In “Environmental Impact of Soil Component Interactions. Volume I Natural and Anthropogenic Organics” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L. Page, Eds.), pp. 21–45. CRC Press/Lewis Publishers, Boca Raton, FL. Hatcher, P. G., Schnitzer, M., Dennis, L. W., and Maciel, G. E. (1981). Aromaticity of humic substances in soils Soil Sci. Soc. Am. J. 45, 1089–1094. Hatzinger, P. B., and Alexander, M. (1995). Effect of aging of chemicals in soil on their biodegradability and extractability. Environ. Sci. Technol. 29, 537– 545. Hayes, K. F., and Traina, S. J. (1998). Metal ion speciation and its signiﬁcance in ecosystem health. In “Soil Chemistry and Ecosystem Health” (P. M. Huang, D. C. Adriano, T. J. Logan, and R. T. Chechai, Eds.), pp. 45–84. Soil Sci. Soc. Am., Madison, WI, SSSA Spec. Publ. 52. Hoffmann, M. R. (1980). Trace metal catalysis in aquatic environments. Environ. Sci. Technol. 14, 1061–1066. Hoffmann, M. R. (1990). Catalysis in aquatic environments. In “Aquatic Chemical Kinetics” (W. Stumm, Ed.), p. 71–111. Wiley, New York. Hooper, D. U., and Vitousek, P. M. (1997). The effect of plant composition and diversity on ecosystem processes. Science 277, 1302–1305. Hsu, P. H. (1988). Mechanisms of gibbsite crystallization from partially neutralized aluminum chloride solutions. Clays Clay Miner. 36, 25–30. Huang, P. M. (1980). Adsorption processes in soil. In “The Handbook of Environmental Chemistry, Vol. 2/Part A. Reactions and Processes” (O. Hutzinger, Ed.), pp. 47 –59. Springer-Verlag, Berlin. Huang, P. M. (1988). Ionic factors affecting aluminum transformation and the impact on soil and environmental sciences. Adv. Soil Sci. 8, 1–78. Huang, P. M. (1990). Role of soil minerals in transformations of natural organics and xenobiotics. In “Soil Biochemistry Volume 6” (J.-M. Bollag and G. Stotzky, Eds.), pp. 29 –115. Marcel Dekker, New York. Huang, P. M. (1991). Ionic factors affecting the formation of short-range ordered aluminosilicates. Soil Sci. Soc. Am. J. 55, 1172–1180. Huang, P. M. (1995). The role of short-range ordered mineral colloids in abiotic transformations of organic compounds in the environment. In “Environmental Impact of Soil Component Interactions. Volume I, Natural and Anthropogenic Organics” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L. Page, Eds.), pp. 135–167. CRC Press/Lewis Publishers, Boca Raton, FL. Huang, P. M. (2000a). Abiotic catalysis. In “Handbook of Soil Science” (M. E. Sumner, Ed.), pp. B303–B332. CRC Press, Boca Raton, FL. Huang, P. M. (2000b). “Environmental Soil Chemistry and Its Impact on Agriculture and the Ecosystem”. Universita` degli Studi di Napoli Federico II, Portici, Italy. Huang, P. M. (2002). Foreseeable impacts of soil mineral–organic component–microorganism interactions on society: ecosystem health. In “Soil Mineral–Organic Matter– Microorganism Interactions and Ecosystem Health”. (A. Violante, P. M. Huang, J.-M. Bollag, and L. Gianfreda, Eds.), Vol. 28A, pp. 1–36. Development in Soil Science. Elsevier, Amsterdam. Huang, P. M., and Bollag, J.-M. (1999). Minerals — Organics–Microorganisms Interactions in the Soil Environment. In “Structure and Surface Reactions of Soil Particles”. (P. M. Huang, N. Senesi, and J. Bufﬂe, Eds.), Vol. 4, pp. 3 –39. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK.

464

P. M. HUANG

Huang, P. M., and Germida, J. J. (2002). Chemical and biological processes in the rhizosphere: metal pollutants. In “Interactions Between Soil Particles and Microorganisms. Impact on the Terrestrial Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 381– 438. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK. Huang, P. M., and Kozak, L. M. (1970). Adsorption of hydroxy-aluminum polymers by muscovite and biotite. Nature (London) 228, 1084–1085. Huang, P. M., and Schnitzer, M. (1986). “Interactions of Soil Minerals with Natural Organics and Microbes”. Soil Sci. Soc. Am., Madison, WI, SSSA Spec. Publ. 17. Huang, P. M., and Violante, A. (1986). Inﬂuence of organic acids on crystallization and surface properties of precipitation products of aluminum. In “Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 159 –221. Soil Sci. Soc. Am., Madison, WI, SSSA Spec. Publ. 17. Huang, P. M., Wang, M. C., and Wang, M. K. (1999). Catalytic transformation of phenolic compounds in the soils. In “Principles and Practices in Plant Ecology. Allelochemicals interactions” (Inderjit, K. M. M. Dakshini and C. L. Foy, Eds.), pp. 287 –306. CRC Press, Boca Raton, FL. Huang, P. M., Wang, M. K., Ka¨mpf, N., and Schulze, D. G. (2002). Aluminum hydroxides. In “Soil Mineralogy with Environmental Applications” (J. B. Dixon and D. G. Schulze, Eds.), pp. 261 –289. Soil Sci. Soc. Am., Madison, WI, SSSA Book Ser. 7. Hughes, M. N., and Poole, R. K. (1989). “Metals and microorganisms”. Chapman and Hall, New York. Ikan, R., Rubinsztain, Y., Nissenbaum, A., and Kaplan, I. R. (1996). Geochemical aspects of the Maillard reaction. In “The Maillard Reaction: Consequences for the Chemical and Life Sciences” (R. Ikan, Ed.), pp. 1–25. Wiley, Chichester, UK. Inoue, K., and Huang, P. M. (1984). Inﬂuence of citric acid on the natural formation of imogolite. Nature (London) 308, 58–60. Jastrow, J. D., and Miller, R. M. (1997). Soil aggregate stabilization and carbon sequestration: feedbacks through organomineral associations. In “Soil Processes and the Carbon Cycle” (R. Lal, et al., Eds.), pp. 207 –223. CRC Press, Boca Raton, FL. Jenkinson, D. S., and Rayner, J. H. (1977). The turnover of soil organic matter in some of the Rothamsted classical experiments. Soil Sci. 123, 298–305. Jenkinson, D. S., Adams, D. E., and Wild, A. (1991). Model estimates of CO2 emissions from soil in response to global warming. Nature (London) 351, 304– 306. Johnston, C. T., and Tomba´cz, E. (2002). Surface chemistry of soil minerals. In “Soil Mineralogy with Environmental Applications” (J. B. Dixon and D. G. Schulze, Eds.), pp. 37–67. Jokic, A., Frenkel, A. I., Vairavamurthy, M. A., and Huang, P. M. (2001a). Birnessite catalysis of the Maillard reaction: its signiﬁcance in natural humiﬁcation. Geophys. Res. Lett. 28, 3899–3902. Jokic, A., Frenkel, A. I., and Huang, P. M. (2001b). Effect of light on birnessite catalysis of the Maillard reaction and its implication in humiﬁcation. Can. J. Soil Sci. 81, 277–283. Ka¨mpf, N., Scheinost, A. C., and Schulze, D. G. (2000). Oxide minerals. In “Handbook of Soil Science” (M. E. Sumner, Ed.), pp. F 125 –168. CRC Press, Boca Raton, FL. Karthikeyan, K. G., Chorover, J., Bortiatynski, J. M., and Hatcher, P. G. (1999). Interaction of 1naphthol and its oxidation products with aluminum hydroxides. Environ. Sci. Technol. 33, 4009–4015. Kennedy, M. J., Pevear, D. R., and Hill, R. J. (2002). Mineral surface control of organic carbon in black shale. Science 295, 657–660. Khan, S. U. (1978). Kinetics of hydrolysis of atrazine in aqueous fulvic acid solution. Pestic. Sci. 9, 39–43. Kirk, G. J. D., and Bajita, J. B. (1995). Root-induced iron oxidation, pH changes and zinc solubilization in the rhizosphere of lowland rice. New Phytol. 131, 129–137.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

465

Kjelleberg, S., Humphrey, B. A., and Marshall, K. C. (1982). Effects of interfaces on small, starved marine bacteria. Appl. Environ. Microbial. 43, 1166–1172. Kodama, H., and Schnitzer, M. (1974). Adsorption of fulvic acid by non-expanding clay minerals. Trans. Int. Congr. Soil Sci., 10th 11, 51–56. Kodama, H., and Schnitzer, M. (1980). Effect of fulvic acid on the crystallization of aluminum hydroxides. Geoderma 24, 195– 205. Krishnamurti, G. S. R., and Huang, P. M. (1990). Kinetics of Fe(II) oxygenation and the nature of hydrolytic products as inﬂuenced by ligands. In “Proc. 9th Int. Clay Conf., Strasbourg” (V. C. Farmer and Y. Tardy, Eds.), pp. 195– 204. Publie´ avec le concours de l’Universite´ Louis Pasteur et du Center National de la Recherche Scientiﬁque, Strasbourg, France. Krishnamurti, G. S. R., and Huang, P. M. (1991). Inﬂuence of citrate on the kinetics of Fe(II) oxidation and the formation of iron oxyhydroxides. Clays Clay Miner. 39, 28– 34. Krishnamurti, G. S. R., and Huang, P. M. (1993). Formation of lepidocrocite from Fe(II) solutions: stabilization by citrate. Soil Sci. Soc. Am. J. 57, 861 –867. Krishnamurti, G. S. R., Huang, P. M., Van Rees, K. C. J., Kozak, L. M., and Rostad, H. P. W. (1995a). Speciation of particulate-bound Cd in soils and its bioavailability. Analyst 120, 659 –665. Krishnamurti, G. S. R., Huang, P. M., Van Rees, K. C. J., Kozak, L. M., and Rostad, H. P. W. (1995b). A new soil test method for the determination of plant-available cadmium in soils. Commun. Soil Sci. Plant Anal. 26, 2857–2867. Krishnamurti, G. S. R., Huang, P. M., and Van Rees, K. C. J. (1996). Studies on soil rhizosphere: speciation and availability of Cd. Chem. Speciat Bioavailab 8, 23–28. Krishnamurti, G. S. R., Cieslinski, G., Huang, P. M., and Van Rees, K. C. J. (1997). Kinetics of cadmium release from soils as inﬂuenced by organic acids: implication in cadmium availability. J. Environ. Qual. 26, 271 –277. Krishnamurti, G. S. R., Wang, M. K., and Huang, P. M. (1999). Role of tartaric acid in the inhibition of the formation of Al13 tridecamer using sulfate precipitation. Clays Clay Miner. 47, 658–663. Kumada, K., and Kato, H. (1970). Browning of pyrogallol as affected by clay minerals. Soil Sci. Plant Nutr. 16, 195–200. Kurek, E. (2002). Microbial mobilization of metals from soil minerals under aerobic conditions. In “Interactions Between Soil Particles and Microorganisms. Impact on the Terrestrial Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 189–225. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK. Kwong, K. F. Ng Kee, and Huang, P. M. (1975). Inﬂuence of citric acid on the crystallization of aluminum hydroxides. Clays Clay Miner. 23, 164–165. Kwong, K. F. Ng Kee, and Huang, P. M. (1978). Sorption of phosphate by hydrolytic products of aluminum. Nature (London) 271, 336–338. Kwong, K. F. Ng Kee, and Huang, P. M. (1979a). The relative inﬂuence of low-molecular-weight complexing organic acids on the hydrolysis and precipitation of aluminum. Soil Sci. 128, 337 –342. Kwong, K. F. Ng Kee, and Huang, P. M. (1979b). Nature of hydrolytic precipitation products of aluminum as inﬂuenced by low molecular weight complexing organic aicds. In “Int. Clay Conf. Proc., 6th, Oxford” (V. C. Farmer and M. M. Mortland, Eds.), pp. 527 –536. Elsevier, Oxford, UK. Kwong, K. F. Ng Kee, and Huang, P. M. (1979c). Surface reactivity of aluminum hydroxides precipitated in the presence of low-molecular-weight organic acids. Soil Sci. Soc. Am. J. 43, 1107–1113. Kwong, K. F. Ng Kee, and Huang, P. M. (1981). Comparison of the inﬂuence of tannic acid and selected low-molecular-weight organic acids on precipitation products of aluminum. Geoderma 26, 179–193. Kyuma, K., and Kawaguchi, K. (1964). Oxidative changes of polyphenols as inﬂuenced by allophane. Soil Sci. Soc. Am. Proc. 28, 371– 374.

466

P. M. HUANG

La˚g, J. (1980). Survey of Geomedical Problems: Including Some Examples from Investigations Carried Out in Norway “Geomedical Aspects in Present and Future Research”. Universitatsforlaget, Olso. La˚g, J. (1994). Geomedicine, an expanded application of soil science. In “Interactions of Soil Components, Agricultural Ecosystems, and Health”. Vol. 3a, pp. 557 –567, Transactions 15th World Congress of Soil Science. Commission II Symposium, Acapulco, Mexico. Lal, R. (1998). “Soil Processes and the Carbon Cycle”. CRC Press, Boca Raton, FL. Laurie, S. H., Tancock, N., McGrath, S. P., and Sanders, J. R. (1991). Inﬂuence of complexation on the uptake by plants of iron, manganese, copper and zinc: Effect of DTPA in a multi-metal and computer simulation study. J. Exp. Bot. 42, 509–513. Li, G. C., and Felbeck, G. T. Jr. (1972). A study of the mechanism of atrazine adsorption by humic acid from muck soil. Soil Sci. 113, 140 –148. Linehan, D. J., Sinclair, A. H., and Mitchell, M. C. (1989). Seasonal changes in Cu, Mn, Zn and Co concentrations in soil in the root zone of barley (Hordeum vulgare L). J. Soil Sci. 40, 103–115. Liu, C., and Huang, P. M. (1999). Atomic force microscopy and surface characteristics of iron oxides formed in citrate solutions. Soil Sci. Soc. Am. J. 63, 65–72. Logan, T. J. (1998). Impact of soil chemical reactions on food chain contamination and environmental quality. In “Future Prospects for Soil Chemistry” (P. M. Huang, D. L. Sparks, and S. A. Boyd, Eds.), pp. 191–204, SSSA Spec. Publ. 55, Soil Sci. Soc. Am., Madison, WI. Lund, W. (1990). Speciation analysis — why and how? Fresenius J. Anal. Chem. 337, 557–564. Lynch, J. M. (1990a). Introduction: some consequences of microbial rhizosphere competence for plant and soil. In “The Rhizosphere” (J. M. Lynch, Ed.), p. 1–10. Wiley, Chichester, UK. Lynch, J. M. (1990b). Microbial metabolites. In “The Rhizosphere” (J. M. Lynch, Ed.), p. 177– 206. Wiley, New York. Mackay, D., and Paterson, S. (1991). Evaluating the multimedia fate of organic chemicals: a level III fugacity model. Environ. Sci. Technol. 25, 427 –436. Maillard, L. C. (1913). Formation de matie`res humiques par action de polypeptides sur sucres. C.R. Acad. Sci. 156, 148 –149. Malcolm, R. L. (1989). Application of solid-state 13C NMR spectroscopy to geochemical studies of humic substances. In “Humic Substances II. In Search of Structure” (M. H. B. Hayes, P. MacCarthy, R. L. Malcolm, and R. S. Swift, Eds.), pp. 339 –372. Wiley, Chichester, UK. Mann, S., Archibald, D. A., Didymus, J. M., Douglas, T., Heywood, R. R., Meldrum, F. C., and Reeves, N. J. (1993). Crystallization at inorganic – organic interfaces: biominerals and biomimetic synthesis. Science 261, 1286–1292. Marschner, H. (1998). Soil– root interface: biological and biochemical processes. In “Soil Chemistry and Ecosystem Health” (P. M. Huang, D. C. Adriano, T. J. Logan, and R. T. Checkai, Eds.), pp. 191 –231, SSSA Spec. Publ. 52, Soil Sci. Soc. Am., Madison WI. Marshall, K. C. (1971). Sorptive interactions between soil particles and microorganisms. In “Soil Biochemistry” (A. D. McLaren and J. Skujins, Eds.), Vol. 2, pp. 409–442. Marcel Dekker, New York. Marshall, K. C. (1980). Adsorption of microorganisms to soils and sediments. In “Adsorption of Microorganisms to Surfaces” (G. Britton and K. C. Marshall, Eds.), pp. 317– 330. Wiley, New York. Martinek, K., and Mozhaev, V. V. (1985). Immobilization of enzymes: an approach to fundamental studies in biochemistry. Adv. Enzymol. 57, 179–249. Masschelyn, P. H., and Patrick, W. H. Jr. (1994). Selecium, arsenic and chromium redox chemistry in wetland soils and sediments. In “Biochemistry of Trace Elements” (D. C. Adriano, Ed.), pp. 615 –625. Science and Technology Letters, Northwood. McBride, M. B. (1989). Reactions controlling heavy metal solubility in soil. Adv. Soil Sci. 10, 1 –56. McBridge, M. B. (1994). “Environmental Chemistry of Soils”. Oxford University Press, Oxford.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

467

McGill, W. B., and Cole, C. V. (1981). Comparative aspects of cycling of organic C, N, S and P through soil organic matter. Geoderma 26, 267 –286. McLaughlin, M. J., Smolders, E., and Merckx, R. (1998). Soil– root interface: physicochemical processes. In “Soil Chemistry and Ecosystem Health.” (P. M. Huang, D. C. Adriano,, T. J. Logan,, and R. T. Checkai,, Eds.), pp. 233 –277, SSSA Spec. Publ. 52, Soil Sci. Soc. Am., Madison, WI. McLean, J. S., Lee, J.-U., and Beveridge, T. L. (2002). Interactions of bacteria and environmental metals, ﬁne-grained mineral development and bioremediation strategies. In “Interactions Between Soil Particles and Microorganisms. Impact on the Terrestrial Ecosystem”. (P. M. Huang, M. J.-Bollag, and N. Senesi, Eds.), Vol. 8, pp. 227 –261. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK. Mingelgrin, U., Saltzman, S., and Yaron, B. (1977). A possible model for the surface-induced hydrolysis of organophosphorus pesticides on kaolinite clays. Soil Sci. Soc. Am. J. 41, 519–523. Morra, M. J. (1996). Bioremediation in soil: inﬂuence of soil properties on organic contaminants and bacteria. In “Bioremediation: Principles and Applications” (R. L. Crawford and D. L. Crawford, Eds.), pp. 35–60. Cambridge University Press, Cambridge, UK. Mortland, M. M. (1986). Mechanisms of adsorption of nonhumic organic species by clays. In “Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 59–76, SSSA Special Publication No. 17. Soil Sci. Soc. Am., Madison, WI. Mortland, M. M., and Halloran, L. J. (1976). Polymerization of aromatic molecules on smectites. Soil Sci. Soc. Am. J. 40, 367–370. Munnecke, D. M., Johnson, L. M., Tablot, H. W., and Barik, S. (1982). Microbial metabolism and enzymology of selected pesticides. In “Biodegradation and Detoxiﬁcation of Environmental Pollutants” (A. M. Chakrabarty, Ed.), pp. 1 –32. CRC Press, Boca Raton, FL. Myneni, S. C. B., Tokunaga, T. K., and Brown, G. E. Jr. (1997). Abiotic selenium redox transformations in the presence of Fe(II, III) oxdies. Science 278, 1108–1109. Naeem, S., Thompson, L. J., Lawler, S. P., Lawton, J. H., and Woodﬁn, R. M. (1994). Declining biodiversity can alter the performance of ecosystems. Nature (London) 368, 734–737. Naidja, A., Huang, P. M., and Bollag, J.-M. (2000). Enzyme–Clay Interactions and Their Impact on Transformations of Natural and Anthropogenic Organic compounds in Soils. J. Environ. Qual. 29, 677–691. Naidja, A., and Huang, P. M. (1996). Deamination of aspartic acid by aspartase — Ca – montmorillonite complex. J. Mol. Catal. 106, 255–265. Naidja, A., Huang, P. M., and Bollag, J.-M. (1997). Activity of tyrosinase immobilized on hydroxyaluminum –montmorillonite complexes. J. Mol. Catal. A: Chem. 115, 305 –316. Naidja, A., Liu, C., and Huang, P. M. (2002). Formation of protein– birnessite complex: XRD, FTIR, and AFM analysis. J. Colloid Interface Sci. 251, 46 –56. Nannipieri, P., and Gianfreda, L. (1999). Kinetics of enzyme reactions in soil environments. In “Structure and Surface Reactions of Soil Particles.” (P. M. Huang, N. Senesi, and J. Bufﬂe, Eds.), Vol. 4, pp. 449 – 479. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK. National Academy of Sciences (U.S.), (1974). Geochemistry and the Environment. “The Relation of Selected Trace Elements to Health and Disease”, Vol. I. Washington, DC. National Academy of Sciences (U.S.), (1997). Geochemistry and the Environment. “The Relation of Other Selected Trace Elements to Health and Disease”, Vol. II. Washington, DC. Nelson, Y. M., Lion, L. W., Ghiorse, W. C., and Shuler, M. L. (1999). Production of biogenic Mn oxides by Leptothrix discophora SS-1 in a chemically deﬁned growth medium and evaluation of their Pb adsorption characteristics. Appl. Environ. Microbial. 65, 175–180. Nielands, J. B. (1981). Microbial iron compounds. Ann. Rev. Biochem. 50, 715 –731. Nielsen, N. E. (1976). The effect of plants on the copper concentration in the soil solution. Plant Soil 45, 679–687.

468

P. M. HUANG

NIST, (1997). Standard Reference Database 46. “Critically Selected Stability Constants of Metal Complexes, Version 5.0”. National Institute of Standards and Technology, US Department of Commerce, Gaithersburg, MD. Oades, J. M., and Waters, A. G. (1991). Aggregate hierarchy in soils. Aust. J. Soil Res. 29, 815 –828. Oades, J. M., Gillman, G. P., and Uehara, G. (1989). Interactions of soil organic matter and variablecharge clays. In “Dynamics of Soil Organic Matter in Tropical Ecosystems”. (D. C. Coleman, J. M. Oades, and G. Uehara, Eds.), pp. 69–95. Nif TAL Project. University of Hawaii, Honolulu. Odum, E. P. (1989). “Ecology and our endangered life-support systems”. Sinauer Assoc., Sunderland, MA. Oscarson, D. W., Huang, P. M., Defosse, C., and Herbillon, A. (1981). Oxidative power of Mn(IV) and Fe(III) oxides with respect to As(III) in terrestrial and aquatic environments. Nature (London) 291, 50 –51. Oscarson, D. W., Huang, P. M., Hammer, U. T., and Liaw, W. K. (1983). Oxidation and sorption of arsenite by manganese dioxide as inﬂuenced by surface coatings of iron and aluminum oxides and calcium carbonate. Water Air Soil Poll. 20, 233–244. Otte, M. L., Dekkers, M. J., Rozema, J., and Broekman, R. A. (1991). Uptake of arsenic by Aster tripolium in relation to rhizosphere oxidation. Can. J. Bot. 69, 2670–2677. Otte, M. L., Kearns, C. C., and Doyle, M. O. (1995). Accumulation of arsenic and zinc in the rhizosphere of wetland plants. Bull. Environ. Contam. Toxicol. 55, 154 –161. Otte, M. L., Rozem, J., Koster, L., Haarsma, M. S., and Broekman, R. A. (1989). Iron plaque on roots of Aster tripolium L.: interaction with Zn uptake. New Phytol. 111, 309–317. Parker, D. R., Chaney, R. L., and Norvell, W. A. (1995). Chemical equilibrium models: applications to plant nutrition. In “Soil Chemical Equilibrium and Reaction Models”. “SSSA Spec. Publ.” (R. H. Loeppert, et al., Eds.), Vol. 42, pp. 163–200. American Society of Agronomy and Soil Science Society of America, Madison, WI. Parton, W. J., Schimel, D. S., Cole, C. V., and Ojima, D. S. (1987). Analysis of factors controlling soil organic matter levels in Great Plains grasslands. Soil Sci. Soc. Am. J. 51, 1173– 1179. Paul, E. A., and Van Veen, J. A. (1978). The use of tracers to determine the dynamics of organic matter. Trans. Int. Soc. Soil Sci. (Edmonton) 3, 61–102. Perdue, E. M., and Wolfe, N. L. (1982). Modiﬁcation of pollutant hydrolysis kinetics in the presence of humic substances. Environ. Sci. Technol. 16, 847–852. Pierzynski, G. M., Sim, J. T., and Vance, G. F. (2000). “Soil and Environmental Quality”. CRC Press, Boca Raton, FL. Pinnavaia, T. J., Hall, P. L., Cady, S. S., and Mortland, M. M. (1974). Aromatic radical cation formation on intracrystal surfaces of transition metal layer silicates. J. Phys. Chem. 78, 994– 999. Preston, C. M., Rauthan, B. S., Rodger, C., and Ripmeester, J. A. (1982). A hydrogen-1, carbon-13, and nitrogen-15 nuclear magnetic resonance study of p-benzoquinone polymers incorporating amino nitrogen compounds ("synthetic humic acids"). Soil Sci. 134, 277 –293. Quiquampoix, H. (1987). A stepwise approach to the understanding of extracellular enzyme activity in soil. I. Effect of electrostatic interactions on the conformation of a b-D-glucosidase adsorbed on different mineral surfaces. Biochimie 69, 753– 763. Reid, C. P. P. (1990). Mycorrhizas. In “The rhizosphere” (J. M. Lynch, Ed.), pp. 281–315. Wiley, New York. Rich, C. I. (1968). Hydroxy interlayers in expansible layer silicates. Clays Clay Miner. 16, 15–30. Robert, M., and Berthelin, J. (1986). Role of biological and biochemical factors in soil mineral weathering. In “Interactions of Soil Minerals with Natural Organics and Microbes. SSSA Spec. Publ. 17” (P. M. Huang and M. Schnitzer, Eds.), pp. 453–495. Soil Sci. Soc. Am., Madison, WI. Ruggiero, P., Sakar, J. M., and Bollag, J.-M. (1989). Detoxiﬁcation of 2,4-dichlorophenol by a laccase immobilized on soil and clay. Soil Sci. 147, 361–370. Ruggiero, P., Dec, J., and Bollag, J.-M. (1996). Soil as a catalytic system. In “Soil Biochemistry Volume 9” (G. Stotzky and J.-M. Bollag, Eds.), pp. 79. Marcel Dekker, New York.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

469

Salts, D. E., Blaylock, M., Kumar, N., Dushenkov, V., Ensley, B. D., Chet, I., and Raskin, I. (1995). Phytoremediation: a novel strategy for the removal of toxic metals from the environment using plants. Biotechnology, 13, 468 –474. Scheffer, F., Meyer, B., and Niederbudde, E. A. (1959). Huminstaffbildung unter katalyischer Einwirkung natu¨rlich vorkommender Eisenverbindungen im Modellversuch. Z. Pﬂanzenernaehr. Bodenkd. 87, 26–44. Schnitzer, M. (1977). Recent ﬁndings on the characterization of humic substances extracted from soils from widely differing climatic zones. In “Soil Organic Matter Studies II”, pp. 77–101. IAEA Bull., Vienna. Schnitzer, M. (1995). Organic–inorganic interactions in soils and their effects on soil quality. In “Environmental Impact of Soil Component Interactions, Vol. I Natural and Anthropogenic Organics” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L. Page, Eds.), pp. 3–45. CRC/Lewis Publishers, Boca Raton, FL. Schnitzer, M. (1986). Binding in humic substances by soil mineral colloids. In “Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 77–101. Soil Sci. Soc. Am., Madison, WI, SSSA Special Publication No. 17. Schnitzer, M. (2000). A lifetime perspective on the chemistry of soil organic matter. Adv. Agron. 68, 1 –58. Schnitzer, M., and Khan, S. U. (1972). “Humic Substances in the Environment”. Marcel Dekker, New York. Schnitzer, M., and Kodama, H. (1966). Montmorillonite: effect of pH on its adsorption of a soil humic compound. Science, 153, 70–71. Schnitzer, M., and Preston, C. M. (1983). Effects of acid hydrolysis on the 13C NMR spectra of humic substances. Plant Soil, 75, 201–211. Schwertmann, U. (1966). Inhibitory effect of soil organic matter on the crystallization of amorphous ferric hydroxide. Nature (London) 212, 645 –646. Schwertmann, U. (1985). The effect of pedogenetic environments on iron oxide minerals. Adv. Soil Sci. 1, 171 –200. Schwertmann, U., and Taylor, R. M. (1989). Iron oxides. In “Minerals in Soil Environments” (J. B. Dixon and S. B. Weed, Eds.), pp. 379–438. Soil Sci. Soc. Am., Madison WI. Schwertmann, U., Kodama, H., and Fisher, W. R. (1986). Mutual interactions between organics and iron oxides. In “Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 223 –250. Soil Sci. Soc. Am., Madison, WI, SSSA Spec. Publ. 17. Senesi, N., and Miano, T. M. (1995). The role of abiotic interactions with humic substances on the environmental impact of organic pollutants. In “Environmental Impact of Soil Component Interactions. Volume I Natural and Anthropogenic Organics” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L. Page, Eds.), pp. 311– 335. CRC/Lewis Publishers, Boca Raton, FL. Shimp, R. J., and Young, R. L. (1988). Availability of organic chemicals for biodegradation in settled bottom sediments. Ecotoxicol. Environ. Safety 15, 31–45. Shindo, H., and Huang, P. M. (1982). Role of Mn(IV) oxide in abiotic formation of humic substances in the environment. Nature (London) 298, 363–365. Shindo, H., and Huang, P. M. (1984). Catalytic effects of manganese (IV), iron (III), aluminum, and silicon oxides on the formation of phenolic polymers. Soil Sci. Soc. Am. J. 48, 927–934. Shindo, H., and Huang, P. M. (1985a). The catalytic power of inorganic components in the abiotic synthesis of hydroquinone-derived humic polymers. Appl. Clay Sci. 1, 71–81. Shindo, H., and Huang, P. M. (1985b). Catalytic polymerization of hydroquinone by primary minerals. Soil Sci. 139, 505–511. Singer, A., and Huang, P. M. (1990). Effect of humic acid on the crystallization of aluminum hydroxides. Clays Clay Miner. 38, 47–52. Sparks, D. L. (2002). “Environmental Soil Chemistry”. Academic Press, San Diego, CA.

470

P. M. HUANG

Sposito, G. (1984). “The Surface Chemistry of Soils”. Oxford University Press, Oxford. Sposito, G., and Reginato, R. J. (1992). “Opportunities in Basic Soil Science Research”. Soil Sci. Soc. Am., Madison, WI. Stevenson, F. J. (1994). Humus Chemistry “Genesis, Composition, Reactions.”, 2nd ed. Wiley, New York. Stone, A. T., and Torrents, A. (1995). The role of dissolved metals and metal-containing surfaces in catalyzing the hydrolysis of organic pollutants. In “Environmental Impact of Soil Component Interactions.” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L. Page, Eds.), Vol. I., pp. 275–298. Natural and Anthropogenic Organics. CRC/Lewis Publishers, Boca Raton, FL. Stotzky, G. (1986). Inﬂuence of soil mineral colloids on metabolic processes, growth, adhesion, and ecology of microbes and viruses. In “Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 305–428. Soil Sci. Soc. Am., Madison, WI, SSSA Spec. Publ. 17. Stotzky, G. (2002). Clays and humic acids affect persistence and biological activity of insecticidal proteins from Bacillus Thuringiensis in soil. In “Soil Mineral–Organic Matter–Microorganism Interactions and Ecosystem Health”. Ecological Signiﬁcance of the Interactions Among Clay Minerals, Organic Matter and Soil Biota. (A. Violante, P. M. Huang, J.-M. bollag, and L. Gianfreda, Eds.), Vol. 28B, pp. 1–16. Elsevier, Amsterdam. Struif, B., Weil, L., and Quentin, K. E. (1975). The behavior of herbicides, phenoxy acetic acids and their esters in water. Vom. Wasser 45, 53 –73. Szmigielska, A. M., Cieslinski, G., Van Rees, K. C. J., and Huang, P. M. (1997). Low molecular weight dicarboxylic acids in rhizosphere soil of durum wheat. Commun. Soil Sci. Plant Anal. 28, 99–111. Tan, K. H. (1986). Degradation of soil minerals by organic acids. In “Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 1 –27. Soil Sci. Soc., Am., Madison, WI, SSSA Spec. Publ. 17. Tebo, B. M., Ghiorse, W. C, van Waasbergen, L. G., Siering, P. L., and Caspi, R. (1997). Bacterially mediated mineral formation: insights into manganese (II) oxidation from molecular genetic and biochemical studies. In “Geomicrobiology: Interactions Between Microbes and Minerals, Reviews in Mineralogy” (J. F. Banﬁeld and K. H. Nelson, Eds.), Vol. 35, pp. 225– 266. Mineralogical Society of America, Washington, DC. Tessier, A., and Turner, D. R. (1995). “Metal Speciation and Bioavailability in Aquatic Systems”. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems, Vol. 3. Wiley, Chichester, UK. Theng, B. K. G. (1979). “Formation and Properties of Clay–Polymer Complexes”. Elsevier, New York. Theng, B. K. G., and Orchard, V. A. (1995). Interactions of clays with microorganisms and bacterial survival in soil: a physicochemical perspective. In “Environmental Impact of Soil Component Interactions”. “Metals, Other Inorganics and Microbial Activities” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. G. McGill, and A. L. Page, Eds.), Vol. II, pp. 123 –143. CRC/Lewis Publishers, Boca Raton, FL. Theng, B. K. G., Tate, K. R., and Sollins, P. (1989). Constituents of organic matter in temperate and tropical soils. In “Dynamics of Soil Organic Matter in Tropical Ecosystems” (D. C. Coleman, J. M. Oades, and G. Uehara, Eds.), pp. 5–31, Nif TAL Project, University of Hawaii, Honolulu. Tiller, K. G. (1996). Soil contamination issues: past, present and future, a personal perspective In “Contaminants and the Soil Environment in the Australasia-Paciﬁc Region” (R. Naidu, et al., Eds.), pp. 1–27. Kluwer Academic Publ., Dordrecht, The Netherlands. Tilman, D., and Downing, J. A. (1994). Biodiversity and stability in grassland. Nature (London) 367, 363–365.

SOIL MINERAL – ORGANIC MATTER– MICROORGANISM

471

Tilman, D., Wedlin, D., and Knops, J. (1996). Productivity and sustainability inﬂuenced by biodiversity in grassland ecosystems. Nature (London) 379, 718–720. Torn, M. S., Trumbore, S. E., Chadwick, O. A., Vitousek, P. M., and Hendricks, D. M. (1997). Mineral control of soil organic carbon storage and turnover. Nature (London) 389, 170–173. U.S. Environmental Protection Agency., (1995). “A Guide to the Biosolid Risk Assessments for the USEPA Part 503 Rule” Ofﬁce of Wastewater Management. USEPA, Washington, DC, USEPA, 832-B-93-005. van der Heijden, M. G. A., Klironomos, J. N., Ursic, M., Moutoglis, P., Streitwolf-Engel, R., Boller, T., Wiemken, A., and Sanders, I. R. (1998). Mycorrhizal fungal diversity determines plant biodiversity, ecosystem variability and productivity. Nature (London) 396, 69–72. Venkateswerlu, G., Yorder, M. J., and Stotzky, G. (1989). Morphological ultra-structural, and chemical changes induced in Cunninghamella blakesleeana by copper and cobalt. Appl. Microbiol. Biotechnol. 31, 204–210. Violante, A., and Gianfreda, L. (2000). Role of biomolecules in the formation and reactivity towards nutrients and organics of variable charge minerals and organo-mineral complexes in soil environment. In “Soil Biochemistry” (J.-M. Bollag and G. Stotzky, Eds.), Vol. 10, pp. 207–270. Marcel Dekker, New York. Violante, A., Huang, P. M., Bollag, J.-M., and Gianfreda, L. (2002a). Soil Mineral –Organic Matter– Microorganism Interactions and Ecosystem Health. In “Dynamics, Mobility and Transformation of Pollutants and Nutrients.” Vol. 28A. Elsevier, Amsterdam. Violante, A., Huang, P. M., Bollag, J.-M., and Gianfreda, L. (2002b). Soil Mineral– Organic Matter– Microorganism Interactions and Ecosystem Health. In “Ecological Signiﬁcance of the Interactions Among Clay Minerals, Organic matter and Soil Biota.”, Vol. 28B. Elsevier, Amsterdam. Vitrousek, P. M., Ehrlich, P. R., Ehrlich, A. H., and Mattson, P. A. (1986). Human appropriation of the products of photosynthesis. Bio. Sci. 36, 368–373. Wada, K. (1989). Allophane and imogolite. In “Minerals in Soil Environments”, 2nd ed. (J. B. Dixon and S. B. Weed, Eds.), pp. 1051–1087. Soil Sci. Soc. Am., Madison, WI, SSSA Book Ser. 1. Wada, K. (1995). Role of aluminum and iron in the accumulation of organic matter in soils with variable charge. In “Environmental Impact of Soil Component Interactions”. “Natural and Anthropogenic Organics.” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L. Page, Eds.), Vol. 1, pp. 47– 58. CRC/Lewis Publishers, Boca Raton, FL. Wada, K., and Higashi, T. (1976). The categories of aluminum– and iron –humus complexes in Ando soils determined by selective dissolution. J. Soil Sci. 27, 357– 368. Wada, S., Ichikawa, H., and Tatsumi, K. (1992). Removal of phenols with tyrosinase immobilized on magnetite. Water Sci. Technol. 26, 2057– 2059. Waisel, Y., Eshel, A., and Kafkaﬁ, U. (1996). “Plant Roots. The Hidden Half.”, 2nd Ed. Marcel Dekker, New York. Walker, S. G, Flemming, C. A., Ferris, F. G., Beveridge, T. J., and Bailey, G. W. (1989). Physicochemical interaction of Escherichia coli cell envelopes and Bacillus subtilis cell walls with two clays and ability of the composite to immobilize heavy metals from solution. Appl. Environ. Microbiol. 55, 2976– 2984. Walton, B. T., and Anderson, T. A. (1992). Plant-microbe treatment systems for toxic waste. Current Opinion Biotech. 3, 267–270. Wang, M. C., and Huang, P. M. (1986). Humic macromolecular interlayering in nontronite through interaction with phenol monomers. Nature (London) 323, 529 –531. Wang, M. C., and Huang, P. M. (1987). Polycondensation of pyrogallol and glycine and the associated reactions as catalyzed by birnessite. Sci. Tot. Environ. 62, 435 –442. Wang, M. C., and Huang, P. M. (1988). Catalytic power of nontronite, kaolinite, and quartz and their reaction sites in the formation of hydroquinone-derived polymers. Appl. Clay Sci. 4, 43– 57. Wang, M. C., and Huang, P. M. (1992). Signiﬁcance of Mn(IV) oxide in the abiotic ring cleavage of polyphenol in natural environments. Sci. Tot. Environ. 113, 147 –157.

472

P. M. HUANG

Wang, M. C., and Huang, P. M. (1994). Structural role of polyphenols in inﬂuencing the ring cleavage and related chemical reactions as catalyzed by nontronite. In “Humic Substances in the Global Environment and Implications in Human Health.” (N. Senesi and T. M. Miano, Eds.), pp. 173 –180. Elsevier, Amsterdam. Wang, M. C., and Huang, P. M. (2000a). Characteristics of pyrogallol-derived polymers formed by catalysis of oxides. Soil Sci. 165, 737–747. Wang, M. C., and Huang, P. M. (2000b). Ring cleavage and oxidative transformation of pyrogallol catalyzed by Mn, Fe, Al, and Si oxides. Soil Sci. 165, 934–942. Wang, T. S. C., and Li, S. W. (1977). Clay minerals as heterogeneous catalysts in preparation of model humic substances. Z. Pﬂanzenernaehr. Bodenkd. 140, 669 –676. Wang, T. S. C., Huang, P. M., Chou, C. H., and Chen, J. H. (1986). The role of soil minerals in the abiotic polymerization of phenolic compounds and formation of humic substances. In “Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 251 –281. Soil Sci. Soc. Am., Madison WI, SSSA Spec. Publ. 17. Wang, T. S. C., Wang, M. C., and Huang, P. M. (1983a). Catalytic synthesis of humic substances by using alumina as catalysts. Soil Sci. 136, 226 –230. Wang, T. S. C., Wang, M. C., Ferng, Y. L., and Huang, P. M. (1983b). Catalytic synthesis of humic substances by natural clays, silts, and soils. Soil Sci. 135, 350–360. Weber, J. B., and Coble, H. D. (1968). Microbial decomposition of diquat adsorbed on montmorillonite and kaolinite clays. J. Agric. Food Chem. 16, 475–478. Weed, S. B., Davey, C. B., and Cook, M. G. (1969). Weathering of micas by fungi. Soil Sci. Soc. Am. Proc. 33, 702–706. Xue, J., and Huang, P. M. (1995). Zn adsorption–desorption on short-range ordered iron oxides as inﬂuenced by citric acid during its formation. Geoderma 64, 343–356.

LOW EXTERNAL INPUT TECHNOLOGIES FOR LIVELIHOOD IMPROVEMENT IN SUBSISTENCE AGRICULTURE Anil Graves,1 Robin Matthews1 and Kevin Waldie2 1

Institute of Water and Environment, Cranﬁeld University, Silsoe, Bedfordshire MK45 4DT, United Kingdom 2 International and Rural Development Department, University of Reading, Reading RG6 6AH, United Kingdom

I. Introduction II. The Technologies A. Intercropping B. Alley Cropping C. Cover Crops and Green Manures D. Biomass Transfer Techniques E. Compost F. Animal Manure G. Improved Fallows III. Generic Issues A. Soil Fertility Management B. Socio-economic Issues IV. Discussion A. Integrated Nutrient Management B. A Systems Perspective C. Modelling V. Concluding Remarks Acknowledgements References

I. INTRODUCTION The global population is currently predicted to reach 9.3 billion people by the year 2050, a 50% increase over the current level, after which it will level off due to falling fertility rates and family sizes (UNPD, 2001). Of these, 84% will live 473 Advances in Agronomy, Volume 82 Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved 0065-2113/03 $35.00

474

A. GRAVES, R. MATTHEWS AND K. WALDIE

in countries that are currently classiﬁed as being in the developing world (Pretty, 1999). This rise in population, together with a desire for a wider variety in diet brought about by greater purchasing power through a steady improvement in incomes, is predicted to increase food demand over the period 1990 – 2050 by 2.4 times in Asia, 1.9 times in Latin America and the Caribbean, and 5-fold in Africa (FAO, 1996). Moreover, the increase in production required to meet this demand will need to be achieved with less water, less labour, and less land, and without adversely affecting the environment (Dowling et al., 1998). How this is to be achieved is the topic of some debate (e.g., Pretty, 1999; Crosson and Anderson, 2002). Over the 40 years since 1960, the global population has doubled; despite this, food production has more than kept pace, resulting in a 24% increase in per capita world food production and a 40% reduction in food prices in real terms (although these ﬁgures do mask some striking imbalances — per capita food production has fallen 20% in Africa, for example). The total number of undernourished people in the world has also fallen signiﬁcantly over the same period. This has been largely achieved by the use of “Green Revolution” technologies, i.e., high-yielding cereal varieties, together with high levels of inputs such as water from irrigation systems, fertiliser to provide the nutrients needed by the varieties, and pesticides to control any associated weeds, pests and diseases. These technologies generally need a relatively high capital investment, either by, or on the part of farmers, and also need a well-functioning economic and physical infrastructure for effective implementation. However, an estimated 30 –35% of the world’s population (i.e., 1.9 – 2.1 billion people) do not have access to such infrastructures, are remote from markets, practice subsistence agriculture on marginal soils, and lack access to knowledge on how to improve their situation (Pretty, 1999). One school of thought is that a similar high external input agriculture (HEIA) approach as used in the last 40 years can also be used to address the demand for food in the next 50 years by improving the productivity of this group of subsistence farmers, perhaps using new emerging technologies such as genetic modiﬁcation (e.g., Crosson and Anderson, 2002). A second school of thought is that such an approach is not sustainable, and moreover, is damaging to the environment as the inputs of fertilisers and chemicals accumulate in neighbouring ecosystems. Thus, technologies using low levels of external inputs readily available either on-farm or from nearby off-farm sources are seen by some experts as more appropriate and sustainable (Pretty, 1995). This approach, often referred to as low external input agriculture (LEIA), emphasises the use of techniques that integrate natural processes such as nutrient cycling, biological nitrogen ﬁxation (BNF), soil regeneration and natural enemies of pests, into food production processes (Pieri, 1995; Snapp et al., 1998). Efforts are also made to minimise losses from the system, such as by leaching or removal of crop residues.

LOW EXTERNAL INPUT TECHNOLOGIES

475

The use of non-renewable inputs such as pesticides and fertilisers that can damage the environment or harm the health of farmers and consumers is also minimised, and more emphasis is placed on the use of such techniques as, for example, intercropping, agroforestry, cover crops, or animal manure. Usually, but not always, such technologies are more labour-intensive than the HEIA approach (Deugd et al., 1998). In many cases, LEIA technologies are not new, but are variations of those practised by farmers for generations, who have sought to make use of resources such as vegetation or animal manure that have always been ready to hand. Wolf (1986) has estimated that about 1.4 billion people (25% of the world’s population), depend on this type of agriculture for their livelihood. Thus, the heart of the debate is not about whether either approach “works,” as clearly both do, and have done, under the appropriate conditions and according to their own criteria. Rather, the central question concerns which approach can best address the future demand for food production while protecting the environment as much as possible. Within this general debate, more speciﬁc questions relate to whether LEIA technologies really have the capability to maintain or increase productivity per unit area above current levels (e.g., Crosson and Anderson, 2002). Certainly, there is evidence to suggest that the relative rate of increase in crop yields through the use of Green Revolution technologies is slowing (Mann, 1997), although Crosson and Anderson (2002) argue that this is more likely due to the practice of quoting annual percentage increases of a constantly increasing baseline rather than absolute annual growth. Proponents of LEIA technologies often claim that the reliance on local sources of inputs is more sustainable, but the analysis of De Jager et al. (2001) suggests there is little difference between the two approaches in this respect, with both mining similar quantities of soil nutrients to generate farm income. However, despite the continuing debate on the relative performance of the two approaches, there are few studies that compare yields and production under the same soil and climatic conditions and over wide areas. With LEIA technologies in particular, there is little in the literature on the issues that need to be faced in scaling up production from plot level to supplying inputs and meeting food demand on a larger scale. The purpose of this review is to examine research that has been conducted in recent years on a number of LEIA technologies in the context of subsistence agriculture — ﬁrstly, to determine our current state of knowledge, and secondly, to assess their potential to improve livelihoods of subsistence farmers and contribute to long-term sustainability of the resource base. The review was prompted by concerns within the United Kingdom Department for International Development (DFID) that, despite considerable research on, and dissemination of, a wide range of LEIA technologies, there had been little clear evidence of widespread uptake by farmers. Although this could partly be ascribed to inadequate attention to promotion pathways and dissemination by research centres, there was some concern that more fundamental reasons may be

476

A. GRAVES, R. MATTHEWS AND K. WALDIE

responsible. In this review, therefore, we have attempted to step back and take a wider view of some of these technologies, examining both their biophysical and socio-economic aspects to try understand the reasons for their variable uptake by farmers. We have focused on those technologies designed to address problems of soil fertility and weed control; these include intercropping, alley cropping, cover cropping and green manuring, biomass transfer, compost, animal manure, and improved and enriched fallows.

II. THE TECHNOLOGIES A. INTERCROPPING Intercropping is the growing of two or more crops on the same piece of land within the same year. Various forms of intercropping have been a central feature of many tropical agricultural systems for centuries. Vandermeer (1989) has proposed that intercropping can be divided into three general categories — full, relay and sequential intercropping — depending on the extent of physical association between the crops. Full intercropping involves complete association between crops planted at the same time, while relay cropping involves only partial association, in which a second crop is planted into an already standing crop before it is harvested. Sequential intercropping, where there is no physical association, is the extreme case where two crops are grown on the same land in the same year but not at the same time. Cover crops are a special case of intercropping and are discussed in more detail in Section II.C; in this section, we discuss cases where the intercrop components are both food crops. The main advantages of intercropping are in reducing the risk of total crop failure, and in product diversiﬁcation — food crops are often mixed with cash crops to help ensure both subsistence and disposable income (Vandermeer, 1989; Singh and Jodha, 1990). There is some evidence to suggest that in intercropping systems, the microclimate surrounding the lower crop is more conducive to plant growth than in a sole crop (Matthews et al., 1991), and that an intercrop is more efﬁcient at using resources such as light and water (Azam-Ali et al., 1990). Cereals and legumes are often mixed, but probably more for dietary reasons than for any beneﬁcial effect from the nitrogen-ﬁxing ability of the legumes. In Zimbabwe, for example, farmers intercrop sorghum with cowpeas, pumpkins, cucumbers and watermelon to provide nutritional and livelihood beneﬁts (Chivasa et al., 2000). Thus, any soil fertility beneﬁts that can be obtained by intercropping leguminous grain crops with other food crops should probably be seen as a useful spin-off rather than the main purpose of the practice. Moreover, main crop yields can even be reduced by intercropping techniques, both as a result of loss of land to the legume, and also to competition for resources

LOW EXTERNAL INPUT TECHNOLOGIES

477

(Vandermeer, 1989; Snapp et al., 1998). In the long-term, cereal/legume intercrops are still likely to require fertilisers for the provision of phosphorus (P), potassium (K) and micro-nutrients in order to maintain satisfactory yields (Coultas et al., 1996; Kumwenda et al., 1997a). Nevertheless, yield advantages through intercropping are well documented. For example, Rao (2000) found that, compared to sole maize, maize intercropped with pigeonpea (Cajanus cajan) in Kenya yielded 24% more and was 49% more proﬁtable, even though the pigeonpea was affected by pests and diseases. In an interesting variation, sequential intercropping of rose (Rosa damascena) with potatoes, maize and cowpea greatly increased the land equivalent ratio (LER) and provided large economic gains (Yaseen et al., 2001). Similarly, positive effects on soil fertility improvement have been observed. Kumara Rao et al. (1981) estimated that leaf abscission during the growth of a pigeonpea intercrop was equivalent to the addition of between 10 and 40 kg N ha21. The root system of pigeonpea may also recycle N from deeper layers, and in some areas, the build-up of sub-surface nitrates at about 1 –3 m has been observed (Farrell et al., 1996; Hartemink et al., 1996). Morris et al. (1990) observed N transfer from arrow-leaf clover to rye grass, and suggested that in mixed stands of legumes and non-legumes, direct transfer of N during the growing season was possible, although this was likely to be 10% or less of the total N ﬁxed. It is also possible that growing non-legumes with legumes encourages legumes to respond by ﬁxing more N than they might do in a pure stand, so long as the legumes dominate the mixture (Marschner, 1995). Snapp et al. (1998) considered soybean (Glycine max), pigeon pea (C. cajan), groundnuts (Arachis hypogaea), dolichos bean (Dolichos lablab), and cowpea to be among the most promising grain legumes in southern Africa for both food provision and fertility enhancement. Although grain legume intercrops can often help to increase the resource use efﬁciency and stabilise yields of the main crop under optimal plant growing conditions, this is not always the case. In India, Indonesia, and the Philippines, Ali (1999) found that although intercropping could help to increase the yield of rice, it also increased the variability of yield. Many green manure crops have been selected on the assumption that maximum production is desirable. However, yield stability under adverse conditions may be more important to many farmers than high productivity under good conditions. Ironically, it is usually under adverse conditions that intercrop competition is most intense, and it is during these conditions that the farmer can least afford the technology to fail. Intercropping is most likely to be practised on small farms, in areas where land is scarce, forcing the simultaneous production of different crops on the same area of land. For example, Ali (1999), in a survey of data from India, Nepal and the Philippines, found that the attractiveness of intercropping increased as land and labour costs grew. Lower rainfall and/or a unimodal distribution of rain may encourage intercropping as farmers try to maximise their use of water, although,

478

A. GRAVES, R. MATTHEWS AND K. WALDIE

in the extreme, this can result in competition for a scarce resource. Mixedcropping techniques are also more likely to be used by farmers relying on handheld implements for tillage (Ruthenberg, 1980). The need for simultaneous production of different food crops and/or cash crops can also encourage intercropping. Relatively better-off farmers with large farms are less reliant on intercropping, being able to fallow and/or control production with other inputs such as water and inorganic fertilisers. The Machobane farming system in Lesotho (IIRR, 1998) is an example of a system incorporating intercropping. Developed in the 1950s by an agronomist, James Machobane, and based on experiments on his own farm, the approach is a complex, integrated farming system designed to improve the productivity of small-scale mountain farms in Lesotho. Based on 0.4 ha (1 acre), an application of 7.5 t FW ha21 of a mixture of animal manure and wood ash (from household ash) is made each year, with the proportion of each depending on soil type. Enough for the 0.4 ha can generally be met from household waste (1 – 2 t year21 of animal manure, 2 t year21 of ash, (Pantanali, 1996)). Wheat, peas, and possibly potatoes as a cash crop, are planted as intercrops in April –May for harvesting the following January – March, and summer crops such as maize, beans, sorghum, and possibly pumpkins and water melons, are planted in August –October for harvesting in November –December (Fig. 1). Crop residues are left in the ﬁeld to allow humus to build up, and the ﬁeld is ploughed only once every ﬁve years. The intercropping/relay-cropping pattern allows food crops to be produced almost all the year round, and there is always some crop cover throughout the year so that erosion, a major problem in Lesotho, is minimised. Despite the crop cover, weeding is essential, and represents a major labour input into the system. Some of the crops (e.g., pumpkins) help to reduce pests, and the use of chemical pesticides is discouraged. Overall, labour inputs are high, and perhaps reﬂecting this, annual productivity is three times higher than the traditional system, allowing a household of ﬁve people to be self-sufﬁcient on 0.4 ha of land (Pretty, 1999). Moreover, the potato crop is a source of cash, and income ﬂuctuations over the year are lower through less reliance on a single crop. Farmers also claim a better resistance of crops to drought. Promotion of the approach emphasises self-reliance, appreciation of the resource base, readiness for hard work, learning by doing, and a duty to help neighbours, on the part of farmers, and lack of success of the system is sometimes blamed on non-compliance with some of those conditions (Pantanali, 1996). Due to the high labour inputs (peak labour demand estimated at 14 days (0.4 ha)21 month21), critics of the system regard it as little more than “gardening,” with limited relevance to Lesotho’s major agrarian problems. Unfortunately, little data appear to have yet been collected to measure the impact of the system in terms of production, gross income and net returns to labour, either compared to traditional cropping methods, or to recommended

LOW EXTERNAL INPUT TECHNOLOGIES

479

Figure 1 Crop calendar for the Machobane farming system in Lesotho (from IIRR, 1998).

“modern” improved technology practices based on monocropping, mechanical power, chemical fertilisers, and pesticides (Pantanali, 1996). Long-term sustainability of the system, however, depends on the production of animal manure, and hence the availability of pasture, which fortunately is available in the mountains. As with many LEIA systems, therefore, there is a reliance on nutrients collected and concentrated from a much wider area. More manure could be made available if less was used for fuel, but in this case, fuelwood would have to be grown, which again could be done on the land saved through intensiﬁcation of production. Thus, through careful integration of crops, livestock and trees, the long-term sustainability of the system seems possible (Pantanali, 1996). In relation to its applicability to other areas, the system’s economic sustainability rests on being able to grow crops all year round, which will not be feasible in areas with a pronounced dry season unless irrigation is available. Even in Lesotho, the system cannot be practised in all areas due to severe winter conditions with snowfall preventing the growth of many of the crops (Pantanali, 1996).

B. ALLEY CROPPING Alley cropping is an agroforestry practice developed in the 1970s at the International Institute for Tropical Agriculture in Nigeria (Kang et al., 1981), in

480

A. GRAVES, R. MATTHEWS AND K. WALDIE

which hedgerows of trees and shrubs are established and annual crops are cultivated in the alleys between the hedgerows. The hedgerows are pruned before planting the crop and periodically while it is growing to prevent shading, with the prunings being applied to the soil as green manure and/or mulch. Between cropping cycles, hedgerows are usually allowed to grow without pruning. It was originally hoped that by incorporating fast-growing nitrogen-ﬁxing woody perennials with crops, their ability to cycle nutrients, suppress weeds, and reduce erosion would create soil conditions similar to those in the fallow phase of shifting cultivation. In this way, the cropping and fallow phases could take place simultaneously on the same land, allowing the land to be cropped for an extended period when long fallow periods are not feasible under the particular socioeconomic conditions. Researchers saw the technology as the combination of farmers’ accumulated traditional wisdom with the efﬁciency of modern science (Kang, 1993). Initial results from on-station experiments were promising. In Nigeria, for example, prunings from Leucaena leucocephala increased maize grain yields from 1.9 to 3.5 t ha21 (Kang et al., 1981), while a Gliricidia sepium alley system on a degraded soil increased maize yields from 1.74 to 2.42 t ha21 (Atta-Krah and Sumberg, 1988). Increases in yields of banana were obtained when alleycropped with Enterolobium cyclocarpum, and of cowpea when alley-cropped with Enterolobium cyclocarpum and Dialium guianense (Oko et al., 2000). In the fourth and ﬁfth years of an alley-crop trial in Burundi, Calliandra calothyrsus increased maize yields by 29 – 63%; Leucaena diversifolia by 27 –43%, and Senna spectabilis by 24 –38% (Akyeampong, 1999). However, these results were obtained in humid regions on soils of high base status. Results from semi-arid regions were less positive. Yields of sorghum, castor and cowpea were found to be lower when alley-cropped with Leucaena than when grown alone (Singh et al., 1989), with the magnitude of the yield depression correlating with distance from the hedgerows. This was attributed to competition between trees and crops for water. Similar depressions of yield under alley cropping were found in Peru (Szott, 1987) and Zambia (Matthews et al., 1992a,b), although, in the latter work, Leuceana was found to have a positive effect on maize yields if lime was applied ﬁrst. In smallholder farms in southern Africa, little beneﬁt was derived from grain legume intercrops, particularly in adverse weather conditions (Mukurumbira, cited in Snapp et al., 1998). Hedgerow intercropping did not increase maize yields in below average rainfall years, indicating that competition by trees predominated over beneﬁts to soil fertility (Snapp et al., 1998). Similarly, Vanlauwe et al. (2001) found reduced crop yields under alley cropping in the absence of mineral fertilizers. Competition, especially for light, is particularly acute with perennial legumes, which are larger than the main crop (Ong, 1994). In general, it appears that where resources are scarce, competition for resources such as water and nutrients is not offset by beneﬁts to fertility (Manu et al., 1994; Rao et al., 1997). Root pruning

LOW EXTERNAL INPUT TECHNOLOGIES

481

has been found to reduce competition for water, but this adds to the labour requirement, and may affect the ability of the tree to access water and nutrients below the crop zone (Ong et al., 2002). In some cases, other negative effects of trees on the crop have been found. For example, Adeorike (2001) found that Inga edulis, Anthonatha macrophylla, and Dactyladenia barterri had allelotrophic effects on maize. In Kenya, decreasing the width between Leucaena hedgerows was found to increase the incidence of angular leaf spot (Phaeoisariopsis griseola) and anthracnose (Colletotrichum lindemuthianum) on beans (Phaseolus vulgaris), with the bean rows closest to the hedgerows the most affected (Koech and Whitbread, 2000). Microclimate changes caused by the hedgerows appeared to best explain the distribution of the diseases on the beans. In some cases, more bird damage has been reported in alley crops than in monocrops (e.g., Hoang Fagerstro¨m et al., 2001). Some control of weeds was achieved in alley crops if the hedgerow canopy was maintained during the fallow period. For example, in Nigeria, uncut hedgerows of Gliricidia and Leucaena decreased the shoot biomass of the weed Imperata cylindrica by about 80% (Anoka et al., 1991). Similarly, Yamoah et al. (1986b) found that unpruned hedgerows of Flemingia macrophylla, Gliricidia sepium, and Cassia siamea were able to reduce weed yields. Shifts in weed composition were also observed — Siaw et al. (1991), for example, reported a signiﬁcant change towards more broadleaf weeds after alley cropping with Leucaena and Dactyladenia barter. In most alley-cropping systems, the weed suppression effect of the hedgerows probably has not been fully exploited, and further studies of the effect of different hedgerow species, fallowing and manipulation of cutting regimes may improve the effectiveness of the system in reducing weed infestation. Alley cropping does seem to have favourable effects on soil physical and chemical properties through the addition of large amounts of organic matter from the prunings. Levels of organic C, total N, extractable P, Mg and K, and pH have been shown to increase under alley cropping under a range of conditions (e.g., Kang et al., 1985; Lal, 1989b; Dalland et al., 1993). Similarly, lower bulk density and penetration resistance, and higher inﬁltration rate and pore volume fraction, were found under Leucaena alley crops in Zambia, which was ascribed to increased levels of soil organic matter (SOM) (Dalland et al., 1993). The magnitude of these effects, however, varied with hedgerow species and management — Leucaena prunings increased crop yields more than Flemingia congesta due to the faster release of nutrients as a result of a lower C/N ratio. There is also some evidence that nutrient recycling is enhanced by alley cropping so that the downward displacement of nutrients is reduced. Hauser (1990), for example, attributed higher concentrations of N, K, Ca and Mg in the surface soil than in the subsoil under Leuceana hedgerows to leaf litter fall and nutrient uptake by the trees from the subsoil. Between the rows, there were lower nutrient

482

A. GRAVES, R. MATTHEWS AND K. WALDIE

levels in the surface soil due to crop uptake and higher levels in the subsoil due to leaching. Alley cropping also seems to have been successful in reducing runoff and soil erosion on sloping land. For example, in Nigeria, Lal (1989a) found a reduction of 73 and 83% in soil erosion under alley crops of Gliricidia and Leucaena, respectively. In the Philippines, Paningbatan (1990) recorded soil erosion over a three-month period of 41 t ha21 with Desmanthus hedgerows and contour cultivation, and only 0.2 t ha21 with hedgerows, application of prunings as a mulch, and zero tillage, compared to 127 t ha21 in the control treatment. In northern Vietnam, contour planting of hedgerows on sloping lands reduced soil loss from 18 to 7.4 t ha21 year21 and also produced 2.5 –12 t DM ha21 year21 for green manure (Nguyen The et al., 2001). Farmers were aware that soil loss resulted from cultivating annual crops on sloping lands without adequate protection, but were constrained by lack of labour and capital (Brodd and Osanius, 2002). Initial uptake of alley cropping by farmers in Nigeria and Benin was good, but nearly half of those who adopted the technology subsequently abandoned it. In Nigeria, no new adoption occurred after 1990 (Douthwaite et al., 2002). Followup studies concluded that the initial enthusiasm shown by farmers was probably more related to the incentives offered by researchers (free establishment of the alley ﬁelds, weeding, provision of animals, animal vaccination, fertilisers and seed of improved crop varieties), and to the prestige of contact with international researchers, than to the characteristics of the technology itself (Whittome, 1994). The Nigerian farmers gave the high labour demand for establishment and management of the hedgerows, and incorporation of the biomass into the soil, as the main constraints. Other studies have shown similar results (e.g., Reynolds et al., 1991; David, 1995; Craswell et al., 1998), with many adopters speciﬁcally citing the labour required for pruning as being the most difﬁcult aspect of alley cropping. Interestingly though, Hoang Fagerstro¨m et al. (2001) note no difference in labour requirements between a monocrop and Tephrosia alley-crop in Vietnam — the extra labour required for hedgerow management was balanced by a reduction in labour associated with crop husbandry, such as weeding. An abundance of available land has also been found to be a factor constraining uptake of alley cropping – Whittome (1994), in his study of farmer experiences in Nigeria and Benin, found that in most cases, land was still sufﬁciently abundant for them not to consider soil fertility decline as a problem, and therefore not to ﬁnd alley cropping attractive. In Kenya, Swinkels and Frankel (1997) found that alley farming was most attractive in areas where the population density was high, farms were small, and labour was plentiful. Access to capital is another key factor. For example, Cenas et al. (1996) have shown that adoption may be higher where farmers have off-farm sources of income, relatively large farms, and were interested in cash cropping. This was not only because of the high labour costs involved in establishment of hedgerows (Nelson et al., 1996), but also because

LOW EXTERNAL INPUT TECHNOLOGIES

483

alley cropping alone could not maintain total soil fertility requirements and money had therefore to be spent on fertiliser for full beneﬁts (Wendt et al., 1994). Moreover, beneﬁts from alley cropping take some time to accrue (Carter, 1995), and resource-poor farmers may feel that these are not realised rapidly enough to meet their current needs and that the long-term beneﬁts do not outweigh the more immediate costs of establishment (Nelson and Cramb, 1998). Indeed, traditional cropping practices often create greater net revenue than alley cropping over the ﬁrst 4– 5 years (Nelson et al., 1996), and this may discourage the use of alley cropping despite greater long-term beneﬁts (Nelson et al., 1998). Security of tenure and long-term access to land are important issues affecting uptake in some countries. Tenant farmers, for example, are unlikely to want to bear the full cost of the technique while the beneﬁt is shared with the landlord (Nelson et al., 1998). Similarly, systems based on revolving cultivation of land amongst family members, short-term tenancy, and share cropping tenancy arrangements may have the same effect. Where farmers have long-term security of tenure over discrete areas of land, alley cropping may be more relevant (Carter, 1995). It is important to note that there are cases of farmers adapting the basic alley cropping practice to ﬁt their own needs. For example, in the Philippines, farmers often increased alley spacing, planted single rather than double hedgerows to reduce planting density, and reduced trimming frequencies and mulch application (Garcia et al., 2002). Some farmers even used alternative tree species so that the hedgerow could be used for other purposes. These modiﬁcations may have reduced the value of alley cropping as a soil fertility-enhancing technique, but have allowed it to ﬁt within the constraints of the farmer and to answer a wider set of needs. In some cases, there has even been an evolution from alley cropping into intercropping two crop species — in eastern Indonesia, for example, Harsono (1996) describes the replacement of hedgerows with strips of grain legumes such as soybean which were shown to increase net proﬁts. In other cases, alley cropping has had some success for reasons other than soil fertility enhancement or erosion control. For example, farmers have planted hedgerows for the provision of poles, medicines, plants, ﬁbre, fruit, and fuel (Cenas et al., 1996), while in the Amarasi district of Indonesia, Field et al. (1992) noted that Leucaena alley crops could provide fodder to allow farmers to develop intensive livestock systems. The return from the sale of cattle could be used for the purchase of food, and as farmers become less reliant on annual crops for subsistence, more appropriate perennial plant systems could be established on steeper land prone to erosion. These modiﬁcations by farmers illustrate an important point regarding adoption of alley cropping, or any technique for that matter — if it is to be successful, it must address a range of farmers’ requirements, some of which may not necessarily be related to the researchers’ original intentions, in this case, those of soil fertility enhancement or erosion control (Douthwaite et al., 2002).

484

A. GRAVES, R. MATTHEWS AND K. WALDIE

Alley cropping (and agroforestry technologies in general) must become more “versatile,” capable of meeting a range of needs in response to changes in the socio-economic circumstances of the farmer (Vosti et al., 1998). Gender issues are also important — adoption is more likely if it is able to meet the needs of both men (e.g., for poles or fodder) and women (e.g., for ﬁbre, fuel-wood and mulch) (Rocheleau and Rocheleau, 1990).

C. COVER CROPS

AND

GREEN MANURES

A cover crop is a crop grown to provide soil cover to prevent erosion by wind and water, regardless of whether it is later incorporated. Green manuring involves the incorporation of a crop while it is still mainly green into the soil for the purpose of soil improvement. Cover crops and green manures are generally annual, biennial, or perennial herbaceous plants grown in a pure or mixed stand during all or part of the year, and as such can be seen as a special case of intercropping. In addition to providing ground cover and, in the case of a legume, producing N, they may also help suppress weeds and reduce insect pests and diseases. Catch crops are cover crops that have been planted speciﬁcally to reduce losses of nutrients by leaching following a main crop. To compete with weeds, which by deﬁnition are aggressive plants, cover crops need to have an appropriate canopy architecture. A spreading cover crop is more likely to suppress weeds than a cover crop with an erect habit. For example, in trials in rice systems, in the Ichilo Sara area of Bolivia, Pound et al. (1999) noted that the performance of Arachis pintoi as a cover crop was highly variable and rejected by farmers due to its inability to suppress weeds (especially Imperata contracta), largely as a result of poor growth and lack of full cover. Similarly, in Ghana, Jackson et al. (1999) found that C. cajan (pigeonpea) was slow-growing, low yielding, and incapable of suppressing weeds due to its poor ground coverage. However, even the use of aggressive cover crops that spread and shade well may fail to suppress the growth of certain shade-tolerant weed species. Pound et al. (1999) found that Mucuna pruriens and Calopogonium mucunoides could not suppress the growth of weed species such as Axonopus compressus, Cyperaceas, Panicum spp., Leersia spp., and shade-tolerant species such as Drymaria, Commelina and Talinum. The duration of the cover crop may also determine its effectiveness in controlling weed growth. If its duration is less than the period between harvest of the main crop and the planting of the next main crop, weeds may proliferate during the time gap, and nutrients may even be released for uptake by weeds rather than by the subsequent crop. For example, in experiments on winter cover crops grown in rotation with rice, Pound et al. (1999) found that certain cover crops with relatively short durations could actually increase the number of weeds in comparison with the traditional practice of leaving a winter fallow. In these

LOW EXTERNAL INPUT TECHNOLOGIES

485

cases, the weeds had time to multiply before the planting date of the next rice crop, possibly taking advantage of N that was ﬁxed by the leguminous cover crops. Cover crops such as alfalfa, cowpea, carioca beans or Mucuna deeringiana had very short durations, and tended to produce comparatively little biomass as a result. Weed counts in these plots were substantially greater than for the traditional fallow. Species with a longer duration, such as Mucuna pruriens, C. cajan, and Canavalia ensiformis, continued growing up to the ﬁrst planting of the rice, produced more biomass, and continued to suppress the growth of weeds over this time, although not signiﬁcantly more than the traditional fallow treatment as far as grass and broadleaf weeds were concerned. Rice yields also showed no signiﬁcant differences between different cover-crop treatments, despite the large variation in grass weed density (Pound et al., 1999). A cover crop with too long a duration may also cause problems. If the use of cover crops in the system is to be sustained, particularly in isolated areas, farmers need to be able to collect seed for the next season. If the crop fails to ﬂower and seed before the next main crop planting, then seed must be obtained externally. Most legume genotypes appear to be adapted to quite narrow biophysical conditions, and, therefore, must be tailored to speciﬁc environments if they are to be successful. Keatinge et al. (1998) showed that Vicia faba, Vicia villosa ssp. dasycarpa, and Lupinus mutabilis would be suitable as autumn-sown cover crops across most of the mid-hills of Nepal if early sowing was possible. Vicia sativa and Trifolium resupinatum, on the other hand, were only likely to mature soon enough at lower elevations. Similar exercises were conducted for hillside regions in Bolivia (Wheeler et al., 1999) in which potential cover crops, not grown locally, were recommended for further trials, and also in Uganda (Keatinge et al., 1999). In some cases, the growth of cover crops grown in association with another crop has been found to be too aggressive, resulting in undue competition with the main crop for nutrients, space, water, and light. If left unchecked this can lead to complete domination of the main crop by the cover crop, i.e., the latter is beginning to behave like a weed itself. For example, Pound et al. (1999) found that Mucuna pruriens was rejected by some farmers in the Ichilo-Sara area of Bolivia because it dominated Bactris gasipaes, a local palm, and banana, limiting their growth and development. Similarly, in on-farm trials of a Calopogonium/ rice intercropping system, farmers found that Calopogonium tended to climb over the rice and cause it too lodge. This occurred especially when rice and Calopogonium were sown simultaneously or when long-duration rice varieties were used (Pound et al., 1999). Nevertheless, cover crops have had some success in addressing problems of soil fertility and weed control. It has been shown that short-term fallows of herbaceous crops such as Mucuna pruriens (velvet bean) and Stylosanthes hamata (Stylo) can help increase main crop yields compared with continuous cropping, and that weed densities can be reduced (Tarawali et al., 1999).

486

A. GRAVES, R. MATTHEWS AND K. WALDIE

Farmers seem to be well aware of these beneﬁts (Buckles and Triomphe, 1999; Franzel, 1999), and because of this, the adoption of cover crops by farmers has been relatively widespread (Sanchez, 1999). In Benin, for example, the ability of Mucuna grown as a cover crop to suppress the weed Imperata cylindrica was discovered almost by accident by researchers with the Recherche Appliquee´ Milieu Ree´l (RAMR) Project, and was subsequently promoted amongst farmers by various formal and informal sector organisations. It was estimated that about 10,000 farmers tested Mucuna between 1988 – 1996 (Tarawali et al., 1999), although Elbasha et al. (1999) report that this was on only 1000 ha. Other estimates put the adoption ﬁgure as high as 100,000 farmers (Versteeg et al., 1998). An adoption study by RAMR showed that about 25% of farmers had adopted the technique (deﬁned as having used it at least twice), whilst about 35% had rejected it despite still having an Imperata problem on their ﬁelds (Versteeg et al., 1998). Adopters cited the need to control Imperata infestations as the primary motivation for their using Mucuna, rather than soil fertility enhancement, although apparently some reported beneﬁts from higher maize yields (3 – 4 t ha21 without application of nitrogen fertilizer and with less labour input for weeding, compared to 1.3 t ha21 without Mucuna (Pretty, 1999)). Non-adopters said that leaving the ﬁeld unproductive during the minor season was a major disincentive, as well as the lack of a use for the grain produced by Mucuna, which is toxic (due to the presence of l-Dopa) unless treated properly (Versteeg et al., 1998). However, in reality, lack of a market was not always a problem, as demand for Mucuna seed grew as use of the technique spread. Interestingly, adoption was lower in areas where land was less scarce, although some farmers in these areas discovered that Mucuna made good fodder for livestock, and could also be used to suppress the parasitic weed Striga hermonthica. The beneﬁt/cost analysis over a period of eight years indicated a ratio of 1.24 when Mucuna was included in the system, and 0.62 for the system without Mucuna, with the ratio as high as 3.56 if Mucuna seeds were sold (UNEP, 1999). Positive returns were achieved in the second year of establishment at both the farm and regional levels. It has been estimated that Mucuna grown as a cover crop can provide more than 100 kg N ha21 to a following maize crop, and, as such, its adoption throughout the Mono Province of Benin would result in savings of about 6.5 million kg of N, or about US$1.85 million year21. However, yearly analysis of the beneﬁt/cost ratio indicated a declining trend over time, suggesting that addition of external inputs (probably P and K fertilizer) are required in order to achieve full sustainability (Pretty, 1999). Widespread adoption of Mucuna as a cover crop has also been evident in Honduras, where, without any extension support, it spread from farmer to farmer since the 1970s, when, due to population pressure on the plains, hillside areas were ﬁrst used for agricultural production (Buckles and Triomphe, 1999). Traditional practice was to crop maize twice a year in the wet and dry seasons for

LOW EXTERNAL INPUT TECHNOLOGIES

487

two years, then return the ﬁeld to fallow for four years, a system which was labour intensive (land preparation, planting, weed control), land extensive (due to the need to have additional plots to meet food requirements during the fallow), and capital intensive (due to outlay on herbicides and labour for weeding). In the new system, farmers plant Mucuna as a relay crop into the maize in the dry season so that it grows alone over the wet season, suppressing other weed species at the same time. At the beginning of the next dry season, the farmer slashes the Mucuna and sows the following maize crop directly into the decaying mulch, and so the cycle continues. The system gives beneﬁts in terms of reduced labour (15 – 20% less) and increased yields after the second year — while the traditional system provided four harvests over six years from a single plot, the maize – Mucuna system produces six harvests with yields 50 –100% higher. On average, those who have adopted the Mucuna technique planted twice as much maize as those who did not. Despite this, the total amount of land occupied by their cropping system was less, as they no longer needed large areas to fallow, although, interestingly, overall deforestation rates continued to increase due to an inﬂux of migrants into the area (Humphries, 1996). Experimental evidence indicated that the system was capable of maintaining soil N and OC, Ca, pH and P levels. This was largely achieved through a large biomass production of about 10– 12 t ha21, and large amounts of N (about 300 N kg ha21) being contributed through this biomass, although, of course, only a proportion of this represents a net addition to the system through nitrogen ﬁxation. By 1992, 65% of farmers were using the maize – Mucuna system, with a further 19% having used it in the past (Buckles and Triomphe, 1999). However, more recently there has been a widespread decline in the number of farmers using the system; such that with abandonment running at 10% per year, only 39% of farmers were still using it by 1997. Neill and Lee (2001) provide interesting insights into why this has occurred. Their surveys showed that there was no single overriding factor involved. Firstly, a minimum farm size of 2– 3 ha is required to meet household food requirements during the wet season while some ﬁelds are under Mucuna, and farmers must have security of land tenure to adopt the system. However, over the last 30 years, there has been a steady decline in farm size and insecurity of tenure has increased. Secondly, the increase in extensive cattle production in the region has decreased the availability of land for rent and off-farm work, both necessary requirements for the smallest farmers if they are to adopt the maize –Mucuna system. Thirdly, improved road access in the region has made other alternatives, such as fruit-trees and off-farm work, more attractive than maize growing. Fourthly, the incursion of the noxious weed Rottboellia cochinchinensis into the area in the last few years has drastically reduced maize yields (by 50 –72%) by establishing in gaps within the Mucuna caused by farmers relying on it to regenerate itself rather than deliberately reseeding it at the beginning of each new season. The presence of Rottboellia has negated the two advantages of the maize – Mucuna system — that of lower

488

A. GRAVES, R. MATTHEWS AND K. WALDIE

labour requirements and higher productivity. Thus, it is clear that the sustainability of a technique is a function both of the agronomic performance of a system, and of the socio-economic context in which it is located, including the knowledge and understanding of the farmers. Neill and Lee (2001) make the point that farmers’ inadequate grasp of agronomic aspects such as reseeding, nutrient cycles, nitrogen ﬁxation and herbicide effects, brought about by too rapid adoption of the system through hearing about it from neighbours, may also have contributed to the decline of the system. In both the West African and Central American cases, adoption of Mucuna cover cropping was due to a speciﬁc set of biophysical and socio-economic circumstances, which may not always be the same in other countries. In Benin, the critical conditions associated with its adoption were a decline in soil fertility, the lack of fertilizer, land scarcity, and weed encroachment, all of which combined to induce farmers to adopt a new technique which might not otherwise have been accepted (UNEP, 1999). In semi-arid Africa, however, establishment and management of cover crops has proven difﬁcult where rainfall is less than 800 mm or on very clayey soils, so the practice is not likely to have much success there (Ganry et al., 2001). In Hondurus, adoption was widespread as farmers appreciated the beneﬁts of the technique, which included higher maize yields, improved soil fertility with ease of land preparation, and moisture conservation, with weed control and erosion control of lesser importance (Buckles and Triomphe, 1999). The seasonality of maize prices also encouraged the uptake of the technique, as maize planted during the dry season commanded a higher than average value. These factors all helped to improve productivity both to land and to labour. The beneﬁts were such that farmers who rented land were even willing to pay a premium on land that had been under the Mucuna technique. This also encouraged landlords who owned more land than they were able to cultivate under maize to invest in the technique. Thus, while rental of land generally discourages investment in techniques by tenants, the availability of land for rent and the value placed on Mucuna-treated ﬁelds encouraged the initial spread of the technique in this particular case. In general, therefore, intermediate intensities of land use are likely to be the best context for cover crops, as the opportunity cost of land will be high in land extensive systems, while the opportunity cost of labour and capital will be high in land extensive systems (Tarawali et al., 1999). Security of land tenure is also important as several years are required to reduce weeds and enhance soil physical and chemical properties to a level that might sustain another arable rotation (Tarawali et al., 1999). Some ﬁnancial and labour investment in herbicides and fertiliser may also be necessary as evidence suggests that cover crops are not capable of indeﬁnitely sustaining arable crop production, even in rotations (Tarawali et al., 1999). Purchase of seed can also be costly and this has been identiﬁed as a major constraint in some areas. Those farmers with intermediate levels of wealth and/or off-farm incomes may prefer to use cover crops in

LOW EXTERNAL INPUT TECHNOLOGIES

489

improved fallows (Tarawali et al., 1999), as the opportunity cost of their labour is relatively high (Franzel, 1999). Poorer farmers are likely to make use of natural fallow unless they are provided with credit facilities and/or other incentives (Tarawali et al., 1999) or have a large labour pool. However, the opportunity cost of capital, land, and labour may be relatively high. Wealthier farmers may decide to use inorganic fertilisers during the cropping cycle, although as they often have fairly large farms, part of the farms may be under fallow. Labour demand and the timeliness of that demand may also be problematic and cover crops will probably have the best chance of being adopted by households with sufﬁcient surplus labour. This can sometimes be difﬁcult as labour availability in rural areas often declines as farmers attempt to broaden their livelihood strategies with off-farm work. In other areas, cover crops have been found in niche uses with high-value tree crops, even though little extension effort has been made for this. For example, Thomas et al. (1991) discuss the use of cover crops as green manures as a source of nutrients for coconut. Similarly, Lehmann (2000) found that Pueraria phaseoloides cover crops in tropical fruit-tree systems enhanced nutrient cycling and reduced leaching, although competition for other nutrients (e.g., P) was a problem. In the tropical rainforest of Nigeria, farmers found that cover cropping was the most effective way of controlling soil erosion (Akinyemi and Kalejaiye, 1998). Another possible use is for feed meal production (Stur et al., 2002). It should be noted that cover crops have not always controlled weeds or increased crop yields, primarily due to competition with the main crop. For example, maize intercropped with the cover crops Pueraria phaseoloides, Vigna unguiculata (cowpea), and Mucuna pruriens in SW Nigeria yielded 2.15 t ha21 (þ 15% increase), 1.92 t ha21 (þ 3%), and 1.71 t ha21 (2 9%), respectively, compared to the 1.87 t ha21 for maize grown with no cover crop, the differences reﬂecting the relative competitive ability of the legume crops (Kirchhof and Salako, 2000). In some cases, the decomposition of green manure (particularly legumes) can release organic acids and/or other compounds which may affect germination, seedling growth, and yield of the following crop (Boddey et al., 1997). In Costa Rica, intercropping maize with Mucuna deeringiana and Canavalia ensiformis helped to control itchgrass (Rottboellia cochinchinensis) infestation to some extent, which usually resulted in improved maize yields, although on occasions yields were lower due to competition from the cover crop, especially Mucuna (Valverde et al., 1999). Similarly, in Benin, although the main reason for the widespread adoption of Mucuna cover cropping was its ability to control weeds, it was found that Imperata, in particular, was suppressed but not eliminated, and that it regained its original strength after one or two years of maize cropping (UNEP, 1999). This was because Mucuna has a limited effect on the rhizomes of Imperata (Akobundu et al., 2000), which suggests that eradication of Imperata may require a more integrated approach

490

A. GRAVES, R. MATTHEWS AND K. WALDIE

using cover crops in combination with other management techniques such as tillage or herbicides. The use of cover crops (traditional or new) may also sometimes encourage the introduction of pests, which may hamper the growth of the main crop or cause harm to the farmers themselves. In Bolivia, Pound et al. (1999) noted the increased presence of rats, snakes, and red spider mites with the use of Pueraria phaseoloides and Arachis pintoi as cover crops, while in West Africa, one of the constraints to adoption of mucuna cover crops was the fear possessed by farmers for the presence of snakes within the canopy (Galiba et al., 1998). Another concern with the use of green manures in ﬂooded systems such as rice agriculture, which is receiving increased attention, is its inﬂuence on methane emissions into the atmosphere and subsequent contribution to global warming (e.g., Matthews et al., 2000).

D. BIOMASS TRANSFER TECHNIQUES In an effort to relocate nutrients from forests to agricultural land, tropical farmers have traditionally used a variety of biomass transfer techniques (e.g., Young, 1987; Nyathi and Campbell, 1993). In most cases, this has involved the use of naturally occurring biomass (i.e., tree or grass material), and rarely biomass that has been speciﬁcally planted for that purpose. Recently, however, the attention of researchers has focused on transfer of biomass from deliberately planted “biomass banks” of species such as Tithonia diversifolia (ICRAF, 1997; Gachengo et al., 1999; Jama et al., 2000), Gliricidia sepium (Rao and Mathuva, 2000), Calliandra calothyrsus, and L. leucocephala (Mugendi et al., 1999) as a means of providing nutrients for crop growth, and organic material for physical improvement of the soil. The use of so-called cut-and-carry grasses is another technique where biomass is harvested and transported, in this case speciﬁcally to provide fodder for animals (e.g., Tanner et al., 1993; Stur et al., 2002). While similar in principle to alley cropping in that plant biomass is cut and incorporated into the soil to release nutrients for crops and to help improve SOM levels, one of the potential advantages of biomass banks is that direct competition between the main crop and that used to supply the biomass is minimised, if not eliminated altogether. The evidence so far suggests that biomass transfer techniques can help to increase soil fertility and sustain or increase crop yields (Mugendi et al., 1999; Rao and Mathuva, 2000). However, for the technique to be successful, the quality of this biomass needs to be high (Snapp et al., 1998), and very large amounts of biomass are required to supply “ideal” quantities of nutrients to crops (Gachengo et al., 1999). Labour for the collection, transportation, and incorporation of the OM into the soil must also be plentiful (Snapp et al., 1998; Gachengo et al., 1999; Jama et al., 2000). For example, in the

LOW EXTERNAL INPUT TECHNOLOGIES

491

subhumid highlands of Kenya, biomass transfer of Caliandra calothyrsus and Leucaena leucocephela prunings was found to increase maize yields compared to both unfertilised and fertilised sole maize (Mugendi et al., 1999). Similarly, Rao (2000) found that transfer of biomass from a Gliricidia stand increased maize yields on an equivalent area by 27%, but the high labour cost for harvesting, transfer, and application of prunings, made the technique uneconomical. These issues are discussed in more detail later in this review. In Vietnam, Hoang Fagerstro¨m et al. (2001) found that Tephrosia grown in a “biomass bank” and used as mulch on rice ﬁelds increased production over a four-year period by 55% compared to the monocropped control, if the extra land used for growing the Tephrosia was not included. This increase dropped to 29% if the total land area was taken into account. The use of the shrub Tithonia diversifolia (Mexican sunﬂower) as a source of P has also received some interest recently (e.g., Jama et al., 2000). Tithonia appears to have the ability to extract relatively high quantities of P from the soil giving it a high leaf P content (. 0.25%), and although it is a non-leguminous species, it also has a relatively high biomass N content of about 3.4% (Gachengo et al., 1998), above the level required to prevent net immobilisation of N (Palm et al., 1997b). In addition, its leaf lignin and polyphenol contents are less than 15 and 4%, respectively, so that it decays relatively easily. As it grows abundantly in the wild in many areas, and indeed, is often cultivated as a farm boundary hedge in Africa and Asia, it has been suggested for use as a biomass transfer technique (ICRAF, 1997; Jama et al., 2000). Experimental evidence suggests that addition of N and P through the application of Tithonia biomass may increase yields more than the use of equivalent quantities of mineral N and P (Jama et al., 2000). This has been attributed to the presence of K, Ca, and Mg in the biomass which might ameliorate deﬁciencies of these nutrients in the soil, and possibly also because of an improvement to soil physical characteristics. In addition to providing nutrients, Tithonia has been shown to reduce P sorption and increase soil microbial biomass (Jama et al., 2000). Various techniques have been found to increase Tithonia biomass production and therefore the quantity of nutrients extracted from the soil and made available for transfer elsewhere. The use of woody cuttings for propagation instead of soft stem cuttings, for example, was found to increase its biomass production, as was the application of mineral P fertiliser (Jama et al., 2000). However, its ability to extract P appears to be less if it is grown as a stand rather than as a hedge (George et al., 2001), suggesting that it is taking up P from a wider area than just that occupied by the hedge. It may, therefore, not be so effective if grown as a biomass bank. The P content of the soil also inﬂuences the concentration in the leaves, and continual P-mining through removal of Tithonia prunings is likely to eventually reduce the quantity and quality of the biomass to the level below which immobilisation of P is likely to occur (, 2.4 g P kg21). Moreover, it needs to be remembered that only the importation of Tithonia biomass will result in a net

492

A. GRAVES, R. MATTHEWS AND K. WALDIE

increase of P on the farm — that taken from boundary hedges will merely redistribute it within the farm. Most literature on established biomass banks appears to be research oriented in nature, and little evidence exists of their deliberate and successful introduction onto farm land, probably for the reasons mentioned above. As with many other LEIA technologies, if biomass transfer techniques are to be used, considerable areas of land will be needed to grow sufﬁcient quantities of plant material, which is clearly a limitation if land is scarce and farms are small. Consequently, it is unlikely that biomass banks for the sole purpose of soil fertility enhancement are likely to be widely adopted by farmers. There is also the disadvantage that there is a relatively long time-lag for beneﬁts on soil fertility of such techniques to accrue (Snapp et al., 1998). Farmers may be interested in biomass banks if they cannot effectively use all their land for cultivation. However, in such cases, it is more likely that they may opt to fallow this land to regenerate soil fertility. Biomass banks could be established on strategically located common land, which would be especially valuable for poor farmers, but for this to be workable, it would be necessary for a system of access agreements to be developed. The question of who was responsible for the establishment and maintenance of such stands would also need consideration, particularly as there are costs involved in purchase of seeds and seedlings, planting, and the care required during the initial establishment (i.e., weeding, etc.). Similarly, the sustainability of the system, with constant removal of nutrients in the biomass, would need to be addressed. Snapp et al. (1998) have noted that even large naturally occurring woodlands, such as the miombo woodlands in southern Africa, cannot be indeﬁnitely mined for nutrients. In most cases, it is unlikely that biomass transfer techniques will be capable of supplying the full fertility needs of a farm, and as with other LEIA fertility enhancing techniques, it may be best to see them as a component of an integrated nutrient management (INM) system involving external supplies of inorganic nutrients. It may be that there are speciﬁc niche roles which will make them useful on small areas within a single farm (e.g., for home gardens), or on degraded common land, although it is more likely in this case that they will fulﬁl other important needs, such as the provision of fuel-wood and/or fodder. Here they would move closer to the role played by natural forests, in which case, they may help relieve some of the pressure on the latter. However, there is a greater likelihood of adoption of biomass transfer techniques where there is some immediate beneﬁt to be obtained by the farmer. The use of cut-and-carry grasses as animal fodder, for example, has the advantage that some animal products, including milk, meat and draught power, are available almost immediately. Improvements to soil fertility through the application of manure and urine may be a secondary result. Farmers are usually well aware of the beneﬁcial effects of manure on soil fertility — there is some evidence that they feed their livestock much more than required for optimal live-weight gain, in

LOW EXTERNAL INPUT TECHNOLOGIES

493

order to provide manure for arable crop production (Tanner et al., 1993). Nevertheless, the essential difference of “processing” plant biomass through animals ﬁrst to gain immediate beneﬁts (e.g., meat, milk, and draft power), rather than using the biomass directly to improve soil fertility, is likely to be a determining factor of whether biomass banks are adopted by farmers or not. Of course, the intensiﬁcation of agriculture with cut-and-carry grasses and/or fodder banks is most likely to occur where animals are already a major component of the agricultural system and where satisfactory and alternative feeding strategies do not already exist (Reynolds et al., 1991; Rachmat et al., 1992). In some areas, the importance of cut-and-carry and zero-grazing techniques may increase as population increases and less and less land is available for free-grazing on communal land (Murwira et al., 1995), or where access to naturally occurring vegetation is limited, either because it does not exist or because access to it is blocked. However, in either case, sufﬁcient land still needs to be found somewhere to grow the biomass (Gashaw et al., 1991). As with other biomass transfer technologies, the evidence indicates that the amount of labour required for cut-and-carry techniques for fodder provision is often a disincentive to adoption (Mogaka, 1993; Wandera et al., 1993; Sanchez and Rosales Mendez, 1999). Cut-and-carry grasses may also not supply the full fodder requirements of livestock, necessitating supplementary feeding. Capital will be needed to pay for this, and for the inevitable veterinary fees associated with keeping livestock healthy (Mogaka, 1993). In addition, the decline of productivity of the fodder banks, as nutrients are removed may require investment in fertilisers to maintain productivity (Wandera et al., 1993). Finding suitable grasses, as well as issues related to land tenure are other important considerations. On the whole, the deliberate establishment and maintenance of fodder banks and cut-and-carry systems is probably most likely where farmers have some surplus capital and labour, but where land scarcity has resulted in limited access to natural vegetation. The ownership of cattle and the establishment of biomass stands involve costs that very poor farmers are unlikely to be able to meet.

E. COMPOST Compost is the aerobic, thermophilic decomposition of organic wastes to a relatively stable humus. Although it makes use of the same decomposition processes occurring naturally, the aim in compost-making is to control the conditions to a level that allows faster decomposition. The biophysical conditions that are required for effective composting are generally those that are required by the micro-organisms at various stages of the composting process, i.e., good moisture levels, moderate temperatures, mixed quality organic matter, and a fairly neutral pH range.

494

A. GRAVES, R. MATTHEWS AND K. WALDIE

Composting is not a new technique for the improvement of soil fertility and structure, and tropical farmers have been aware for centuries of its impact on crop yields, soil structure and fertility, crop growth and vigour (Diop, 1999; Onduru et al., 1999). For example, Oue´draogo et al. (2001) observed that yields of sorghum in Burkino Faso could be tripled by the application of 10 t ha21 of compost. Another beneﬁt noted is the reduced need for capital inputs (Onduru et al., 1999), although some capital may be necessary for farmers to adopt the technology (Girish and Chandrashekar, 2000; Slingerland and Stork, 2000). Compared to techniques such as alley cropping, composting in relation to subsistence agriculture seems to have received little attention by researchers. Much of the literature available tends to be written from a purely technical standpoint, and is often from the perspective of agriculture in developed countries. Relatively few studies have considered the on-farm issues of using compost, although the few that do give a good indication of the constraints involved. The major problem associated with the use of compost is the high labour requirements (Onduru et al., 1999). In particular, female-headed households can have considerable difﬁculty in undertaking some of the heavier tasks involved in composting, such as preparing compost pits (Diop, 1999). In Uganda, Briggs and Twomlow (2002) found that the poorest households did not make compost at all because of the labour and time requirements. Transportation of biomass and compost is also problematic (Apiradee, 1988; Adeoye et al., 1996). Also, like the other LEIA techniques already discussed, large quantities of biomass are required, and questions arise as to where farmers can obtain this (Onduru et al., 1999; Oue´draogo et al., 2001). This is particularly relevant where there are competing demands for such resources, for example as mulch, fuel or fodder (Drechsel and Reck, 1998), and where land to produce the biomass is scarce. Resource-poor farmers may have problems providing land for processing of “ideal” quantities of biomass, although this is not generally cited as a limitation, probably because fairly small quantities of compost are usually produced. On the other hand, some systems seem capable of producing quite large quantities; Briggs and Twomlow (2002) found that smallholder households in Uganda produced 40 kg of fresh organic waste per day, or about 9.2 t DM year21, 25% of which was used to make compost, with the rest either fed directly to livestock or applied directly to household plots. Composting may sometimes be constrained by lack of water (Apiradee, 1988; Diop, 1999), which is needed to aid decomposition, and by lack of biostarter (Apiradee, 1988), although it appears that animal manure and inorganic fertiliser may be used. Lack of tools for composting may also be a problem in some areas (Oue´draogo et al., 2001). An example of the use of compost in subsistence agriculture is provided by the work of the Rodale Institute, which has been working since 1987 with farmers in the Peanut Basin region of Senegal to help offset the rising cost of fertiliser due to removal of subsidies (Diop, 1999). Composting is not a new technique in

LOW EXTERNAL INPUT TECHNOLOGIES

495

Senegal, so research has focused on attempting to improve existing techniques involved in its production and use. Farmers are encouraged to collect crop residues for compost making rather than burning them, and to incorporate the resulting compost in the soil rather than leaving it lying on the surface. Pit composts are also being developed — these are about 1 m deep, lined with cement and bricks, and covered to prevent contamination from wind-blown sand. In the simplest and most efﬁcient approach, which does not require turning, rainfall in the wet season is used as a water source, with compost being produced within 18 months. In a more labour-demanding approach during the dry season, water is added manually when the compost humidity drops below a certain level. The compost must be turned every 15 days, but is ready within 45 days. For larger-scale compost making with this method, three pits are used with compost being transferred from one pit to the next every 15 days. Yields of both groundnut and millet, the traditional crops, have been approximately tripled through the application of 2 t ha21 of compost (Diop, 1999), although the area of crop involved in this case is not stated. Another example is the Zai system, traditionally used by farmers in Burkina Faso in times of drought (Roose et al., 1999). The system involves digging holes (20,000– 25,000 ha21), typically 30 cm wide and 20 cm deep, ﬁlling them with compost, and planting seeds of sorghum, millet, and cowpea into them. The compost is made from farmyard manure, plant residues, garbage, and rock phosphate — a natural product from mines in Burkina Faso. Crop yields can be more than 10-fold higher using the system than otherwise (from 150 to 1700 kg ha21), and the holes can be reused for 3 years. Trees may also sprout spontaneously in the holes. As composting is labour intensive, it is probably most appropriately used close to the homestead, on speciﬁc crops (e.g., Briggs and Twomlow, 2002). Evidence also suggests that progress may be made by improving the technical knowledge of farmers, so that composting practice is improved (e.g., Sutihar, 1984; Adeoye et al., 1996; Wakle et al., 1999). For example, the quality of compost can be greatly enhanced by mixing it with a combination of inorganic chemicals (van den Berghe et al., 1994), or by combining it with manure (Onduru et al., 1999). During processing, protecting it from heat and direct light may reduce volatilisation of the nutrients, while protecting it from rain may prevent similar losses by leaching (Diop, 1999). It seems unlikely that composting would be a new technique in many areas, unless a rapid transition from an abundance of land to scarcity is occurring. Progress is more likely to be made by determining whether composting practices can be improved within the socio-economic constraints of such farmers, such as with the Rodale Institute work. Recommending the use of animal manure in compost, for example, would not be appropriate if animal manure is the only source of fuel. Thus, as with other OM management techniques, composting is likely to provide only a partial solution to the problem of decline in soil fertility and soil

496

A. GRAVES, R. MATTHEWS AND K. WALDIE

structure in subsistence agriculture, but should be an important component in the basket of options that a farmer could use. Since composting is very labour intensive, it will probably continue to be most used close to the homestead, or on small areas of high-value crops. Extending the use of composting to large or distant ﬁelds may require unrealistic levels of labour for most farmers.

F. ANIMAL MANURE As with several of the other LEIA technologies, the beneﬁcial impacts of applying animal manure to land are well known to many tropical farmers and are also well documented in the scientiﬁc literature (e.g., Prudencio, 1993). As already mentioned in relation to the Machobane System, the use of animal manure as a source of crop nutrients is often a nutrient-harvesting technique in which nutrients are gathered through grazing of a relatively large area and concentrated on a smaller area where crops are grown. Even where animals are stall-fed, the nutrients they consume must be brought to them from elsewhere, either as collected fodder or as purchased concentrates. The rumen provides an ideal environment for the decomposition of organic matter, and is a way of improving the rate of decomposition in suboptimal environments, such as those with low temperatures or dry conditions. Livestock in developing countries generally feed on low-quality grasses and/or crop residues, which have signiﬁcant fractions of indigestible materials such as lignin and cellulose – lignin complexes, as well as low N contents (Leng, 1990). Resource-poor farmers often cannot afford to purchase supplements, and instead use tree and forage legumes to improve poor-quality diets (Coppock and Reed, 1992). The quality of a diet can inﬂuence feed intake, feed digestibility, and the partitioning of N between the urine and faeces (e.g., Somda et al., 1995). This latter characteristic is important, as a large proportion of the N excreted in the form of urine can be lost from the system through volatilisation and leaching, whereas that in faeces can be collected, stored and used to beneﬁt crop growth (Delve et al., 2001). The manure of animals which are fed highly digestible diets is more susceptible to N losses than that of animals fed a greater amount of roughage (Powell and Williams, 1993), while feeding on browse species has been found to shift the balance of excreted N away from the urine more towards the faeces (e.g., Reed, 1986). For example, Coppock and Reed (1992) found that animals fed Acacia tortilis pods excreted only half as much N in urine as those fed cowpea hay. Similarly, inclusion in the diet of Flemingia congesta, which has a high tannin content, was found to reduce N excreted in the urine by up to 35% (Fassler and Lascano, 1995). Hypothetically, a farmer with access to organic material is faced with a choice of whether to apply it directly to a crop as an organic fertiliser, or to use it to feed livestock ﬁrst with the manure produced then being applied to the crop.

LOW EXTERNAL INPUT TECHNOLOGIES

497

The optimal choice depends on the quality of the biomass available. It is generally accepted that material with a C/N ratio of more than 20 kg C (kg N)21 results in immediate immobilisation of N by microbial biomass when incorporated into the soil (Senesi, 1989) which is likely to reduce crop yields by reducing the amount of mineral N available. Passing this type of material (e.g., cereal straws) through an animal rumen ﬁrst can result in a large reduction in its C/N ratio, thereby reducing the amount of immobilisation. For example, Delve et al. (2001) found that cattle faeces from a diet of barley straw had a C/N ratio of 27 kg C (kg N)21 compared to that of 86 kg C (kg N)21 of the plant material only; consequently on incorporation, the faeces had a faster mineralisation rate and showed less net N immobilisation than did the fresh barley straw alone. Crop uptake of N was also greater from incorporated faeces than from the fresh straw. On the other hand, incorporation of high-quality plant material usually results in a higher recovery of N than of faeces derived from the same plant materials (e.g., Catchpoole and Blair, 1990). Delve et al. (2001) found that leaves of Calliandra calothyrsus with a C/N ratio of 13 kg C (kg N)21 resulted in steady net N mineralisation, but that manure from a diet containing 30% Calliandra calothyrsus resulted in net immobilisation between 3 and 16 weeks. The results from these studies indicate that the effect of passing material with C/N ratios ranging from 13 to 86 kg C (kg N)21 through an animal rumen is to make the C/N ratio converge to a value between 20 and 27 kg C (kg N)21. Thus, from a nitrogen supply point of view, it would seem better to apply high-quality organic matter (C/N ratio , 20 kg C (kg N)21) directly to the soil, but to pass low-quality biomass through an animal ﬁrst, and then apply the manure to the soil. Of course, this does not take into account other beneﬁts such as a better supply of milk and meat that might come from feeding animals high-quality fodder. The beneﬁcial effects of animal manures on crop yields, when applied in sufﬁcient quantities, are well documented. Not only can they enhance immediate crop yields (Selvarajan and Krishnamoorthy, 1990; Ali, 1996; Drechsel and Reck, 1998), they can also provide residual beneﬁts for following years (Singh and Desai, 1991). However, Giller et al. (1997) note that, although animal manures are of major importance in nutrient cycling, they are generally of poor quality as a supply of plant nutrients. In Rwanda, for example, Roose et al. (1997) found that the application of manure at 10 t ha21 had no effect on maize yields. Even the effect of applying manure at the rate of 20 t ha21 was limited to a single season, as the yield of the second crop (sorghum), showed no signiﬁcant difference compared to the other treatments, which was ascribed to the low amount of P supplied in the manure. The beneﬁts to the soil from applying manure, therefore, may be more related to improvements in physical characteristics rather than the provision of nutrients, especially in the quantities that farmers can supply (Singh and Desai, 1991). It is important to remember that the use of manure is related to its role as part of a larger system. In many subsistence-farming systems, there is a close

498

A. GRAVES, R. MATTHEWS AND K. WALDIE

interdependence between crop production, animal production, and possibly neighbouring forests and/or rangelands. In Nepal, for example, animals are grazed in the forest, crop residues and forage collected from the forest are used to feed housed animals, animal manure is applied to cropped areas, crop residues and forest litter are used for animal bedding, and animals are used to provide tillage and transport (Pilbeam et al., 1999b). In certain areas of western Africa, some arable farmers make arrangements with itinerant herdsmen to corral livestock on their land (Waldie, 1990; Enyong et al., 1999). Farmers may also move their homesteads from place to place so that crops can be grown on that land to beneﬁt from the manure left over by livestock (Ruthenberg, 1980). In upland Java, livestock may be fed with far greater amounts of biomass than is needed for optimal live-weight gain, the rationale being the production of manure-compost that is collected for intensive upland agriculture (Tanner et al., 1993). In smallholder farms in Uganda, Briggs and Twomlow (2002) calculated that an average of 4.6 t of manure was produced annually from a household with 2.2 ha of land — this was mainly from goats grazing on communal land during the day, but tethered near the household at night — a ﬁgure that could be increased if better manure management was practised. In many cases, there may be a high opportunity cost of using manure as a fertiliser, and farmers may often value it more for uses other than soil fertility maintenance. Beneﬁts obtained from manure include the provision of material for plastering and building, and fuel for heating and cooking (Jeffery et al., 1989; Murwira et al., 1995). Provision of milk for human consumption from keeping livestock has already been mentioned. Many of these activities (e.g., the production of fuel cakes or milk) have direct economic value in themselves (Jeffery et al., 1989). As discussed for the Machobane system, competing uses of manure inevitably reduce the amount available for soil chemical and physical improvement. If manure is to be extensively used to enhance soil fertility, it will need to be culturally acceptable to farmers, which is most likely to occur where livestock are an integral part of the farming system already. This is more likely to be in areas where population pressure is higher, labour availability is higher, and land is scarce, such as in Nepal (Murwira et al., 1995). Where land is not scarce, farmers may ﬁnd it more practical to improve soil fertility and structure through natural or improved fallow, although they may still use manure on plots or kitchen gardens close to the homestead if they own poultry, smallstock, or cattle. Both the production and use of manure are labour intensive (Enyong et al., 1999). Households without adequate labour, or the means to procure it (e.g., in communal work groups or through purchase), may only be able to use small amounts of manure. In general, investment in manure to improve soil increases where an enabling environment for agricultural production is provided. For example, in areas where there are market outlets and effective extension networks providing technical guidelines on good manure practice, farmers may be

LOW EXTERNAL INPUT TECHNOLOGIES

499

encouraged to increase productivity, stimulating the use of manure as one of the various options available for increasing crop production (Enyong et al., 1999). However, the fragmentation of ﬁelds occurring in many developing countries may also make it more difﬁcult to transport manure, reducing farmers’ willingness to apply it to ﬁelds at a distance from the homestead (Enyong et al., 1999). In general, as with other LEIA techniques, the use of manure should be seen as a component in an INM system. Indeed, Sherchan et al. (1999) found that the highest yields of maize and wheat in Nepal were obtained from a combination of manure and inorganic fertiliser. Alternatively, it could be used in certain niches, say around the homestead with specialised high-value crops, where the distance to transport it is not a major constraint.

G. IMPROVED FALLOWS Shifting cultivators have traditionally alternated periods of crop production with periods of fallow in order to restore soil fertility and suppress weeds. In some cases, the cropping period only lasts 4– 5 years while the fallow period may be as long as 30 years (e.g., Matthews et al., 1992a), during which time the land is usually unproductive in terms of generating a livelihood. In recent years, researchers have focused on ways to shorten this period, and/or to make some use of the land while it is fallow. Thus, an “accelerated fallow” is where speciﬁc fastgrowing leguminous or non-leguminous trees, shrubs, legumes, and other plants are used to improve soil fertility faster than would occur otherwise, while an “enriched fallow” is where trees or shrubs of economic value are planted into the fallow so that the farmer can derive some income from them while the land is regenerating (Garrity and Lai, 2000). Szott et al. (1999) have reviewed the processes involved in fallowing from an ecological perspective. Under a natural fallow, the time needed to restore N to original levels can be less than two years, but can be up to 20 years for other nutrients such as Ca. The rapid restoration of N is likely due to N ﬁxation (up to 300 kg N ha – 1 year21 on high base status soils, (Giller, 2001)) and retention in the growing biomass. Most evaluations of the performance of improved fallows have compared crop yields after the fallow period with those after a natural fallow of the same duration. In general, it has been found that short-term fallows (i.e., , 3 years) growing leguminous trees or shrubs can increase crop yields compared to the natural fallow control (Szott et al., 1999). Nitrogen ﬁxation is the main reason for this, but the retrieval of subsoil N by deep-rooted species from below the crop root depth can also be important, and may be of the same magnitude as N-ﬁxation, particularly if nitrate has accumulated in the subsoil during the cropping phase before the fallow (e.g., Mekonnen et al., 1997). Improvement in soil structure may also be important — Salako et al. (2001), for example, found that soil physical properties improved under fallows of

500

A. GRAVES, R. MATTHEWS AND K. WALDIE

Pueraria phaseoloides, Acacia auriculiformis, Leucaena leucocephela, Senna siamea, and Acacia leptocarpa, as well as under a natural fallow. Grass fallows do not usually increase crop yields to the same degree, if at all, probably due to immobilisation of N (Barrios et al., 1997) although in some cases, they do improve soil structure (Padwick, 1983). It is noteworthy that crop yields were signiﬁcantly lower following fallows of Ageratum conizoides, Tithonia diversifolia, Sida rhombifolia, or Rottboellia cochichinensis than following a natural fallow, despite having more N and P in the fallow biomass (Szott et al., 1999). On soils that are severely deﬁcient in P, short-duration improved fallows have been shown to need additions of fertiliser P to maintain crop yields (Jama et al., 1998). Beneﬁts of short-duration fallows to crop yields are sometimes related to the amount of biomass accumulated during the fallow. Maroko et al. (1998), for example, reported no increase in yields of the second maize crop after a Sesbania fallow, which had accumulated 14 t ha21, but a signiﬁcant increase in crop yield with a 23 t ha21 Sesbania fallow. In most cases, there is little residual effect of the fallow biomass after two crops (Drechsel et al., 1996). Interestingly, the use of short-term cover crops such as Mucuna in fallows may be detrimental to long-term accumulation of nutrients and biomass by impeding the establishment of trees and shrubs (e.g., Szott and Palm, 1996). Kang (1997) has recommended that short fallows be used where the duration of the cropping period is at least half of the total cropping/fallow cycle. There may also be other beneﬁts derived from improved fallows. For example, Brodd and Osanius (2002) report reduced soil erosion under an improved fallow of Tephrosia candida in northern Vietnam compared with shifting cultivation. Similarly, Gallagher et al. (1999), in a review of literature, have shown that improved fallows of woody perennials and herbaceous cover crops could suppress weeds, particularly over a number of years, and might be an important component of integrated weed management (IWM) strategies. Perhaps the main advantage of accelerated and enriched fallow systems is that that they are modiﬁcations of an existing system, requiring only minor changes to existing farmer practice. Accelerated fallows can be seen as a natural progression from shifting cultivation and other long-fallow techniques, and may therefore be an appropriate choice where these are no longer sufﬁcient to maintain productivity due to population pressure (Sanchez, 1999). From the biophysical point of view, due to the deeper-rooting characteristics of the woody species usually used in accelerated fallows, nutrients from below the rooting depth of arable crops can be made available again, and the use of appropriate leguminous species can result in improved rates of addition of N to the system. Also, there is no direct competition for resources with main crops as with some of the other LEIA techniques such as intercropping or alley cropping (Sanchez, 1999). The downside, compared with continuous cropping systems at least, is that production is lost from the land set aside for fallow. Hoang Fagerstro¨m et al. (2001) found that the overall rice production from both a natural fallow and a Tephrosia fallow

LOW EXTERNAL INPUT TECHNOLOGIES

501

with 2-year crop/2-year fallow cycle were 18 and 6% less, respectively, than production from continuous cropping over the four-year period, despite annual yields in the latter being considerably lower than in the fallow treatments. Similar results were obtained by Swinkels et al. (1997) in western Kenya. Accelerated fallows are more likely to be used by farmers in areas where increases in the population density is starting to make the long periods required by natural fallow impracticable. At higher population densities, however, scarcity of land means that there is a high opportunity cost in putting land to fallow, and intensive continuous cultivation systems may dominate (Drechsel et al., 1996). Accelerated fallows are, therefore, most relevant in the intermediate stage between extensive and intensive land use (Franzel, 1999). However, if they are to be adopted more widely, farmers need to be aware that there is a problem to be addressed. This may be declining yields (Franzel, 1999) or fertility (Degrande and Duguma, 2000), or controlling weeds (Tarawali et al., 1999). Security of land tenure is also an important consideration, as farmers are unlikely to be willing to invest time and effort in establishing accelerated fallows if they are not the ones to receive the beneﬁts (Seif El Din and Raintree, 1987; Long and Nair, 1999; Tarawali et al., 1999). Institutional support, in the form of seed programmes and training of extension agents and farmers, has been found to be important in improving the adoption of accelerated fallow techniques (Franzel, 1999). Other important requirements may also be to provide adequate and/or improved germplasm (Place and Dewees, 1999). A major disadvantage of both natural and accelerated fallow systems is the length of time it takes for any ﬁnancial beneﬁts to accumulate (Grist et al., 1999), particularly with the latter, since natural fallows may provide resources that accelerated fallows do not. Kaya et al. (2000) concluded that improved fallows in Mali were not attractive to farmers if their sole purpose was soil fertility improvement. Enriched fallows address this problem to some extent, in that species that are able to provide some economic beneﬁt, such as fruit or nuts, are planted in preference to species that only improve soil fertility (Cairns and Garrity, 1999; Franzel, 1999; Sanchez, 1999). Other practical beneﬁts to farmers may include production of fodder (Kaya et al., 2000), honey (Franzel, 1999), ﬁrewood, or bean poles (Drechsel et al., 1996), or light timber for construction (Franzel, 1999). From an ecological viewpoint, Styger et al. (1999) have suggested that the judicial identiﬁcation, selection, and domestication of preferred forest fruit trees could be used as a means of preserving biodiversity in forest margin areas that are under pressure. Where such multipurpose tree (MPT) techniques are successful, farmers may be encouraged enough to develop them into permanent agroforestry systems (Cairns and Garrity, 1999), and, indeed, in several places many have done this already. Farmers in Benin, for example, plant oil-palm trees in a fallow of about 12– 15 years (Versteeg et al., 1998). This restores soil fertility, but also provides subsistence and cash income, even upon clearance of the trees when “palm wine”

502

A. GRAVES, R. MATTHEWS AND K. WALDIE

is produced. However, removal of harvestable products (and the nutrients they contain) lengthens the time taken for regeneration of soil fertility, and in extreme situations, may eventually mine the soil of nutrients (Franzel, 1999). These deﬁcits will have to be made up with other sources of nutrients. Pradeepmani (1988) has discussed some of the issues affecting farmers’ decisions to plant MPT species. These include having adequate land, time, labour, knowledge, and inputs, being able to protect trees properly, and success with tree survival. Security of tenure and access to markets are also important (Mahamoudou and Meritan, 1998; Hellin et al., 1999). A certain level of access to capital also appears to encourage adoption. Where these factors are not in place, farmers tend to increase the rate at which they discount future beneﬁts, making such techniques socio-economically unviable, and reducing both the “action time horizon” and the “planning time horizon” (Vosti and Witcover, 1996). Efforts to encourage planting of MPTs species through training visits, extension of effective methods for protecting trees (which is often expensive), and government land tax incentives, were also noted as important factors (Pradeepmani, 1988). Although the multipurpose nature of many trees may serve as “pull” factors, strong “push” factors can also operate at the localised scale. For example, shortages of agricultural labour, high cost of agricultural inputs, and shortages of power and water have all been reported to encourage farmers to plant MPTs (Dasthagir et al., 1996). In most cases, economic factors are more important than ecological factors in inﬂuencing farmers’ decisions to plant MPTs (Mahamoudou and Meritan, 1998). Certainly, in a study of the reasons for the adoption of MPTs on homestead land by farmers in Bangladesh, direct economic concerns were reported as being uppermost in the minds of farmers (Salam et al., 2000). Other factors in order of importance were the provision of fruit, ﬁrewood, and building materials for subsistence, emergency cash needs, maintenance of ecological balance, and protection from strong winds. Further analysis showed that tree planting increased with increases in the amount of land owned, the level of non-agricultural income, the market costs of fuel-wood, the male membership of the household, and the knowledge of extension activities. Compared to natural fallow systems, more labour is required for both accelerated and enriched fallows, primarily for planting of the fallow species and weeding to ensure their establishment (Drechsel et al., 1996; Grist et al., 1999). Hoang Fagerstro¨m et al. (2001) give values of 386 and 600 labour days ha21 over a 4 year period for a natural fallow and a Tephrosia fallow, respectively. There is a clear danger that this additional labour demand could clash with those for planting and management of other crops (Franzel, 1999). On the other hand, the labour demand of improved fallows is relatively ﬂexible, certainly compared to alternative systems such as alley cropping where the timeliness of pruning is very important. Some capital input is needed for improved fallows, mainly for the purchase of seeds or seedlings of the species to be planted. If farmers have insufﬁcient capital, they tend to use natural fallows for soil fertility improvement,

LOW EXTERNAL INPUT TECHNOLOGIES

503

whereas better-off farmers will tend to purchase inorganic fertilisers (Franzel, 1999). Improved fallows may therefore, be most appropriate for farmers at an intermediate level with some disposable income. Adoption of improved fallow techniques by farmers has been signiﬁcant in some regions. Citing various sources, Sanchez (1999) claims that there is largescale adoption of improved short-term fallows (i.e., , 5 years in duration) occurring in Central America, Brazil, Southeast Asia, East Africa, and southern Africa, with perhaps hundreds of thousands of farmers using the technique. The majority of species used are Sesbania, Leucaena, Mucuna, Centrosema, Pueraria, Crotalaria, Cajanus, Indigofera, and Mimosa. An example is the “Qezungual System,” which has been indigenously developed in western Honduras (Hellin et al., 1999). The technique can be described as a triple-level agroforestry system, combining crops such as maize, sorghum and beans, numerous pollarded trees and shrubs (about 1.5 m high) and high-value trees, particularly fruit trees and timber trees. The system has generally developed on land that has been under secondary vegetation, or less commonly, on land that is under primary forest. A main characteristic of the technique is the reduction of labour requirements for the establishment of the valuable fruit and timber trees species, as these are simply selected when the land is cleared for agriculture. Other less valuable trees and shrubs are pollarded and the land is prepared for cultivation. This also reduces the time required for beneﬁts to accrue to the farmer, and may reduce the need for inputs requiring capital (for example, seedlings) and labour (maintenance of vulnerable seedlings). Competition between perennial plants and food crops is greatly reduced by the pollarding, and can be manipulated by gradual clearing of perennial plants, if necessary. Natural regeneration is managed by selecting speciﬁc trees for production, whilst others are pollarded. Within the pollarded areas, crops may be rotated and areas left fallow to control pests. From discussions with farmers, there appear to be several beneﬁts from the use of this technique. Pollarded plots have higher agricultural production and can also be cultivated for a longer time than unpollarded plots. Soil moisture is conserved because of reduced soil evaporation due to the mulch from the pollards, and perhaps because the pollards improve soil physical structure and allow increases in WHC. The system provides multiple beneﬁts for subsistence and cash income (fruit, food crops, timber and ﬁrewood). Disadvantages are that moisture levels can become too high in times of high rainfall, leading to fungal attacks on crops. Birds may also be attracted by the trees and pollards, causing reduction of food crop yields. Animal or mechanised traction cannot be used to cultivate the land due to the random presence of the trees and pollards. Various factors have encouraged farmers to adopt the Qezungual system. Land scarcity is important (most farms are about 2.5 ha) as, where land is abundant, farmers generally continue to use natural fallows. Absence of ﬁre as a management tool is also important, otherwise the trees are destroyed. Lack of

504

A. GRAVES, R. MATTHEWS AND K. WALDIE

animal or mechanised traction is another factor, as the pattern of trees and pollards is fairly random, although with selective thinning, it might also be possible to develop pathways for animal and mechanical traction tillage. Possibly the most important factor is that it addresses a problem that the farmers ﬁnd important — soil moisture. In dissemination of the use of the system, it is promoted as a soil moisture technology, rather than an erosion control technology, although erosion is deﬁnitely a problem in some areas.

III. GENERIC ISSUES A. SOIL FERTILITY MANAGEMENT 1.

Biomass Quantity

Relatively little quantitative information exists on the ideal level of SOM. Brady (1990) suggests that it should be around 5%, corresponding to a soil organic carbon fraction (SOC) of about 3%. In sandy southern African soils, 1– 1.5% SOC has been recommended as the long-term agroecologically viable minimum (Araki, 1993). Similarly, research from western African countries suggests that when SOC levels fall below 1%, severe physical soil degradation can be expected to take place (Pieri, 1995). The threshold level of SOC required to prevent severe physical degradation of a soil is also related to the soil’s texture — for soils with a low-sand content, an SOC content of 0.9% may be adequate, but for sandy soils, this may have to be as high as 1.5% (Araki, 1993). Thus, although there appears to be some variation as to an “ideal” level, somewhere between 1 and 3% SOC is what many researchers would consider necessary. Young (1989) has estimated indicative quantities of plant biomass required for maintenance of SOC in soils in various agroecological zones, based on typical topsoil organic carbon levels and approximate oxidation and erosion losses. Above-ground plant biomass required inputs were estimated to be about 8.4 t DM ha21 year21 for humid regions, 4.2 t DM ha21 year21 for subhumid regions, and about 2.1 t DM ha21 year21 for semi-arid areas, with below-ground biomass inputs about 70% of these ﬁgures. These are broad estimates and other evidence suggests different quantities are required. For example, in southern Africa, Snapp (1998) estimated that annual applications of about 10 t DM ha21 year21 of high-quality plant biomass (see below for discussion on biomass quality), or 7 t DM ha21 year21 of low-quality residue, were necessary to maintain a minimum level of 1% SOC in sandy loam soils in the subhumid tropics, assuming a decomposition rate of 0.05 year21 (Janssen, 1993). Thus, assuming a linear relationship between SOC levels and biomass inputs, the “ideal” level of 3% SOC mentioned previously may require as much as 30 t DM ha21 year21 in

LOW EXTERNAL INPUT TECHNOLOGIES

505

similar conditions. Even if the relationship is not linear, the data suggest that on the whole, large amounts of biomass are required to maintain the physical condition of the soil at a level that can support continuous and sustained crop production. The supply of N and P to meet crop nutrient requirements also requires large quantities of biomass, not only because of the relatively low concentration of nutrients in biomass, but also because the fraction actually recovered by crops is generally low for organic inputs due to losses to the atmosphere, surface waters, or ground water (Gregory et al., 2002), although this can vary substantially depending on biophysical conditions. Giller and Cadisch (1995) have suggested a value of 20% for most organic inputs. Gachengo et al. (1999) found that the recovery of N in Tithonia prunings was about 25% by a ﬁrst maize crop in western Kenya. Other evidence suggests that N recovery by the ﬁrst crop after OM incorporation is generally between 9 and 28% of the N supplied in the OM, but may be as low as 2 –10% in a second crop (Snapp et al., 1998). However, the rate of N recovery can be improved by the addition of limiting nutrients to ensure that the growth of the main crop is not limited. For example, Snapp et al. (1998) found that the recovery fraction of organic N increased from 25 to 46% in the ﬁrst year when 25 kg P ha21 was applied to correct the P deﬁciency at the site. To supply adequate N for a typical 5 t ha21 maize crop removing about 100 kg N ha21, therefore, more than 14 t DM ha21 of biomass would be required (assuming a 20% recovery rate and 3.5% leaf biomass N content (Jama et al., 2000)). For animal manure with 1.5% N content (Lekasi et al., 1998), 33 t DM ha21 would be required. These ﬁgures would equate to 60 and 67 t ha21 of fresh material, respectively, assuming a 20% dry matter content for fresh leaf biomass and a 40% dry matter content for fresh manure. For P, Jama et al. (2000) calculated that 5 t DM ha21 would be required to supply the 18 kg P ha21 needed to overcome moderate P deﬁciencies. With a 20% recovery rate, this is equivalent to an application of 25 t DM ha21 or 125 t ha21 fresh weight. In severely P-deﬁcient soils, even more would be required. The quantities required for the effective management of other soil nutrients is also large. For example, Jackson et al. (1999) found that to correct zinc deﬁciencies of soils near Wenchi in Ghana, about 20 t ha21 (dry or fresh weight not speciﬁed) of poultry manure was required. The amount of sheep or cattle manure required was estimated to be between 40 and 60 t ha21. If these levels of biomass are required to have an appreciable effect on SOC, the question arises as to how easily these quantities can be produced by smallholders. In the studies where alley cropping has been shown to beneﬁt crop yields, tree biomass production has been in the order of 6 –8 t ha21 year21, using L. leucocephala (Kang et al., 1985). For Flemingia congesta, Budelman (1988) recorded an annual dry matter production of 12.4 t ha21 year21 in the Ivory Coast, while Yamoah et al. (1986a) measured 16.9 t DM ha21 year21 for Flemingia congesta in Nigeria. However, in many cases, biomass production is not this high. In an alley cropping system in Costa Rica (humid), the net primary

506

A. GRAVES, R. MATTHEWS AND K. WALDIE

production of Calliandra calothyrus was about 4.4 t DM ha21 year21. Of this, about 2.8 t DM ha21 year21 was estimated to be leaf production (Baggio and Heuveldorp, 1988) and of possible use as a green manure. In Nepal, C. cajan was able to produce about 3– 4 t DM ha21 year21, Centrosema pubescens about 5 t DM ha21 year21, and Calopogonium mucunoides about 4 t DM ha21 year21 (Pande, 1997). In Malawi, Saka et al. (1995) showed that the leaf biomass production of three hedgerow species (Gliricidia sepium, Leuceana leucocephala, and Senna spectabilis) varied between 0.5 and 2 t DM ha21 year21, and did not affect the SOC level over a 1 year period. Also in Malawi, Kanyama-Phiri et al. (1997) found that Sesbania sesban produced about 2– 3 t DM ha21 year21 of high-quality leaf biomass. This was in addition to the fuel-wood produced during the 10 months of growth between January and October. In Zambia, although Flemingia congesta produced a maximum of 3 t DM ha21 year21 in one trial (Table I), the mean production of all the species was only 1.3 t DM ha21 year21 (Matthews et al., 1992a). In many cases, therefore, biomass production does not seem high enough to have any appreciable effect on SOC levels. Cover crops are likely to produce even less biomass annually due to their shorter duration of growth compared to woody perennials. Increasing SOC to “ideal” levels, therefore, will in most cases, necessitate the importation of additional amounts of organic material. The question is, therefore, where is this biomass to come from? If the farmer is to grow it, can it be produced in sufﬁcient quantities to have an appreciable effect

Table I Mean Annual Biomass Production (t ha21 y21) of Different Tree Species in Agroforestry Trials at Kasama, Northern Province, Zambia. (Developed from Matthews et al., 1992a,b) Trial no. D11 D21 D22

D31

D32

D33

Species Flemingia congesta Flemingia congesta Flemingia congesta Tephrosia vogelii Cassia spectabilis Calliandra calothyrsus Leucaena leucocephala Albizzia falcataria Flemingia congesta Gliricidia sepium Flemingia congesta Cassia spectabilis Sesbania sesban Flemingia congesta Cassia spectabilis Sesbania sesban

1987

2.46 0.67 0.94 0.83 0.54 0.72 1.05 0.56 0.86

1988

1989

1990

1.40 0.28 1.06 2.19

2.45 2.22 1.89 0.33

2.91 1.09 1.41

2.23

0.64 2.47

0.44 0.79 0.74 1.31 0.44 1.13 2.14 0.66

1.09 1.08 0.51 0.84 0.07 1.36 1.40 0.81

0.56 1.51 1.48 0.73 0.47 0.60 1.12 0.82 0.97

LOW EXTERNAL INPUT TECHNOLOGIES

507

on SOC levels? Assuming an annual biomass production of 2 t DM ha21 year21, 3 ha of land would be required to supply 1 ha of cropped land area with enough biomass to maintain the SOC level at just 1%. Signiﬁcantly more would be required to raise it to the 3% level indicated earlier. In the initial years of hedgerow intercrop systems, when farmers are most likely to reject or accept a new technique, tree biomass production is likely to be well below 2 t DM ha21 (e.g., Matthews et al., 1992a). This is clearly insufﬁcient to maintain SOC at even the 1% level. From the farmer’s perspective, if this biomass is to be grown on-farm, the area available for food crop production must necessarily be less. If it is to be supplied from outside the farm, the transport of such large quantities of organic material would require considerable labour and physical effort (Rao and Mathuva, 2000; Hoang Fagerstro¨m et al., 2001). Furthermore, labour will be required for incorporation of the biomass into the soil. In many situations, therefore, it may simply not be possible to use plant biomass to increase SOC within farmer constraints. Supplying adequate amounts of organic matter through animal manure is also difﬁcult. The dry matter quantities required for soil physical improvement are similar to the amounts required when using plant biomass (Euroconsult, 1989). Pilbeam et al. (1999b), in deriving an N balance for a hypothetical household with 1 ha of agricultural land in the mid-hills of Nepal, estimated that the total feed requirement for buffalo, assuming a live-weight of 450 kg, was about 2.6 t DM year21. For cattle, assuming a live-weight of 250 kg, feed requirements were estimated to be 1.8 t DM year21. Thus, to balance the SOM losses given by Young (1989), about 3 buffaloes or 5 cows would need to be kept for every 1 ha in humid regions and one buffalo or 1.5 cows per hectare in semi-arid regions, assuming that nearly 100% of the consumed biomass passes through the animals. For comparison, Nandwa and Bekunda (1998) calculated that between 2 and 8 cattle would be needed to supply enough nutrients for a 2—3 t ha21 maize crop. For a typical farm in the mid-hills of Nepal, the analysis by Pilbeam et al. (1999b) suggests that about 2.5 t of animal feed comes from dry and green crop residues, presumably from on-farm sources. However, to maintain SOC levels of one hectare of land through the use of animal manure alone, this still leaves a requirement of about 6 t DM of fodder from off-farm sources. These biomass requirements are rough guides, but serve to show that the quantity of animal manure needed to maintain soil physical characteristics and nutrient levels effectively are substantial. In most cases, it is unlikely that resource-poor farmers would have access to on-farm sources of manure in sufﬁcient quantities to supply the total OM requirements of their cropped land, particularly as there are competing demands for its use, such as for building material or fuel (e.g., Nandwa and Bekunda, 1998). In such cases, reliable access to off-farm land for fodder collection will be a major requirement for the use of animal manure. In the Embu District of the Kenyan highlands, application rates of 5– 8 t FM ha21 (~2 –3 t DM ha21) are recommended to farmers

508

A. GRAVES, R. MATTHEWS AND K. WALDIE

(Lekasi et al., 1998), but average rates applied by farmers are often much higher at 11 t FM ha21 (~4.4 t DM ha21), and could even exceed 17 t FM ha21 (~7 t DM ha21). Despite these relatively high rates of application, many farmers felt that they would use even more if it were available. In general, the smaller the animal, the higher is the nutrient concentration in its manure. For example, poultry manure has much higher levels of N and P (up to 4.8 and 1.8%, respectively) than cattle manure (1.5% N and 0.14% P) (Reddy et al., 2000). The amount of poultry manure needed to supply the requisite amount of N and P to a crop would, therefore, be correspondingly lower, as would the labour required for transporting it. For example, Swift et al. (1994) estimated that 1 –2 tonnes of poultry manure would be required to fertilise a 2 tonne maize crop, compared to 7 tonnes of low-quality animal manure or 10 tonnes of straw. On the other hand, smaller animals produce lower quantities of manure per animal, and ﬁnding poultry manure in sufﬁcient quantities could prove difﬁcult, as it is unlikely that the numbers of poultry typically found on resource-poor farms would supply more than a few kilograms of manure annually. Overall, therefore, the quantity of biomass available to small farmers, either from plants or from animal manure, is likely to constrain the degree to which soil fertility can be maintained or even improved. Production of this biomass on farm must be weighed against the loss of any other use the land may be put to, particularly the growing of crops. Off-farm sources may be an option, as in Nepal, but this will not always be the case. However, an important effect of the addition of OM to the soil, and one that may be immediately appreciated by farmers, is an improvement in the workability of the soil, so that farming operations, particularly ploughing or hand hoeing, are eased. This is likely to be particularly important where continuous cultivation of land is already practised and SOC has decreased, and bulk density has increased as a result. For example, in on-farm trials in Ghana and Bolivia, farmers reported the beneﬁts in terms of the “softness,” “looseness” or “coolness” of soils after using OM from cover crops and animal manure (Kiff et al., 1999; Pound et al., 1999). The extent, to which these beneﬁts were a result of the incorporation of the OM in the soil, or the growth of cover crop, was not reported. However, what is important in that soil workability was a key factor by which farmers appreciated the effect of OM techniques in the short-term rather than in the long term. 2. Biomass Quality The “quality” of biomass is a function of its nutrient content and its resistance to breakdown. Biomass quality has two opposing effects, in that lower quality OM is likely to have a larger impact on SOM levels than high-quality material, which mineralises more rapidly. On the other hand, higher quality organic material

LOW EXTERNAL INPUT TECHNOLOGIES

509

contributes more to the nutrient status of the soil (N, P, K, and micronutrients), and is important for maintaining soil microbial activity and the soil buffering capacity. Successful OM management depends on ﬁnding the correct balance between these two effects. This applies to plant OM, whether green manure or crop residues, as well as to animal manure. Crop residues and other low-quality organic material, particularly if added in large quantities, may temporarily induce N or P deﬁciencies in the soil due to microbial immobilisation, thereby reducing crop yields. Palm et al. (1997b), for example, have shown that addition of OM containing less than about 0.25% P to the soil is likely to cause net immobilisation of P. Such deﬁciencies may have to be overcome through the use of inorganic fertilisers (Muriwara and Kirchmann, 1993). The ratio of carbon to nitrogen (i.e., the C/N ratio) in organic material is often used as a measure of its quality. More recently, the concentrations of lignin and polyphenols have also been found to be important, particularly the lignin/N, polyphenol/N, and (lignin þ polyphenol)/N ratios (Snapp et al., 1998). Highquality organic inputs are low in lignin and polyphenol and high in N (Palm et al., 1996), and as such decompose more quickly. It has been estimated that they release between 70 and 95% of their N within a season under tropical conditions (Giller and Cadisch, 1995). The quality of plant biomass is not constant, but varies with age and whether a plant is leguminous or non-leguminous. In general, young plant material has low C/N ratios ensuring that its nutrients will be released quickly when it is incorporated into the soil, while material from older plants (or plant organs) of the same species generally has a higher C/N ratio. Data on C, N, P, and K contents for a range of organic materials used in lowexternal input agriculture are available in the Organic Resources Database developed as part of the Tropical Soils Biology and Fertility Programme (TSBF) (Palm et al., 2001). The nutrient content of animal manure depends on the species (Table II), the diet of the animal, and on how the manure is collected, stored, and applied. Diet is particularly important in relation to the partitioning of N between the faeces and the urine (Snapp et al., 1998). High-quality diets (low in lignin and polyphenols) result in more N being excreted in the urine than in the faeces (Somda et al., 1995). N that is excreted in the urine is much more quickly volatilised, and urine is also more difﬁcult to collect. Animals fed with a tannin-rich diet tend to excrete more of the N in their faeces. However, recent results suggest that this kind of N is very resistant to mineralisation (Mafongoya et al., 1997a). The beneﬁcial 2 effects of N released from manure as NHþ 4 and NO3 appear to be directly after application. However, poor-quality manure has been found to result in prolonged periods of N immobilisation, and under these conditions, N availability has been increased with inorganic sources (Muriwara and Kirchmann, 1993). Table II shows that for manures from most domesticated animals, P content is above the threshold at which P immobilisation is likely to occur, but that N content is often below the critical limit required to prevent N immobilisation.

510

A. GRAVES, R. MATTHEWS AND K. WALDIE

Table II Mean Content of C, N, P and K in Manure from Various Domestic Animals in Muranga and Kiambu Districts, Kenya

Manure type

C (%)

N (%)

P (%)

K (%)

C/N ratio (kg C (kg N)21)

Cattle Cattle and compost Goat Pig Poultry (broilers) Poultry (local) Rabbit Sheep

35 25 32 40 41 22 33 33

1.4 1.3 1.5 2.0 2.4 1.2 1.6 1.5

0.60 0.44 0.40 1.19 1.60 0.91 0.40 0.33

0.59 0.36 0.53 0.49 0.41 0.26 0.50 0.44

26 21 22 21 17 19 20 22

(From Lekasi et al., 1998).

In the absence of inherently fertile soils and inorganic fertilisers, improved crop yields are usually achieved only with high-quality OM. Low- or even medium-quality residues have generally been unsatisfactory (Snapp et al., 1998). In Kenya, for example, Nandwa (1995) found that incorporation of maize stover (low-quality OM) reduced maize grain yield by between 3 and 30%. Maize stover has also been shown to reduce crop yields in Zimbabwe (Rodell et al., 1980; Muriwara and Kirchmann, 1993). In India, Goyal et al. (1992) found that a combination of wheat straw (low-quality OM) and urea reduced yields, while a combination of Sesbania green manure (high-quality OM) and urea increased yields compared with the application of urea alone. The reduction in yields with low-quality OM is generally attributed to the immobilisation of N by microbial growth. Added organic material with a high C/N ratio (e.g., . 20 kg C (kg N)21) provides adequate C substrate, but as the C/N ratio of microbial biomass is lower (3—14 kg C (kg N)21), N is taken up from the mineral N pool in the soil to meet the shortfall, thereby reducing that available for crop uptake. This can be a major problem if it coincides with critical growth stages of the crop. Mafongoya et al. (1997b) have shown that N immobilisation also occurs when the lignin and polyphenol content of the residues incorporated into the soil were over 15 and 3%, respectively. Interestingly, N immobilisation resulting from high polyphenol levels seems to last much longer than that resulting from low C/N ratios (Palm et al., 1996). As the C content of dry biomass is usually around 0.4 kg C (kg DM)21, the critical C/N ratio of 20 kg C (kg N)21 corresponds to a biomass N content of between 2.0 and 2.5% (Palm et al., 1997b). Table III shows the N contents for a range of organic materials, from which it can be seen that some cover crop species, the leguminous tree species, and Tithonia diversifolia all contain N levels above 2.5%. The animal manures, on the other hand, particularly

LOW EXTERNAL INPUT TECHNOLOGIES

511

Table III Mean, Maximum and Minimum N Percentages for Various Organic Materials Organic matter source Non-legume Legumes (cover crops)

Legumes (trees)

Animal manure

Crop residues

Tithonia diversifolia Crotalaria juncea Canavalia ensiformis Mucuna pruriens Leuceana leucocephala Gliricidia sepium Sesbania sesban Cattle Cattle (Lekasi et al., 1998) Pigs Pigs (Lekasi et al., 1998) Poultry Poultry (Lekasi et al., 1998) Maize Sorghum

Mean

Max

Min

3.38 3.47 2.89 3.23 3.68 3.38 3.54 1.04 1.40 3.79 2.00 4.02 2.40 1.01 0.63

4.59 6.30 4.74 6.05 6.32 5.33 4.81 4.15 2.00 4.25 2.20 6.73 2.60 3.07 0.63

1.10 0.80 0.23 0.83 1.04 1.33 1.39 0.30 0.50 3.08 1.50 1.85 2.30 0.25 0.63

Compiled from the Organic Resources Database (Palm et al., 2001) and Lekasi et al. (1998).

from cattle, tend to have mean N contents that are likely to lead to net immobilisation of N in the soil. Data from Lekasi et al. (1998) also suggest the pig and poultry manures are marginal in terms of N levels, although the data from the Organic Resources Database show more favourable levels of mean N for these animals. The low N contents in most cereal crop residues also suggest that there may be negative yield effects on the crops to which they are added, due to net N immobilisation. These issues are important when adding OM to the soil, whether it is to improve the physical or chemical characteristics of the soil. Where long-term improvement to soil physical characteristics is important, the requirement will be for moderate- to low-quality OM to be applied. However, this may result in N and P immobilisation, particularly without supplementary use of inorganic fertilisers, and crop production may therefore decline. Where an improvement to the soil nutrient status in the short-term is needed, high-quality OM with low C/N ratios should be applied.

3. Nutrient Mining In view of the large quantities of OM required and also because of the resulting uptake of nutrients during the production of this biomass, there is the question of whether such production is sustainable in the context of subsistence agriculture in the tropics. For example, biomass transfer systems,

512

A. GRAVES, R. MATTHEWS AND K. WALDIE

such as cut-and-carry grasses or fodder banks, may function effectively for a certain period of time and relieve pressure on other sources of biomass, in particular, forest biomass and on-farm sources of biomass. However, it is unlikely that such systems can exist long without additional inputs, as nutrients are “mined” from the soil in which the biomass is being produced. In Kenya, it was found that the introduction of fodder banks was effective for a short while, but it soon became apparent that their productivity was declining due to the continuous removal of biomass and hence of nutrients from the soil (Wandera et al., 1993). Similar considerations apply to the use of Tithonia diversifolia as a supply of P to crops. Tithonia is able to scavenge relatively large quantities of P from the soil and to provide biomass with relatively high concentrations of P for incorporation as OM. However, this does not add to the net amount of P in the soil, but rather is a way of redistributing it spatially. Clearly, Tithonia will eventually mine the soil of P and other nutrients, which is unsustainable at the level of the whole farm. Unless resource-poor farmers have access to large areas of off-farm land growing Tithonia, it is unlikely that the system can be sustained over long periods of time. One solution might be to fertilise the on-farm biomass banks or hedges, but in this case farmers may as well fertilise the crops directly. The sustainability of the System for Rice Intensiﬁcation (SRI) technology developed in Madagascar in the 1980s by Fr. Henri de Laulani´e (Stoop et al., 2002) and promoted as a successful example of low-external input agriculture (e.g., Pretty and Koohafkan, 2002) can also be questioned from this point of view. The approach focuses on early transplanting, planting of single plants rather than hills, a square arrangement of plants rather than in rows, frequent weeding, and non-ﬂooding of the soil during the vegetative period, resulting in less seed and water being required. Compost can be added as a source of nutrients, although this is not an essential part of the package. Huge increases in rice yields up to 21 t ha21 without the use of purchased inputs of fertiliser and pesticide have been claimed (Uphoff, 1999). Some of the individual components (e.g., adequate spacing and weeding) are commonly recommended practices, but the remarkable yield increases are claimed to arise from the “synergistic” effect of all of these used in combination (Stoop et al., 2002). Part of the success of the approach appears to be in maintaining the soil in an aerobic state rather than completely ﬂooded as in traditional irrigated rice cultivation, resulting in higher rates of organic matter mineralisation (Stoop et al., 2002), so that more nutrients can be supplied from the soil rather than from fertiliser applications. Certainly, anaerobic conditions are known to depress mineralisation rates, and there are also reports of increases in rice yields in well-drained soils compared to ﬂooded soils (e.g., Ramasamy et al., 1997). While the yields of 21 t ha21 are undoubtedly an extreme (and indeed questionable physiologically), average yields of 8.8 t ha21 in farmers ﬁelds have been reported (Stoop et al., 2002). Such yields will remove about 130 kg N ha21 along with other nutrients at each harvest.

LOW EXTERNAL INPUT TECHNOLOGIES

513

Even if such low-fertility soils as reported (Uphoff, 1999) are able to supply such quantities of N, it is difﬁcult to see how such yields could be sustained for long without further addition of nutrients. If compost is to be the main source of nutrients, then quantities in the order of 33 t DM ha21 (at 2% N content and 20% recovery rate) would be required to replace the N removed by the rice crop. The 21 t ha21 rice yields were reportedly achieved with applications of 40 t FW ha21 (~16 t DM ha21) compost made from leaves of Tephrosia, Crotalaria, and banana along with rice straw (Uphoff, 1999). It is not easy to see how such quantities of compost could be produced and transported by a single household on a sustainable basis, except for very small areas of crop. Estimates of annual household waste production (potentially available for making compost) vary widely from ~100 kg DM year21 in Uganda (Wortmann and Kaizzi, 1998) to as high as 9.8 t DM year21 for households of 8 –14 people, also in Uganda (Briggs and Twomlow, 2002), although at between 700 and 1225 kg DM person21 year21, this last value appears somewhat high. Pantanali (1996) gives an intermediate value of 2 t DM year21 for the typical annual production of household waste in Lesotho, which, assuming a family of eight persons, represents about 250 kg DM person21 year21. Estimates for low-income households in South Africa are as low as 40 kg person21 year21 (Durban Metropolitan, 1999). For comparison, average household waste produced per person in England is about 500 kg person21 year21. Whatever ﬁgure is used, it would seem that household waste alone is unlikely to be able to meet the nutrient requirements of such highyielding crops, and that nutrients, either in inorganic or organic form, would have to be collected from a wider area for the enhancement of the cropped area.

4. Biological Nitrogen Fixation Biological nitrogen ﬁxation (BNF) by legumes is a key process in LEIA technologies as it potentially results in a net addition of N to the system. However, the quantity of N ﬁxed by legumes is difﬁcult to quantify and varies according to the species involved and location. Webster (1998) noted that estimates of the amount of N ﬁxed by groundnuts and grain legumes range from about 25 to 200 kg N ha21 during growing seasons of 60 –120 days. Bouldin et al. (1979) found that some legumes seemed to ﬁx N at relatively high rates — for example, values up to 535 kg N ha21 have been recorded for Crotalaria spp. and 400 kg N ha21 for other pure legume green manure crops. Moore (1962) reported that Centrosema pubescens (star grass) ﬁxed N at the rate of 280 kg ha21 year21. Reviewing the contribution of BNF in alley cropping, Sanginga et al. (1995) quote measured values of BNF of between 100 and 300 kg N ha21 year21 for some tree species such as L. leucocephala, Gliricidia sepium, and Acacia mangium, but values as low as 20 kg N ha21 year21 for others such as

514

A. GRAVES, R. MATTHEWS AND K. WALDIE

Faidherbia albida and Acacia senegal. Values compiled by Brady (1990) indicate that rates of N ﬁxation by a range of legumes vary between 5 and 300 kg N ha21 year21, with an average of about 100 kg N ha21 year21. The amount of N ﬁxed biologically that is actually taken up by the main crop is difﬁcult to determine with accuracy. In their review, Sanginga et al. (1995) estimate the contributions of N from trees in alley crops to an associated food crop to range from 25 –102 kg N ha21 season21 from root turnover and nodule decay, with a further 40 –70 kg N ha21 season21 being supplied through prunings. This represents recoveries of N of about 30%, although values as low as 5 –10% were reported. Thus, between 65 and 172 kg N ha21 season21 could potentially be supplied to the main crop over a season, representing enough N to support maize crops yielding between 3.5 and 8.5 t ha21. These ﬁgures, however, are based on BNF rates towards the high end of the range. Taking the average value of 100 kg N ha21 year21, and assuming that the recovery by the crop was the same 20% suggested by Giller and Cadisch (1995), with the rest lost through immobilisation, adsorption, volatilisation, leaching, and denitriﬁcation, BNF would provide around 20 kg N ha21, or roughly the N requirements of a 1 t ha21 maize crop. Moreover, this ﬁgure assumes a pure stand of the legume — if it is a component of mixed species system, then the addition of N to the system would be correspondingly less (Snapp et al., 1998). Kumara Rao et al. (1981) showed that both relay and sequential intercrop systems in southern Malawi produced insufﬁcient quantities of biologically ﬁxed N to maintain a maize biomass yield of 4 t ha21 on infertile land. They estimated the N contribution of Sesbania sesban when relay intercropped in low-fertility areas to be 28 kg N ha21, and that from a pigeonpea/groundnut intercrop to be about 20 kg N ha21. In comparison, the quantity of N required to sustain maize yields was estimated to be about 100 kg N ha21. Similar calculations were made by Matthews et al. (1992a) for alley crops in northern Zambia — a net amount of 14 kg N ha21 year21 was estimated to be added to the system through BNF, only 13% of the local recommended fertiliser N application rate for maize. Despite this, maize crop yields of 2 – 2.5 t ha21 were achieved for 4 years, which, with an annual removal of 30 – 35 kg N ha21 year21 in the harvest, suggests that some nutrient mining may have been occurring. The growth rate of the legume is a strong determinant of the rate of BNF. Any factor that reduces this growth, therefore, will also reduce the introduction of N into the system through BNF. In particular, competitive main crops such as maize may reduce the growth, and hence BNF rates, of legumes growing in an intercrop (Shumba et al., 1990; Muza, 1995; Kumwenda et al., 1997b). For example, Patra and Poi (1998) noted that intercropping maize with various legumes caused the number of nitrogen-ﬁxing nodules on the legume crops to decrease due to shading, while Fujita et al. (1993) found that artiﬁcial shading reduced BNF in several pasture legume species. In agroforestry systems, competition for light can be managed with timely pruning of the perennial species. It has been suggested

LOW EXTERNAL INPUT TECHNOLOGIES

515

that competition for water and nutrients can also be managed with pruning, although the effectiveness of this is still debated (Snapp et al., 1998). Nair (1993) has pointed out that controlling the below-ground competition of roots is far more difﬁcult than controlling the above-ground competition of canopies for light. Reduced light intensity as a result of interception from a dominant main crop has also been found to reduce nodulation (Webster and Wilson, 1998). Another factor inﬂuencing the net contribution of N to the system is the amount that is removed if the legume crop is harvested for grain or forage. Where the legume produces a useable product, farmers are likely to harvest and use the product. In the case of edible grain legume crops grown either as intercrops or in rotation, harvesting the grain can result in lower than expected residual N effects on the following crop, particularly if the grain legume has a high harvest index. In Brazil, residual N beneﬁts from using soybean as a leguminous green manure were greatly reduced by harvesting (Boddey et al., 1997). Farmers are likely to leave grain for incorporation as a green manure only where they do not market or consume it themselves. Boddey et al. (1997) have suggested that the solution may be to develop or use varieties of soybean that produce extra biomass while producing the same grain yield, thereby reducing the harvest index and allowing more N to be incorporated into the soil with the extra plant biomass. Grain legumes with a naturally low harvest index, such as pigeonpea or groundnut, or plants with no commercial or subsistence value could also be used. However, such solutions are only likely where improved N management is more important than short-term subsistence or cash beneﬁts. Legumes may also have other important functions within the smallholder’s strategy. For example, in the semi-arid areas of northern Namibia, McDonagh and Hillyer (2000) found that for cowpea, bambara and groundnut intercrops to be able to make any contribution of N to the system, there should be no grazing or burning of legume residues; but as cattle were an integral part of the system, this was unlikely to be a popular option with the farmers. Increasing the legume plant density just to the point where it began to affect the growth of the pearl millet contributed only about 4 kg N ha21 to the system, which increased millet yields by 80 kg ha21, a somewhat insigniﬁcant amount. They concluded that grain legumes alone are unlikely to be able to improve soil fertility in the area substantially, and that external fertiliser inputs would be necessary, although the uncertain rainfall makes investment in soil fertility unattractive for farmers there. It would appear, therefore, that in many cases BNF approaches are unlikely to provide more than a small fraction of the N requirements of main crops, unless low yields are accepted, or unless the farmer has access to other sources of leguminous biomass. Certainly, the use of these techniques to supply the full N requirements of the crop will require more land under the legume than under the main crop, which deprives the farmer of land that could otherwise be used for subsistence or cash crops. The use of mixed species intercropping is likely to decrease main crop yields through competition, particularly in difﬁcult

516

A. GRAVES, R. MATTHEWS AND K. WALDIE

biophysical conditions, where legume growth is limited, for example, by lack of P, or where lack of rainfall exacerbates competition for water. Nevertheless, Snapp et al. (1998) suggest that farmers may contemplate the use of such systems as part of the solution to N management, providing that the area available from growing the main crop is not reduced too much.

B. SOCIO-ECONOMIC ISSUES As the technologies we have discussed are ultimately used by people, social and economic factors are likely to be as important, if not more so, than the biophysical characteristics of the technologies. Amede (2001), for example, noted that in the Ethiopian highlands, adoption of various techniques such as cover crops and crop residue incorporation related as much to farm size and availability of labour as to the conditions of the soil. In this section, we discuss various socio-economic factors that can inﬂuence the successful uptake of LEIA technologies.

1.

Land

The amount of land required to produce enough biomass to maintain or improve the SOC level of cropped land or to supply sufﬁcient nutrients to meet crop requirements has already been discussed. The calculations described previously indicate that in general, a minimum of 3 ha of land is required to produce enough plant biomass to maintain SOC at 1% or to meet the nutrient requirements of 1 ha of cropped land. Essentially, in such a system, nutrients are being harvested from a wider area to enhance productivity within a smaller area, as in some slash-and-burn cultivation systems (e.g., chitemene and fundikila in Zambia, (Matthews et al., 1992a)). In many cases, depending on the type of biomass, climate and soil conditions, this ratio may be considerably more than 3:1. In the case of animal manures, where animals are used to gather nutrients from a wider area, the ratio is likely to be higher because of the losses of nutrients during manure storage and N losses in the urine not incurred in direct incorporation of biomass in the soil. Palm et al. (1997a) estimated that between 14 and 42 ha of miombo woodland would need to be grazed to provide enough N in manure for a 2 t ha21 maize crop, while ratios of up to 45:1 have been estimated for other extensive systems (Turner, 1995). In many situations, it is very unlikely that resource-poor farmers will have access to sufﬁcient land, over and above that needed for crop production, to produce the appropriate quantities of biomass. Thus, if organic materials are to be used solely for soil fertility maintenance or improvement, in most cases, it would seem that they must be

LOW EXTERNAL INPUT TECHNOLOGIES

517

obtained from off-farm sources such as forests or, possibly, as waste from urban areas. Even with intercropping or alley cropping, where a leguminous species is grown on the same piece of land as the main crop, there are costs; ﬁrstly, in that the area taken up by the legume cannot be used by the main crop, and secondly, due to potential competition for resources such as light, water, and nutrients. Few studies have demonstrated the possibility of achieving increases in main crop yield per unit of total system area while maintaining soil fertility. Relay cover cropping with annual legumes, or perennial legumes like C. cajan for a single season, may address the issue of direct competition, but is often not feasible due to lack of rainfall outside the main cropping season. Thus, legume growth and hence BNF rates do not compensate for the N that is removed in the harvest of the main crop. On the other hand, if rainfall does continue, farmers may be more likely to grow a second food crop rather than an unproductive legume. Improved fallows may be an option for resource-poor farmers where fallows are already an accepted part of the cropping system. However, where land is limited and continuous cropping exists, improved fallows would have to replace a main crop, which is not feasible for most resource-poor farmers. Furthermore, resource poor-farmers are unlikely to be able to meet the ﬁnancial and labour costs required for the establishment of improved fallows. A natural fallow, particularly one of several years, may provide resources that an improved fallow cannot. Plants, animals, or residual germination and growth of cultivated crops provide important products for farmers and soil regeneration under a natural fallow may not be very different compared to that under an improved fallow, especially if the natural fallow contains leguminous plants. Suggestions have been made that ﬁeld boundaries could be used to source nutrients, particularly P, from biomass transfer species such as Tithonia diversifolia (e.g., Briggs and Twomlow, 2002). While this may make some contribution, a quick calculation sufﬁces to show how much P, for example, might be supplied in this way. If we assume that a landholding is 1 ha in area consisting of 3 –4 ﬁelds, (typical for many small-holdings in Nepal and Ghana), there would be about 600 m of ﬁeld boundaries. Tithonia growing on these boundaries producing 1 kg DW m21 from biannual pruning (Jama et al., 2000), would therefore supply only 0.6 t DM ha21, or about 2.2 kg P ha21, equivalent to about 15% of the 15 kg P ha21 removed in the harvest of a typical 5 t ha21 maize crop. Alternatively, this amount of P would only support a 0.7 t ha21 maize crop in P-deﬁcient soils, and probably far less if due consideration is given to the usual crop recovery rates of nutrients from organic material. It must also be borne in mind that the Tithonia hedge has probably extracted some of this P from the adjoining cropped ﬁelds in the ﬁrst place. In some situations, there may be ways of making better use of land that is unproductive at certain times of the year. For example, in the Barind Tract of Bangladesh, land is often left fallow during the dry season following the main

518

A. GRAVES, R. MATTHEWS AND K. WALDIE

rice crop, as it is often difﬁcult to establish a second crop during this time due to drying of the soil surface layers (Musa et al., 2001). This is despite the fact that there is usually sufﬁcient residual water further down in the soil proﬁle left from the irrigated rice crop. In recent years, drought-tolerant crops such as chickpea have been introduced, but establishment is not certain, and complete crop failure may result. Recent research has shown that “priming” chick-pea seed, by soaking it overnight in water, can result in a marked improvement in crop establishment, making the difference between a healthy crop and no crop at all (Musa et al., 2001). Attractions of the technique are that it is simple for farmers to implement, requires no expensive inputs, and is aimed at resource-poor farmers rather than at those with mechanised systems (Harris et al., 1999). Farmers welcome the ability to gain an extra crop in the sequence, particularly of chickpea, which currently commands good prices in the market, at little extra cost in land, labour, or capital. Of course, there are questions of whether such systems of increased intensiﬁcation are sustainable, or whether soil fertility decline and possibly weed encroachment is enhanced, although the experience in Bangladesh would suggest that this is not the case. The area in question was converted from forest about 150 years ago, and although current SOM levels are very low (0.5 –0.8%), reasonable main crop yields seem to be obtained year after year with appropriate inputs of inorganic fertilisers. Whether a further crop in the sequence will reduce SOM levels even lower remains to be seen. Land tenure is also an important issue inﬂuencing the use of LEIA technologies. It is generally thought that land users who do not own their land have less incentive to invest in technologies that take some time for soil fertility beneﬁts to accrue. Share-cropping is one such example, where a farmer exchanges a proportion of farm output in exchange for the right to crop an area of land (Ellis, 1988). The system has generally been considered as being less economically efﬁcient than if the land is owned, as share-croppers are thought to input labour only at a level that maximises their own perceived share of farm production, which is less than what they would be prepared to give if they receive the total production from the farm (Todaro, 1999). Nelson et al. (1998) analysed the economics of upland agriculture in the Philippines, concluding that share-cropping would reduce the economic attractiveness of alley cropping techniques compared with alternative techniques (Fig. 2). This was because it was assumed that, under the particular share-cropping arrangement, landlords would not contribute to the establishment costs of the hedgerows, while a portion of the main crop would be given to them as part of the tenancy agreement. However, other analysis suggests that share-cropped farms are not necessarily inherently less efﬁcient than ownermanaged farms (Reid, 1976). Also, share-cropping may at least give very poor farmers the opportunity to farm, and some evidence suggests that it is the poorest and most unskilled who stand to beneﬁt from share cropping (Reid, 1976), especially where the landlord also wishes to maximise the productivity of the farm. Similarly, Ayuk (2001) in reviewing a number of studies, concluded that

LOW EXTERNAL INPUT TECHNOLOGIES

Figure 2 (a) The impact of sharecropping on the net present value of open ﬁeld, fallow and hedgerow intercropping in upland Philippine agriculture at a discount rate of 25%. (b) The net present value without the impact of sharecropping is also shown for all three systems at a discount rate of 25%. (Source: Nelson et al., 1998).

519

520

A. GRAVES, R. MATTHEWS AND K. WALDIE

the relationship between land ownership and willingness to invest in technologies to improve SOM was not a simple one. While in some cases, there was a strong correlation between long-term use rights and soil fertility improvement in the form of manuring and mulching (e.g., Blarel, 1994), in other cases, there was no correlation at all (e.g., Golan, 1994). Interestingly, a number of studies indicate that traditional tenure systems do not inhibit land users from investing in improving their land, and that a greater danger may be the granting of titles to urban dwellers who have no interest in agriculture (Ayuk, 2001). Fragmentation of land is another factor that may inﬂuence the adoption of certain LEIA technologies. Fragmented landholdings may result in a single farmer having to transport inputs to several isolated plots of land in several different locations. This difﬁculty is particularly great with the use of biomass transfer techniques, where several tonnes of biomass per hectare may be required. In the mountainous terrain found in Nepal or Bolivia, for example, transporting heavy loads to small isolated plots of land is extremely arduous. This is an extremely labour-intensive process if the farmer has to do it alone, or expensive if he/she hires labourers. Moreover, land fragmentation may result in decreasing ﬁeld sizes, which makes the implementation of certain techniques impractical. Trees used for green manure or fodder, for example, may shade out the crop if planted on the borders of small ﬁelds.

2.

Labour

The labour required to make use of LEIA technologies may also constrain their adoption. Often household labour may need to be supplemented with that purchased from off the farm. While such contributions are seldom accounted for in analyses of technology costs and beneﬁts, especially when female labour is involved, it is important to recognise that the use of domestic labour represents a real opportunity cost. The need for external labour, which will generally involve a cash transaction and therefore directly affect household ﬁnances, is more readily acknowledged. Both aspects are important. Ali (1999), in a study of farmers in Asia, found the cost of labour to be partly responsible for making nutrient supply through organic matter less cost-effective than through mineral fertilisers, a situation which is likely to get worse due to rapidly rising wages. In India, the requirement for a pair of bullocks and a ploughman was 10.5 days ha21 in a rice/green manure/rice rotation, while in Nepal, the number of days needed in a wheat/green manure/wheat system was 11 days ha21. In both cases, the cost of this was about US$40 ha21 which largely accounted for the differences in economic performance between green manure and mineral fertilisers. This occurred despite the reduced need for weeding and the increased yields obtained. Ali (1999) also analysed the economic

LOW EXTERNAL INPUT TECHNOLOGIES

521

performance of using grain legumes rather than green manures, and found that these were more proﬁtable, despite increased labour requirements, because the legume produced grain that could be sold. However, acceptance was not as high as expected because the incorporation of the legume biomass delayed the planting of the monsoon crop and increased the burden of labour at a time when it was already high. The grain legume was also susceptible to pests and diseases. Similarly, the cutting and carrying of biomass, such as from Tithonia, is extremely labour intensive, particularly if it is to supply the full crop P requirements in a P-deﬁcient soil (Buresh and Niang, 1997). For a typical crop requirement of 15 kg P ha21, the application of about 20 t DM ha21 of Tithonia biomass would be needed (assuming a 20% recovery rate of the P by the crop), equivalent to about 100 t ha21 of fresh biomass. Harvesting at the rate of 100 kg FW man-day21 (ICRAF, 1997), 1000 man-days of labour would be needed just to harvest the biomass! These are not the only costs. Tithonia also needs to be propagated and prepared for incorporation in to the soil (ICRAF, 1997), and, although it does not have thorns, it is difﬁcult to handle because it is sticky and exudes a pungent smell (Jiri and Waddington, 1998). In addition, because of its ability to regenerate, it may invade farmland (Jama et al., 2000), thereby increasing the labour required by a farmer to control it. The implication is that either labour must be plentiful and cheap, or that the crops fertilised with Tithonia should be high-value crops. As an example, Jama et al. (2000) cite data from ICRAF showing that under farmer-managed conditions in western Kenya, investing in Tithonia fertilisation was viable for high-value kale (Brassica olecacea), but uneconomical when used with a low-value crop such as maize. Supplying sufﬁcient P through animal manure also requires substantial labour. For example, the supply of 15 kg P ha21 through cattle manure (using a mean concentration of P of 0.138% and a 20% recovery rate of P by the crop) would require 55 t DM ha21 of manure, equivalent to 275 t FW ha21. Assuming that the farmer had to transport the manure manually, that 20 kg of manure per load could be carried at 5 km hr21, and that this load had to be transported 100 m from the source, transporting this 15 kg P would require about 550 man-hours of labour. In comparison, transporting the same amount of P in poultry manure, the same distance would take only about 43 man-hours. Loads may often have to be carried much further than this, and where large amounts of manure have to be transported long distances, farmers may struggle to provide the labour required. Where the manure has to be transported in mountainous terrain, such as in Nepal, the amount of time required to transport manure will be even greater. For many farmers, weeding is one of the most labour-demanding activities undertaken. Gill (1982) noted that hand-hoe weeding in India required between 200 and 400 man-hours ha21 and that two weedings were needed during the growth and development of many ﬁeld crops. Van Tienhoven et al. (1982) found that between 13 and 37 man-days ha21 of labour were required to weed a maize – bean production system in the Jinotega region of Nicaragua.

522

A. GRAVES, R. MATTHEWS AND K. WALDIE

This accounted for between 21 and 35% of family labour. Ruthenberg (1980) has compiled data of labour requirements for weeding from various sources — in Ghana, for example, weeding in a maize system required 31% of the total labour (about 186 man-hours ha21), while in Columbia, weeding a cassava crop required about 55% of total farm labour (about 408 man-hours ha21). Other traditional agricultural systems cited were less intensive, although they all required at least 20% of total labour requirements for weeding. In the Ichilo-Sara area of Bolivia, Pound et al. (1999) found that weeding could require from 35 man-days ha21 of labour in the ﬁrst year of a cropping cycle to 53 man-days ha21 in the third year as weeds started to dominate the system. The increase in weed cover was associated with large decreases in rice yield, with yields in the third year only about 30% of those in the ﬁrst year. This was probably due to the combined effect of weed growth and declining soil fertility. The labour requirement for weeding was reduced when Calopogonium was sown as an intercrop 25 days after the rice planting, but not if the Calopogonium was sown 45 days after the rice planting or as a cover crop after the rice was harvested. Pound et al. (1999) make the point that these reductions in labour were probably not great enough to be of practical signiﬁcance to subsistence farmers, and clearly, would not make a substantial difference to their livelihoods. The system for rice intensiﬁcation (SRI) discussed above (Stoop et al., 2002), provides an interesting example of how labour requirements can limit the uptake of an improved practice. Although labour requirements are high (38 – 54% more than traditional methods), returns to labour are also high ($3.87 day21 compared to $2.61 day21), a characteristic which has been seen as a major advantage of the technology compared to traditional approaches. Despite these apparent advantages, farmer adoption of SRI in the areas where it was promoted has been low, abandonment of the method by those farmers who originally adopted it has been high, and those who continue to practice the method rarely do so on more than half of their land (Moser and Barrett, 2002). Participatory surveys showed that this was because the recommended technological package required signiﬁcant additional labour inputs due to the extra weeding and water management involved (the latter on a daily basis), during the time of year when poorer farmers need to seek employment with other farmers to earn cash to meet immediate consumption needs. Those who did try SRI were less likely to rely on agricultural day labour as a source of income, and were also more likely to have larger farms due to the economies of scale in offsetting the initial costs involved in levelling their ﬁelds and in redesigning their irrigation systems to allow more precise water control. Interestingly, adopters were also less likely to have a relatively high salaried income as the opportunity cost of foregoing their salary to supervise hired labour for the SRI was greater than the returns gained. Thus, poorer farmers for whom the technology was designed to help were less able to take advantage of it. This example highlights the need to evaluate new technologies in the context of the whole agricultural system — even

LOW EXTERNAL INPUT TECHNOLOGIES

523

considering returns to labour in this case was not sufﬁcient to predict the uptake of SRI, as the timing of labour and income requirements throughout the year was the most important factor. Assertions by the proponents of SRI that “farmer perceptions and practices are not necessarily wise or optimal” (Uphoff, 1999) may, therefore, be somewhat premature when the wider picture is taken into account. Moser and Barrett (2002) make the point that opportunity cost is often overlooked in evaluating LEIA technologies, i.e., even though the ﬁnancial cost of inputs may be low, this does not mean that the technology is without other costs. They offer an interesting comparison with the technique of off-season cropping (OSC), which was introduced in the same areas in Madagascar as SRI, in which crops such as potatoes are grown in the winter season after the rice harvest. Adoption of the OSC technique has been high at 84% of households (across a range of wealth classes), and with no disadopters to date (2002). The key to its success appears to be the ease with which it ﬁts into the existing agricultural system. Despite relatively high input costs for the purchase of seed and fertiliser, these occur at a time when farmers have just completed the rice harvest, and have more time and money. Moreover, the OSC harvest provides them with food and working capital at the beginning of the rice season, freeing the household from needing to do off-farm work to earn wages to purchase food. Farmers also perceive a carry-over beneﬁt of the fertiliser applications on soil fertility in the following rice crop. A large proportion of the farmers adopting the technique had learned of it from other farmers, suggesting that it was easy to learn. The importance of the labour proﬁles of new technologies ﬁtting in with existing labour patterns has also been noted by Hoang Fagerstro¨m et al. (2001) — biomass banks of Tephrosia were not popular with farmers in Vietnam as the labour required for cutting and transporting the biomass coincided with busy periods for other farm activities. On the other hand, a two-year crop/twoyear fallow cycle for upland rice ﬁtted in well with the existing split of work between upland and lowland cultivation.

3.

Economics

In addition to its labour requirements, the adoption of a particular technique will also depend on its economic beneﬁts in relation to other options, regardless of its biophysical attributes. For example, in a study of the economic viability of combined fertiliser, green manure and grain legume techniques, Ali (1999) found that the short-term beneﬁts of green manure were “negative or trivial” compared with inorganic fertiliser, despite on-farm experiments showing that yields were higher with combined fertiliser and green manure treatments. On the other hand,

524

A. GRAVES, R. MATTHEWS AND K. WALDIE

farmers using grain legumes obtained short-term beneﬁts from the sale of the grain as well as long-term beneﬁts, compared to inorganic fertiliser systems. Ali (1999) concluded that recommended techniques need to have short-term economic beneﬁts, or risk being rejected by farmers despite any long-term advantages they might have. The relative costs of inorganic sources of nutrients is a major determinant of the use of LEIA techniques. In Nepal, for example, Pilbeam et al. (1999a) showed that margins were generally negative when manure was applied either alone or in combination with fertilizer, but positive with applications of fertilizer. Similarly, in Zambia, subsidies on N fertiliser in the 1980s stimulated many subsistence farmers to adopt high-input maize production based on mineral fertiliser inputs (Matthews et al., 1992a). An economic analysis showed that under these conditions, the best net returns were obtained from applying as much mineral fertiliser N as possible (up to 120 kg N ha21 in the study) rather than using alley cropping (Matthews et al., 1992b). However, when the fertiliser subsidies were removed in 1990 through donor pressure, farmers reverted to their traditional low-input chitemene and fundikila shifting-cultivation systems. Under these conditions, alley cropping with Leucaena always gave a small positive return, and indeed, the highest net return was obtained from Leucaena and 60 kg N ha21 fertiliser. This analysis did not, however, include the cost of labour; if it had, the positive net returns from alley cropping would probably have disappeared. Sometimes, however, the yield increase under alley cropping may be economically viable. Chianu (2002) used partial budget analysis to show that compared to a bush fallow, alley cropping L. leucocephala becomes advantageous during longer fallow periods due to the production of fuelwood. However, it was noted that yield variability, labour scarcity, and risk aversion could inﬂuence the technology choice of the farmer. Similarly, novel alley cropping systems may be economically viable to farmers. In India, introducing geranium (Pelargonium spp.) into alleys of Eucalyptus citriodora did not affect the essential oil yield of the latter, but resulted in higher monetary beneﬁt over sole Eucalyptus plots (Singh et al., 1998). Intercropping with Java citronella and lemongrass also resulted in higher net beneﬁts than from Eucalyptus alone, although lemongrass did reduce Eucalyptus yields. In India, Pakistan, the Philippines, and Indonesia the prices of both land and labour have increased relative to inorganic fertiliser since the 1970s; consequently, the use of the latter has increased at the expense of labourintensive, land-extensive, organic matter approaches to maintaining soil fertility (Ali, 1999). In Taiwan, for example, green manure crops have decreased from an area of 153,000 ha in 1954 to 11,000 ha in 1991, while in India, Nepal, and Pakistan, green manure is no longer widely used. Ali (1999) calculated the economics of using Sesbania, Azolla, and rice straw as green manures in India, Indonesia, and the Philippines, assuming that a Sesbania green manure crop would provide 70 kg N ha21, Azolla 30 kg N ha21, and rice straw 18 kg N ha21.

LOW EXTERNAL INPUT TECHNOLOGIES

525

Ignoring the opportunity cost of growing green manure on productive land, he found that the beneﬁt/cost ratio of all three techniques was less than 1.0. This indicated that at these labour/fertiliser price ratios, green manure was not a costefﬁcient option in any of the three countries compared to inorganic fertiliser. Ali (1999) calculated that for green manure to be competitive with fertiliser and with a zero opportunity cost of land, the ratio of wages (US$ day21) to fertiliser price (US$ kg21 N) should not exceed 3.0 for Sesbania and Azolla, and 2.0 for rice straw. As the ratio was higher than this in India, Indonesia, and the Philippines, this explained the low use of green manure techniques there, while a more favourable ratio accounted for the greater use of green manure in Myanmar and China. Interestingly, Swinkels et al. (1997) found that farmers in the high population density areas of western Kenya were more likely to practice fallowing due to the availability of off-farm income, contradicting the view that continuous cropping on depleted soils is the ﬁnal stage in the land intensiﬁcation process (e.g., Ruthenberg, 1980). Cropping depleted land gave farmers poorer returns to labour, and, therefore, it was more rational to take work to buy food and allow the soil fertility to regenerate. Only a 21% increase in crop yield following a one-year fallow was necessary to compensate for the loss of returns during the fallow, mainly due to the savings in crop husbandry costs. Similar behaviour has been observed in Zambia (Low, 1988), Kenya and India (Dewees and Saxena, 1995), and Indonesia (Nibbering, 1991). As land and labour prices rise, there is pressure to shift away from green manure towards intercropping, grain legumes, compost, and animal manure. In India, farmers growing Crotalaria juncea and Tephrosia purpurea as green manure after winter rice moved to grain legumes such as C. cajan and Vigna spp., particularly if irrigation water was available. In the central terai of Nepal, Crotalaria juncea and Sesbania rostrata have given way to grain legumes such as mung-bean, particularly if water is available during the dry season. In many cases, only dramatic increases in fertiliser prices due to scarcity of fossil fuels may make green manure viable in these countries, because transport networks are well developed (Ali, 1999). Grain legumes, on the other hand, have potential if the grain has economic value, although the removal of nutrients in the grain will largely undermine their utility as a source of N. Economic analyses of improved fallow systems have given conﬂicting results. In upland northern Vietnam, a Tephrosia fallow had a negative net present value (NPV) and was judged not to be a rational choice where natural fallow was still viable (Hoang Fagerstro¨m et al., 2001). In Mali, Kaya et al. (2000) concluded that improved fallows were only ﬁnancially attractive if fodder had a value and if subsequent crop yields exceeded the regional average of 2500 kg ha21. This is in contrast to the example in western Kenya, where improved fallows were seen to be a rational choice because of the relatively small increase in crop yields required to break even, provided labour demands of establishing the fallow were low (Swinkels et al., 1997). A study in West Africa showed that cumulative net

526

A. GRAVES, R. MATTHEWS AND K. WALDIE

Figure 3 Cumulative discounted net beneﬁts for continuous maize with no fertiliser, continuous maize with recommended, one-year Sesbania fallow fertiliser, two-year Sesbania fallow, and threeyear Sesbania fallow. (Drawn from data of Kwesiga et al., 1999).

beneﬁts were $1664 ha21 under a Pueraria fallow, $1121 ha21 under Chromlaena odorata natural fallow and cover crops, and $1113 ha21 under continuous cropping at farmer input levels (Tian et al., 2001). In Cameroon, Adesina and Coulibaly (1998) found that improved fallows of Tephrosia, Sesbania, Mucuna, and Calliandra were all proﬁtable whether used alone or in conjunction with inorganic fertilisers. In Zambia, Kwesiga et al. (1999) showed that one- and two-year Sesbania fallows gave higher net returns than continuous unfertilised maize crops, while a three-year Sesbania fallow was the same (Fig. 3). Fertilised continuous maize, however, gave more than twice the net return of the best fallow treatment. These authors make the point that the timing of cash ﬂow from improved fallows is important for subsistence farmers — even though the net return of the 2-year fallow was higher than the unfertilised maize after 4 years, for the ﬁrst 2 years the net returns were negative, and the farmer would need to have other sources of food and income.

IV. DISCUSSION A. INTEGRATED NUTRIENT MANAGEMENT In general, LEIA technologies aim at reducing losses, while also introducing inputs from organic sources, while HEIA technologies aim at ensuring that

LOW EXTERNAL INPUT TECHNOLOGIES

527

agronomic inputs (i.e., fertilisers, pesticides, water, energy) into the production system are maintained at a high level, with less attention being paid to reducing losses from the system. Gregory et al. (2002) refer to these two approaches as Type I and Type II intensiﬁcation, respectively. It is clear from the analysis in this review that the LEIA techniques we have discussed cannot alone meet crop nutrient requirements if potential yields are to be obtained. On the other hand, sole reliance on HEIA technologies is also not an option for many resource-poor farmers due to cost and availability of the inputs required, and is also not desirable because of the associated health and environmental pollution problems. There is, therefore, a growing consensus amongst researchers that the debate on whether HEIA or LEIA technologies are the most appropriate is largely irrelevant, and that the best way forward is through an INM approach, in which a combination of techniques from the two extremes is used (e.g., Sanchez et al., 2001). INM has been deﬁned as the “judicious manipulation of nutrient stocks and ﬂows, in order to achieve a satisfactory and sustainable level of agricultural production” (e.g., Deugd et al., 1998). Certainly, there is ample evidence (e.g., Palm et al., 1997a; Jadhao et al., 1999; Prasad et al., 2002) that the highest yields and returns can be obtained by a combination of maximising inputs into the system, both from organic and inorganic sources, and at the same time, reducing losses. Inorganic fertilisers have the advantage that nutrient concentrations are much higher than in organic material, so that handling and incorporation into the soil is greatly facilitated. On the other hand, organic matter in the soil acts as both a “binder” for added inorganic nutrients so that they are less likely to be lost by leaching and volatilisation and more likely to be taken up by a crop, and also as a source of nutrients in its own right through decomposition. It is also critical in improving the physical structure of the soil. The resulting greater efﬁciency of nutrient use through the combined use of both inorganic and organic sources of nutrients and reduced losses is referred to as Type III intensiﬁcation by Gregory et al. (2002). The challenge, therefore, is to identify those techniques from the continuum between the HEIA and LEIA extremes, which can complement each other to achieve this goal of sufﬁcient and sustainable production, regardless of whether they are labelled “organic” or “inorganic.” In general terms, Sanchez et al. (2001) have proposed a combination of (1) biological N ﬁxation by short-term leguminous fallows, (2) applications of mineral P fertilisers, (3) enhanced P cycling using Tithonia, (4) use of trees to maximise nutrient cycling, (5) return of crop residues, (6) soil erosion prevention, (7) improved crop management practices such as the use of better varieties, and (8) improved availability and timeliness of supply of inorganic fertilisers. For this approach to be feasible, ways need to be found to extract the indigenous deposits of rock phosphate present in many countries, particularly those in sub-Saharan Africa (Ayuk, 2001), and make it available to farmers at reasonable cost. At the farm level, improved management of organic resources, such as in the storage and application of

528

A. GRAVES, R. MATTHEWS AND K. WALDIE

compost and manure, may help reduce losses by volatilisation (Pilbeam et al., 1999b). Similarly, the development of alternative energy sources such as woodlots and the introduction of more efﬁcient stoves may help to free up crop residues currently used as fuel sources, which could then become available for SOM maintenance and improvement (Ayuk, 2001). Many of the nutrient management approaches in subsistence systems involve moving nutrients from one part of the landscape at various scales to another where they are more useful. For instance, several of the techniques we have reviewed involve collection of nutrients from a wider area through grazing or biomass harvesting and concentrating these on a smaller cropped area. In Nepal, for example, the fertility built up in forested areas over long periods of time is transferred gradually to nearby farms through collection of fodder for animals and biomass for enhancing on-farm soil fertility (Pilbeam et al., 1999b). In much of the analysis of farming systems, however, this heterogeneity is often ignored and the fertility of a farm is assumed to be constant across all ﬁelds or parts of the farm. In reality, farmers are usually very adept (consciously or subconsciously) at manipulating this heterogeneity to improve their livelihoods. Both Wortmann and Kaizzi (1998) and Briggs and Twomlow (2002) describe ﬂows of organic material from distant maize ﬁelds to higher value banana plantations nearer the household on smallholder farms in Uganda. In many countries, the soil fertility of small areas used as home gardens is enhanced by incorporating household waste containing nutrients gathered from a wider area, both from other parts of the farm through consumption of crops harvested there, and from off-farm sources such as food bought in a market. Higher-value crops such as vegetables or fruit, which would not grow well elsewhere on the farm, are often grown in these highfertility areas. Indeed, Sanchez et al. (2001) have suggested that the growing of high-value crops may be the most direct way out of poverty. For example, they quote highvalue vegetables such as kale, tomatoes, and onions in Kenya as being able to increase net proﬁts from US$91 to US$1665 ha21 year – 1. Whether this is viable on a large scale will depend on broader economic development and the availability of markets, storage and processing facilities, and urban population growth rates. High-value tree crops may also be promising. For example, extractions from the bark of Prunus africana can be used to treat prostate glandrelated diseases, and has an annual market value of $220 million per year (Sanchez et al., 2001). The demand has been so high that the species is now on the CITES (Convention on International Trade in Endangered Species) list, but is now in the process of being domesticated. Other examples include bush mango (Irvingia gabonensis) in West Africa, and Sclerocarya birrea from the miombo woodlands of southern Africa, which are used to make liqueurs. The danger is, however, that without replenishment in some form or another, eventually the fertility of those parts of the farm providing nutrients for these high-value crops will decline to a point where it is no longer possible for them to act as a nutrient source.

LOW EXTERNAL INPUT TECHNOLOGIES

529

Clearly, such systems are complex, and there are questions regarding overall sustainability, but with the advent of new crops, new markets, and better infrastructure in many regions, it may be worthwhile re-examining these traditional systems of nutrient redistribution carefully to see if there are any possibilities of adding value to the overall production system through innovative practices that farmers have not yet discovered. The NUTMON project is an example of an approach to develop INM strategies at the farm level (de Jager et al., 1998). The prime objective of the project was to investigate technologies that mitigate against nutrient depletion and possibly add nutrients which are economically viable and socially acceptable. This was done by monitoring nutrient inﬂows and outﬂows of a farm, calculating the balances, and quantifying the impact of INM practices on soil fertility, and hence agricultural production and sustainability. The approach included a diagnosis phase during which both qualitative and quantitative assessments of nutrient management and stocks were made from farmer interviews and by taking and analysing soil samples. Then, the most appropriate technologies for a speciﬁc farming system were determined from a combination of the quantitative nutrient ﬂows and balances, the economic performance indicators, and farmers’ perceptions, which were subsequently tested through onfarm trials. Existing indigenous or science-based technologies as well as any new ideas or modiﬁcations of existing technologies were considered. The approach allowed the study of a farm in a holistic way, taking into account the effects of many different household activities on the nutrient stocks and ﬂows and the economic performance of the farm, providing a way to assess the constraints to adoption of alternative INM technologies in relation to economic viability and demand for labour. Briggs and Twomlow (2002) followed a similar approach in determining ﬂows of organic material within smallholder farms in southwest Uganda. By helping farmers to conceptualise and draw diagrams of these ﬂows within their farms, several sources of organic material, such as hedgerows, weeds, fallowed areas, and ash residues, were identiﬁed which farmers had not previously recognised as potentially contributing to the household’s fertility management practices. At the national level, Cuba provides an interesting example of how INM practices can work (Carney, 1993). With the collapse of the Soviet empire at the end of the Cold war in 1989, the country was deprived of a source of imported fertilisers, pesticides, and fossil fuel. To cope, it was forced to look for other ways of sustaining its agricultural production, and focused on more efﬁcient nutrient recycling and reuse of organic urban waste, and biological pest control. Although food production initially dropped by 30%, it has since risen steadily, and to date (2003) there are no food shortages despite soils being severely degraded, a high population, and a continuing economic blockade. Better integration between rural and urban areas may be a key focus for other countries — in general in developing countries, there is a net ﬂow of nutrients from the former to the latter

530

A. GRAVES, R. MATTHEWS AND K. WALDIE

in the form of produce for human and animal consumption, which is not often returned due to the inability or reluctance to process human and other waste. In Ghana, for example, poultry manure produced in urban areas is considered to have little value and is often dumped on the roadside (Quansah et al., 2001). Thus, ways in which these nutrients can be retained and returned to the rural areas, rather than being lost to the groundwater or sea as usually occurs, need to be developed, taking into account, of course, the costs involved in processing and transport, and possible disease risks. Currently, the economics are not favourable — Palm et al. (1997a) showed that in Kenya, N and P in urban compost cost US$0.5 kg21 and US$1.2 kg21, respectively, compared to US$0.42 kg21 and US$0.18 kg21 in purchased inorganic fertilisers. It is unlikely that there are any techniques that will provide universal solutions — it is much more likely that progress will be made by farmers taking an idea and adapting it to their own particular “microniche.” These microniches will be unique for every farmer — not only will the biophysical environment vary, but each farmer will also have different perspectives based on their interests and experiences and be inﬂuenced by his/her own particular worldview (Scoones and Toulmin, 1998). It is perhaps more important that farmers have a good understanding of the principles involved in nutrient management, pest management, and crop interactions, rather than a detailed knowledge of a particular technique, and know how to apply this understanding to their own situation. Deugd et al. (1998) have emphasised the importance of any improvements to nutrient management being through farmer participation and learning. Thus, improvements should not be seen as optimal solutions to scientiﬁcally welldeﬁned problems, but more as stages in an adaptive learning process within a complex and changing environment.

B. A SYSTEMS PERSPECTIVE It is evident from many of the examples we have reviewed, that a major reason for the lack of widespread uptake of particular technologies is the failure to see them as part of a larger system. For example, the decline in numbers of farmers using the maize –Mucuna system in Honduras was found to be due largely to external socio-economic factors independent of the agronomic performance of the system itself (Neill and Lee, 2001). Similarly, despite the large increases in rice yields and high returns to labour claimed for the System of Rice Intensiﬁcation, it has not been adopted widely by farmers, and a sizeable fraction of those who did adopt it are now in the process of abandoning it (Moser and Barrett, 2002). This was ascribed to the large amount of labour required by the technology at the time of the year when poor farmers, facing cash shortages, need to work for other farmers to earn enough to buy food. It remains to be seen whether provision of credit facilities can ease this ﬁnancial pressure to seek

LOW EXTERNAL INPUT TECHNOLOGIES

531

off-farm work so that technologies such as SRI can be adopted. It is also questionable whether the large quantities of biomass required for making the compost used in the system would be available for the sustainable uptake of the technology on a large scale, both in terms of the numbers of farmers involved, and in terms of the actual cropped area under the technology. This biomass must of necessity, come from a wider area than that actually cropped. Widespread uptake of the Machobane system in Lesotho (Pantanali, 1996), and the Zai system in Burkino Faso (Roose et al., 1999), both of which also rely on large supplies of compost for their success, is likely to face the same constraint. Thus, failure to consider the whole livelihood system, including the inﬂuence of temporal ﬂows of labour and money on farmer decision making, can lead to false conclusions and unrealistic expectations on the part of the scientists about the appropriateness of a particular technology. Ironically, the SRI technology has been more widely adopted by middle-income farmers so far than by the poor farmers it was designed to help (Moser and Barrett, 2002). It is important, therefore, for researchers to try and view the situation from a farmer’s perspective (Douthwaite et al., 2002). In general, most researchers conceptualise problems associated with LEIA systems in developing countries in purely biophysical terms, particularly in relation to soil fertility decline and weed encroachment; less attention has been paid to the socio-economic issues (Ayuk, 2001). While this approach is useful from the perspective of scientiﬁc research, farmers seldom think in these terms. Rather, they are more concerned with how particular practices relate to their broader livelihoods. In considering whether or not to adopt a particular research product such as alley cropping, they are more likely to be inﬂuenced by how their livelihood will beneﬁt in terms of the extra food, cash, or quality of life it is likely to provide, than they are from technically framed arguments concerned with nitrogen ﬁxation and the like. This is not an argument about a “reductionist” versus an “holistic” approach. The fallaciousness of drawing a dichotomy between the two approaches has been pointed out by Kline (1995). Neither is superior to the other, and we would argue that a “reductionist” approach is essential, provided it is contextualised within a broader framework of analysis. An example of such a framework is the sustainable livelihoods (SL) approach currently being promoted by a number of aid agencies as a way of thinking more broadly about the objectives, scope and priorities for development, in order to enhance progress on the elimination of poverty (Ashley and Carney, 1999). The main feature of the SL approach is that it places people at “centre stage” of these discussions, rather than natural resources or commodities as has been the case in the past, and considers their assets (natural, human, ﬁnancial, physical, and social capital) and their external environment (trends, shocks, and transforming structures and processes). A key concept is that of “sustainability” — a livelihood is deﬁned as sustainable where it can cope with (and recover from) stresses and shocks, and maintain or enhance its capabilities and assets both now and in the future, while not undermining

532

A. GRAVES, R. MATTHEWS AND K. WALDIE

the natural resource base (Carney, 1998). An important point is that agriculture is seen as a system, rather than as simpliﬁed issues arising from single discipline perspectives. We would argue, therefore, that a more appropriate approach might be to consider the farmers’ problems from a livelihoods perspective, identify potential solutions to these problems, and match or develop relevant techniques in order to achieve these solutions. An illustrative example is provided in Table IV. The three broad problems we have deﬁned in Table IV are included in the livelihood outcomes of the SL framework, while the various techniques we have been reviewing contribute to the various livelihood strategies that subsistence farmers can adopt to achieve these outcomes. We do not claim that Table IV is exhaustive, but have attempted to present a different way of looking at possible interventions, one that perhaps corresponds more closely to problems perceived by farmers. Whether or not a particular technique is adopted will depend on the balance between the perceived beneﬁts and the costs of obtaining these beneﬁts, particularly, but not exclusively, in terms of the land, labour, and capital that is required. Although the abilities of a technique to meet researchers’ expectations, such as soil fertility enhancement or better weed control, is important, these abilities are not necessarily how the farmers value them. This approach also allows the consideration of other options besides natural resource management techniques. For example, a cash-generating activity might be for some of the household members to seek work in a local town or abroad. In many cases, this could be a better option than trying to grow a cash crop for this purpose, as returns

Table IV Possible Approach to Addressing Relevant Problems of Subsistence Farmers and Matching of LEIA Techniques and Practices to Problem Solutions Problem being addressed More food for the household

Solution Increased yields Extra food source

More cash generated for the household

Increase sales of surplus produce

Enhanced quality of life for members of the household

Reduce labour

More varied diet Ease of cultivation Fuel source Aesthetic value

Technique Improved varieties, intercropping, animal manure, composting Multipurpose trees, cover crops, intercropping Enriched fallow, cover crops, animal manure, multipurpose trees, composting, crop residues Improved fallows

Crop diversiﬁcation, cover crops, multipurpose trees Animal manure, cover crops Animal manure, multipurpose trees, crop residues Tithonia hedgerows

LOW EXTERNAL INPUT TECHNOLOGIES

533

to labour may be greater. Thus, not only is there a much wider range of possible solutions that can be evaluated, but also any improved solutions developed are much more likely to be adopted by farmers. Gender issues are also important. Women make up a sizeable proportion of tropical farmers, and it is they who most often focus on subsistence crops, generally using lower inputs of organic and inorganic fertilisers than men (Gladwin et al., 1997). In a study on constraints faced by women using organic agriculture, Gladwin et al. (1997) found that lack of capital prevented them from investing in either organic or inorganic fertilisers, lack of land limited their use of low-input organic techniques, and lack of labour limited their ability to undertake the activities that were required to implement such techniques, particularly as most women were also solely responsible for household duties and child care. In Senegal, women have even less rights to long-term use of land than do men (Golan, 1994), and therefore have no incentive to make long-term investments to improve soil productivity. Another gender-related problem that may act against the uptake of any improved technique is that additional incomes arising from sales of produce may go directly to the men in households, who are less likely than women to invest in children and the household as a whole (Pretty and Koohafkan, 2002). Cultural traditions may also restrict the uptake and use of particular techniques, despite any other beneﬁts they may have. For example, in relation to the potential use of animal manure in Ghana, Kiff et al., (1997) found from surveys in a number of villages that farmers knew that manure could be used, but generally found its use unattractive due to its supply being unreliable, too much effort involved in its collection, and because they felt that manuring techniques were regressive and old-fashioned. In certain areas, there is also the cultural problem of persuading farmers who have no tradition of cattle husbandry to develop the knowledge and interest to keep them on their farms (Dickson and Benneh, 1995). In Nepal on the other hand, there is a long tradition of keeping livestock, and the integration of animal manure into farm nutrient management is well developed. Animals are multifunctional in Nepalese agricultural systems, and provide meat, milk, ghee, curd, and draught power, as well as manure. Many families may own more than a single species of animal, with the most common combination (60% of those owning animals) being cattle, goats, chickens and buffaloes (Pilbeam et al., 1999b). Often there are competing demands for manure produced by livestock, particularly for use as fuel or in construction, which may make its availability as a nutrient source scarce. The use of the livelihoods-oriented approach described above may help to identify the real limitations of agricultural production systems more clearly. For example, the destruction of primary forest at the forest margins in Bolivia and Brazil is driven by other more powerful factors than soil fertility or weed encroachment issues. There, the underlying causes are economic or political in nature — aided by government subsidies, wealthy landowners buy out the

534

A. GRAVES, R. MATTHEWS AND K. WALDIE

frontier colonists to obtain land for cattle ranching, so that the latter then move on and clear new land. Converting primary forest to pasture is therefore an effective “cash generating” option for resource poor farmers that ﬁts neatly within wider economic and political realities. Consequently, the introduction of LEIA soil fertility and weed management techniques in an attempt to stabilise agricultural production at the forest margin, whilst technically feasible, is rendered almost irrelevant by these broader political and economic factors (Muchagata, 1997).

C. MODELLING To integrate all the various enterprises on a farm and to understand the often complex differences between farms and farmers, tools are needed. The SL framework just described is a useful start in this direction, but is limited in that it does not capture the dynamic and spatial nature of farming systems, particularly in relation to nutrient ﬂows (Scoones and Toulmin, 1998). Simulation modelling is a tool that offers the potential to integrate knowledge from a range of different disciplines and to explore the complex relationships between them in a dynamic and spatial way. For this reason, we believe that it is important that effort is made in developing simulation models of subsistence agricultural systems so that the processes involved are made explicit and to identify gaps in our knowledge. Because of the long-term nature of many of the underlying processes, such modelling offers a cost-effective and relatively quick way of obtaining answers to questions regarding potential interventions. Such models could also help to explore some of the wider global issues such as climate change, deforestation, and desertiﬁcation from the livelihoods perspective discussed above. The type of modelling we propose to be the most appropriate for this task is an integration of the key biophysical and socio-economic processes at the level of a household. A large number of household models incorporating these aspects already exist (e.g., de Jager et al., 1998; Pagiola and Holden, 2001), but most of these are static models providing only a snapshot of the state of a household at an instant in time and do not capture the dynamic characteristics of household activities (Scoones and Toulmin, 1998). In a recent attempt to make progress in this area, Shepherd and Soule (1998) developed a farm simulation model to assess the long-term impact of existing soil management strategies on the productivity, proﬁtability, and sustainability of farms in west Kenya. The model ran in time-steps of one year and linked soil management practices, nutrient availability, crop and livestock productivity, and farm economics. Crop types included weeds, fodder, grass, shrubs, and two types of grain crops. Growth of the plants was determined by N and P availability. A wide range of soil management options were simulated, including crop residue and manure management, soil erosion control measures, biomass transfer, improved fallow, green manuring, crop rotations, and N and P fertiliser application.

LOW EXTERNAL INPUT TECHNOLOGIES

535

The model was used to examine the sustainability of existing farming systems in the Vihiga district in western Kenya. Three household types were simulated to represent the range of resource endowments of households in the study area. Data to initialise the model for the different household types were collected through participatory research in the area, which indicated large differences in farm size, quantity and quality of livestock, soil and plant management, food consumption patterns, and sources of income. It was shown that the low and medium resource endowment farms had declining SOM, negative C, N, and P budgets, and low productivity and proﬁtability. The high resource endowment farms, on the other hand, had increasing SOM, low soil nutrient losses, and were productive and proﬁtable. This disaggregation into household types according to resource endowments highlighted the dangers of relying on nutrient balances of an “average” farm type — most previous studies in Africa have generally shown negative nutrient balances using this approach. Nevertheless, in this particular case, the low and medium resource households represented around 90% of the total in the study area, suggesting that overall a negative nutrient budget was likely. There was also the question of where the increasing nutrients of the high resource households were coming from — their greater purchasing power may have meant that there was nutrient ﬂow from the poorer households to the richer ones. Shepherd and Soule (1998) concluded that the ability of the high resource households to manage their farms proﬁtably and sustainably indicates that it is possible, but that capital is required. Strategies they suggested to improve livelihoods included: (1) an increase in the value of farm input, (2) an increase in high quality nutrient inputs at low cash and labour costs to the farmer, and (3) an increase in off-farm income. In another example, as part of a project evaluating integrated agriculture – aquaculture farming systems in the Philippines, Schaber (1997) developed a whole-farm model called FARMSIM to quantify ﬂows of nutrients between the different farming enterprises. The ORYZA_0 rice simulation model was combined with a ﬁsh-pond model (for Nile Tilapia Oreochromis niloticus (L.)) and models of pigs, chickens and buffaloes. After the rice was harvested, the straw was assumed to be composted, the bran to be fed to the pigs and to the ﬁsh, and broken rice to be used as chicken food. If this supply of food was less than the demand, then more had to be purchased externally. Weeds from the ﬁeld bunds were fed to the buffaloes. Manure from the pigs and buffaloes was fed to the ﬁsh, although buffalo manure could also be applied to the rice ﬁelds. The model was then used to evaluate three different scenarios in terms of the efﬁciency of N use (deﬁned as the output N as a ratio of the input N) of the farm as a whole. In the ﬁrst scenario, a conventional farming system with a mono-cultural rice ﬁeld, two buffaloes and ﬁfteen chickens was assumed. High levels (200 kg N ha21) of commercial fertiliser were applied to the rice ﬁeld, and the output of rice grain was high. In the second scenario, diversiﬁcation increased as pigs were introduced. The third scenario represented a fully integrated, diversiﬁed farming

536

A. GRAVES, R. MATTHEWS AND K. WALDIE

system, including a ﬁsh-pond, with all reusable products from each farm enterprise being used as inputs to other enterprises where appropriate. The predicted efﬁciencies of N use in each of the three scenarios were 13.3, 18.7, and 21.6% respectively. It was concluded that the greater the number of enterprises there were on the farm, the greater the efﬁciency of N use of the farm as a whole. A similar household model is currently being developed as a tool to help evaluate the relevance of potential soil fertility-enhancing techniques to livelihoods of farmers in the mid-hills of Nepal (Matthews, 2000). In addition to the biophysical processes of crop and animal growth, and water and nitrogen ﬂuxes through the household, economic and labour ﬂows have also been incorporated, along with household resources such as food, money, manure, fodder, and fertiliser. The model, therefore, incorporates elements of the natural, human, and ﬁnancial capitals in the SL framework described above. Various types of household can be accommodated, ranging from resource poor to resource rich. The model will be used in the ﬁrst instance to evaluate potential interventions in the existing system and the likelihood of uptake of these interventions, using criteria such as their contribution to household ﬁnances, food production, alleviation of risk, and labour demands in relation to other farm enterprises. So far, these models are not spatially explicit, but they do have the capability to become so. This is important because, as we have already discussed, the spatial relationship between different parts of the landscape is often central to the functioning of a farming system, with some areas acting as nutrient sources and others as nutrient sinks. The net movement of nutrients from ﬁelds far from the household to those close by on smallholder farms in Uganda (Wortmann and Kaizzi, 1998; Briggs and Twomlow, 2002) has already been mentioned, while in Burkino Faso, Prudencio (1993) found that 85% of household manure was applied to nearby plots and only 15% to distant plots. Similarly, farmers often consciously manipulate erosion and run-off from dry toplands as a means of concentrating moisture and nutrients on the wetter and more accessible bottomlands where they can beneﬁt from them (Scoones and Toulmin, 1998). Thus, extrapolation of results from measurements at a single site can be highly misleading. The household models described above could take this into account by incorporating a number of associated land management units (e.g., ﬁelds) that are linked spatially, both in the horizontal and vertical directions if necessary. Heterogeneity at the village level could then be described by linking a number of such individual household models together (Matthews and Stephens, 2002). Such models would then provide insights into the movement of nutrients within the landscape, and possibly suggest ways in which this could be optimised to beneﬁt people’s livelihoods. At some point in the future, these models should also incorporate household decision-making processes, based on a labour and economic analysis each year of the various household enterprises (crops, livestock, off-farm work, etc.), also

LOW EXTERNAL INPUT TECHNOLOGIES

537

taking into account subsistence needs and attitude to risk. However, considerable thought needs to be given to the dynamic processes involved in household decision-making, and how these are inﬂuenced by the biophysical environment. Some progress has been made by Pagiola and Holden (2001) and Angelsen and Kaimowitz (2001) in determining when forest clearing is likely to be a rational decision for farmers. The emerging ﬁeld of multi-agent systems simulation, in which artiﬁcial intelligence approaches are incorporated and interactions between individuals are central (e.g., Bousquet et al., 1999), may be one way in which further progress can be made in this direction. This is one area where multidisciplinary research involving biophysical scientists and social scientists is likely to be fruitful. It is important to emphasise that while such models cannot be used to predict the behaviour of speciﬁc households precisely, they are useful as tools for understanding and testing hypotheses regarding the processes involved in interactions between the biophysical and socio-economic environments of subsistence farmers, and how these processes relate to their livelihoods and poverty. Exploration of viable pathways out of poverty is more important than the prediction of ﬁnal endpoints. In the context of LEIA technologies, some of the types of questions that can be addressed with such models are as follows: 1. The potential of LEIA techniques: While LEIA techniques such as the ones we have discussed in this review can make a useful contribution to maintenance of soil fertility, they are unable to supply enough nutrients required to achieve the genetic potential of high-yielding crops. However, it would be useful to know what level of crop yields could be sustained by the sole use of such techniques in different environments and contexts, and how farmer livelihoods are affected by this. 2. Evaluation of fallow types: Natural fallows offer a way of regenerating soil fertility, but land must be set aside for long periods of time. Where land is relatively plentiful, natural fallowing is a rational strategy. However, where population increases and the availability of land decreases, the opportunity cost of setting aside land for long periods of time rises signiﬁcantly. Improved fallows may be able to speed up the regeneration process, but at what level of land availability does it become worthwhile for a farmer to consider the technique? Similar questions can be asked for enriched fallows, where the regenerative process is accompanied by income generation, taking into account the possibly slower regeneration rate due to removal of harvested material. Experimental determination of these issues is time-consuming and expensive. 3. Managing variation in natural resources: By concentrating resources in one area at the expense of another, higher value crops may be grown, leading to an improvement in cash income for the household, some of which could be re-invested in the poorer areas of the farm, thereby improving the overall

538

A. GRAVES, R. MATTHEWS AND K. WALDIE

fertility of the farm in the long term. Many of the LEIA systems we have discussed (e.g., the Machobane system, SRI, the Zai system) essentially concentrate nutrients from a wider area into a smaller area that can be cropped. What is the optimum way to do this for a given biophysical and socioeconomic situation? 4. Effect of a change in farmer perceptions: Recent work in Ghana has shown that in evaluating different practices, farmers do not always value the opportunity cost of their own labour, though they more readily assess the ﬁnancial cost of hiring others to work on their farms (Galpin et al., 2000). Participatory interaction, however, has brought some of them to consider that their own time and labour should be factors taken account of in the evaluation. It would be interesting to compare the likelihood of adoption of various techniques (both traditional and researcher-generated) with and without consideration of the labour involved. Would patterns of development be different in each case? Do more sustainable practices result from taking labour into account? Or is the concept of opportunity cost of labour meaningless when there are so few options available in which it could be deployed, anyway? 5. Effect of current socio-economic trends: In Nepal in recent years, a decline in soil fertility has been ascribed by farmers themselves to a decline in manure applications, due in turn to a decline in livestock numbers brought about by a reduction in the household labour pool with more and more children going to school (Ellis-Jones, pers. comm.). School leavers are not interested in returning to work on the farm, preferring to ﬁnd jobs in the towns and cities. What effect is this likely to have on the fertility of the soil in the ﬁrst instance, and on the overall livelihood of the household, bearing in mind that urban jobs represent a potential source of cash income for the household in the future? Is it a good livelihood strategy to invest in the education of one’s children, and at what cost is this to the biophysical environment? Should government policies aim to encourage the educated to take up farming, or is it desirable that hill agriculture continues to decline? 6. Trajectories out of poverty: The question of whether there are “natural” processes (in the broadest sense, including both biophysical and socioeconomic processes) that can lead to the evolution of one agricultural system into another needs to be explored. Given that it is perfectly rational for poor people to adopt short-term strategies that attempt to maximise their livelihood outcomes, can improved (or even new) strategies be developed or promoted to hasten the change from shifting cultivation systems to more settled patterns of agriculture? Can LEIA techniques improve livelihoods even if they are used efﬁciently, or are external inputs essential? Could governments adopt particular policies that would facilitate the transition process? How is the distribution of wealth inﬂuenced by different processes of transition? What are the long-term environmental consequences of such transitions?

LOW EXTERNAL INPUT TECHNOLOGIES

539

V. CONCLUDING REMARKS For any agricultural system to be sustainable, regardless of whether it is classiﬁed as LEIA or HEIA, inputs must not be less than outputs — if they are, then mining of existing stocks occurs, and the system must deteriorate in time. Where there is constant removal of nutrients in crops harvested for human consumption and these are not returned to the system, then no such system can be sustainable without some further input from outside. No group of technologies, whether based on organic or inorganic sources of nutrients, can contravene this basic law of mass conservation. In this review, we have attempted to analyse the biophysical and socioeconomic characteristics of a number of LEIA techniques that have been evaluated as potential improvements to subsistence agricultural systems. These techniques included intercropping, alley cropping, cover crops and green manures, biomass transfer, compost, use of animal manure, and improved and enriched fallows. It is important to remember that farmers have been practising variations of these techniques for generations, using organic materials that have been available within their immediate vicinity. While population density has been low, and land has been abundant, these systems have been sustainable, but with the population explosion in the 20th century, the parameters have now changed, and there is no guarantee that these systems will continue to function in a sustainable way. Indeed, many are beginning to break down as cropping periods are extended and fallow periods are shortened (e.g., Chidumayo, 1987). There is no doubt that there is an urgent need to ﬁnd ways to make these systems more productive and sustainable, since it is upon the outputs of such systems that the lives and livelihoods of so many of the world’s poor ultimately depend. While there may be some scope in making existing systems more efﬁcient by reducing losses of nutrients, identifying new sources of organic material, or spatially manipulating nutrient concentrations within farms (Wortmann and Kaizzi, 1998; Briggs and Twomlow, 2002), most of the LEIA techniques we have reviewed, when used alone, appear to have limited potential to increase food production dramatically. In the case of those techniques aimed at maintaining or improving soil fertility, the nutrient content and the quantity of biomass that can be produced within the resources available to such farmers is insufﬁcient to meet the requirements of most crops, certainly at a reasonable yield level, although where existing yields are very low, substantial relative yield increases may be possible (e.g., Pretty et al., 2003). The contribution that these large relative, but small absolute, increases in yields make to overall global food production needs to be determined by weighting each yield level with an appropriate land area. Similarly, those LEIA techniques aimed at weed control in the studies reviewed, while often being able to reduce weed populations, had little effect on crop yields. In both cases, there is also the possibility that crop yields may actually be depressed,

540

A. GRAVES, R. MATTHEWS AND K. WALDIE

either as a result of competition for resources, particularly in difﬁcult conditions, or as a result of N or P immobilisation if low-quality organic material is incorporated into the soil. While LEIA technologies are often promoted as “natural” (e.g., Pretty, 1999), as we have seen, in reality many of them have very high demands on land and labour. Thus, the debate over the relative merits of LEIA and HEIA technologies essentially hinges on whether the inputs should be capital intensive or land and/or labour intensive. Ironically, all of these are assets that by deﬁnition, resource-poor farmers have in short supply. Any hopes of promotion of their use to bring about radical changes to existing cultivation systems to address the problem of future global food demand, and to provide impetus to lift such farmers out of the poverty spiral, is, therefore, likely to be more utopian than realistic. A similar conclusion was also reached by Sanchez et al. (2001), who argued that sole use of LEIA technologies is only likely to perpetuate food insecurity and poverty. The future, therefore, must surely lie in the integration of the two approaches, using LEIA technologies when organic sources of nutrients are available, but also being prepared to supplement these with external supplies if necessary and when it is economic to do so. Taking a systems perspective is essential — there is no doubt that higher inputs of either inorganic or organic forms of nutrients can increase crop yields of individual plots or ﬁelds — the real issue, however, is whether the supplies of these inputs are sustainable at higher scales, and whether there is sufﬁcient labour and capital in the system for their transport and handling. There is little point in increasing crop yields per se if these cannot be feasibly scaled up or maintained at that level for long. This applies to both LEIA and HEIA technologies — while promotion of the use of external supplies of fertiliser has been rightly criticised for being out of reach for many subsistence farmers, so might the large amounts of land required to provide sufﬁcient nutrients in biomass, or the large amounts of labour at the right time to harvest, transport and incorporate this biomass. There is a tendency by researchers and proponents of both LEIA and HEIA approaches to quote only yield increases and to ignore the overall constraints of the system required to produce these increases (e.g., Uphoff, 1999; Singh and Sharma, 2001), which can be misleading. We would make a plea, therefore, that as well as providing adequate biophysical information on soils and climate, all future reports of yield increases in LEIA techniques state the area of land on which the crop yield has been measured, accompanied by estimates of the amount of land from which the particular technique has gathered its nutrients from, and the amount of labour required to harvest and transport these nutrients to the cropped area. Similar analyses should be made for existing farmer practice, and for HEIA techniques in terms of capital costs of purchasing, transporting, and applying the external inputs. The economic costs and beneﬁts of the technique should also be determined, preferably dynamically, so that it is possible to see whether resource demand matches resource availability throughout the year. In this way, a more objective appraisal of all techniques may be possible.

LOW EXTERNAL INPUT TECHNOLOGIES

541

Future research to improve subsistence agriculture needs, therefore, to focus more on developing interventions to meet farmer realities, such as increased food security, improved cash generation, reduced risk, and enhanced quality of life. This may necessitate consideration of a broader range of interventions than just natural resource management options. In this process, both “holistic” and “reductionist” approaches are necessary and valuable in an iterative process of investigation and analysis. The starting point should be from a holistic viewpoint, the analysis of problems in the system and development of solutions should be reductionist, and any successful solutions to the problems should be evaluated holistically again. Many of the processes, both biophysical and socio-economic and their interactions, are poorly understood, and it is essential that future research addresses this. This lack of knowledge is compounded by the large degree of heterogeneity of the production systems involved, both at the system level with different cultivation systems in the different countries, and also at the individual farm level with between-farm variability in terms of farmer aspirations and attitudes, and withinfarm variability in resources. However, it is important that this heterogeneity is preserved as it contributes to the resilience, and hence the sustainability, of the production system. This was highlighted in the example of the maize – Mucuna system in Honduras in which over-reliance on one cover crop species was probably a major factor allowing the incursion of Rottboellia as Mucuna as more effective against broad-leaf weeds than grasses (Neill and Lee, 2001). Of all of the LEIA and HEIA technologies on offer, there is no single one that is a panacea for the problems faced by subsistence farmers. Each has particular strengths and weaknesses, and the challenge is to identify these strengths and combine them into integrated systems capable of adapting to changing circumstances when necessary.

ACKNOWLEDGMENTS We are grateful to the Natural Resources Systems Programme of the United Kingdom’s Department for International Development (DFID) for funding this work. The review is condensed and updated from the Final Report of DFID Project R7560.

REFERENCES Adeorike, V., Ogburia, M. N., and Anegbeh, P. (2001). Evaluation of the allelopathic inﬂuence of selected multipurpose tree species on maize (Zea mays) under a simulated ﬁeld condition. Tropicultura 19(4), 191 –193.

542

A. GRAVES, R. MATTHEWS AND K. WALDIE

Adeoye, G. O., Sridhar, M. K. C., and Oyenekan, O. T. (1996). A survey of compost utilization for ornamental plants in Nigeria. Compost Sci Utilization 4(4), 81–83. Adesina, A., and Coulibaly, N. (1998). Policy and competitiveness of agroforestry-based technologies for maize production in Cameroon: an application of policy analysis matrix. Agric. Econ. 19, 1–13. Akinyemi, J. O., and Kalejaiye, A. S. (1998). The strategies for implementing the soil conservation practices among the rural farmers in the tropical rain forest zone of Nigeria. Int. J. Tropic Agric. 14(1–4), 35–48. Akobundu, I. O., Udensi, U. E., and Chikoye, D. (2000). Velvetbean (Mucuna spp.) suppresses speargrass (Imperata cylindrica (L.) Raeuschel) and increases maize yield. Int. J. Pest Manag. 46(2), 103–108. Akyeampong, E. (1999). Suitability of six hedgerow species for alley cropping on an acid infertile soil in Burundi. Tropic. Agric. 76(2), 82 –87. Ali, M. (1996). Quantifying the socio-economic determinants of sustainable crop production: an application to wheat cultivation in the Tarai of Nepal. Agric. Econ. 14(1), 45–60. Ali, M. (1999). Evaluation of green manure technology in tropical lowland rice systems. Field Crops Res. 61, 61 –78. Amede, T., Belachew, T., and Geta, E. (2001). Reversing the degradation of arable land in the Ethiopian Highlands. Managing Africa’s Soils 23, 23. Angelsen, A., and Kaimowitz, D. (2001). When does technological change in agriculture promote deforestation? In “Tradeoffs or Synergies? Agricultural Intensiﬁcation, Economic Development, and the Environment” (D. R. Lee and C. B. Barrett, Eds.), pp. 89 –114. CAB International, Wallingford, UK. Anoka, U. A., Akobundu, I. O., and Okonkwo, S. N. C. (1991). Effects of Gliricidia sepium and Leucaena leucocephala on growth and development of Imperata cylindrica. Agrofor. Syst. 16, 1–12. Apiradee, K. (1988). Factors affecting the adoption of farmers’ compost making and application in Amphoe Khao Yoi, Changwat Phetchaburi, Thailand. Kaesart University, Bangkok. Araki, S. (1993). The effect of burning, ash fertilisation, and soil organic matter on productivity of the chitemene shifting cultivation in northern Zambia. In “Soil Organic Matter Dynamics and Sustainability of Tropical Agriculture” (K. Mulongoy and R. Merckz, Eds.), pp. 367–375. John Wiley & Sons, New York. Ashley, C., and Carney, D. (1999). Sustainable Livelihoods: lessons from early experience, p. 55. Department for International Development, London, UK. Atta-Krah, A. N., and Sumberg, J. E. (1988). Studies with Gliricidia sepium for crop/livestock production systems in West Africa. Agrofor. Syst. 6, 97 –118. Ayuk, E. T. (2001). Social, economic and policy dimensions of soil organic matter management in sub-Sahara Africa: challenges and opportunities. Nutrient Cycling in Agroecosystems 61, 183–195. Azam-Ali, S. N., Matthews, R. B., Williams, J. H., and Peacock, J. M. (1990). Light use, water uptake and performance of individual components of a sorghum/groundnut intercrop. Exp. Agric. 26, 413–427. Baggio, A., and Heuveldorp, J. (1988). Initial performance of Calliandra calothyrus Messn. in live fences for the production of biomass. Agrofor. Syst. 2, 19–29. Barrios, E., Kwesiga, F., Buresh, R. J., and Sprent, J. I. (1997). Light fraction soil organic matter and available nitrogen following trees and maize. Soil Sci. Soc. Am. J. 61, 826–831. van den Berghe, C., Sota, P., and Mujawayezu, A. (1994). Study on mineral complementation in the composting process in farmers conditions of Mugamba, Burundi. Tropicultura 12(2), 55– 60. Blarel, B. (1994). Tenure and agricultural production under land scarcity: the case of Rwanda. In “Searching for Land Tenure Security in Africa” (J. W. Bruce and S. N. Migot-Adholla, Eds.), pp. 71– 96. Kendall/Hunt Publishing Company, Dubuque, Iowa, USA.

LOW EXTERNAL INPUT TECHNOLOGIES

543

Boddey, R. M., de Sa, J. C. M., Alves, B. J. R., and Urquiaga, S. (1997). The contribution of biological nitrogen ﬁxation for sustainable agricultural systems in the tropics. Soil Biol. Biochem. 29(5/6), 287 –799. Bouldin, D. R., Mughogho, S., Lathwell, D. J., and Scott, T. W. (1979). Nitrogen Fixation by Legumes in the Tropics. (International Agricultural Mimeograph). Report No. 75. Department of Agronomy, Cornell University, New York. Bousquet, F., Barreteau, O., Le Page, C., and Weber, J. (1999). An environmental modelling approach: the use of multi-agent simulations. In “Advances in Environmental and Ecological Modelling” (F. Blasco and A. Weill, Eds.), pp. 113–122. Elsevier, Paris, France. Brady, N. (1990). In “The Nature and Properties of Soils”. Macmillan Publishing Company, New York. Briggs, L., and Twomlow, S. J. (2002). Organic material ﬂows within a smallholder highland farming system of south west Uganda. Agric. Ecosyst. Environ. 89, 191–212. Brodd, P., and Osanius, M. (2002). Erosion and land use at plot and catchment levels — a case study in northern Vietnam, p. 62. Minor Field Studies International Ofﬁce, Swedish University of Agricultural Sciences, Uppsala, Sweden. Buckles, D., and Triomphe, B. (1999). Adoption of mucuna in the farming systems of northern Honduras. Agrofor. Syst. 47, 67 –91. Budelman, A. (1988). Leaf dry matter productivity of three selected perennial leguminous species in humid tropical Ivory Coast. Agrofor. Syst. 7, 42–62. Buresh, R. J., and Niang, A. I. (1997). Tithonia diversifolia as a green manure: awareness, expectations and realities. Agrofor. Forum 8(3), 29–31. Cairns, M., and Garrity, D. P. (1999). Improving shifting cultivation in Southeast Asia by building on indigenous fallow management strategies. Agrofor. Syst. 47, 37–48. Carney, J. (1993). Low input sustainable agriculture in Cuba. Agric. Human Values 10(3) (Special Issue). Carney, D. (1998). Implementing the sustainable rural livelihoods approach, Sustainable Rural Livelihoods — what contribution can we make?, pp. 3–23. Department for International Development, London, UK. Carter, J. (1995). Alley farming: have resource-poor farmers beneﬁted? Natural Resources Perspectives No. 3. Overseas Development Institute, London. Catchpoole, D. W., and Blair, G. J. (1990). Forage tree legumes. III. Release of nitrogen from leaf, faeces and urine derived from Leucaena and Gliricidia leaf. Aust. J. Agric. Res. 41, 539–547. Cenas, P. A. A., Pandey, S., and Pandey, S. (1996). Determinants of adoption and retention of contour hedgerows in Claveria, Misamis Oriental. Philippine J. Crop Sci. 21(1–2), 34 –39. Chianu, J. N., Akintola, J. O., and Kormawa, P. M. (2002). Proﬁtability of cassava-maize production under different fallow systems and land-use intensities in the derived savanna of southwest Nigeria. Exp. Agric. 38(1), 51 –63. Chidumayo, E. N. (1987). A shifting cultivation land use system under population pressure in Zambia. Agrofor. Syst. 5, 15 –25. Chivasa, W., Harris, D., Chiduza, C., Nyamudeza, P., and Mashingaidze, A. B. (2000). Biodiversity on-farm in semi-arid agriculture: a case study from a smallholder farming system in Zimbabwe. Zimbabwe Sci. News 34(1), 13–18. Coppock, D. L., and Reed, J. D. (1992). Cultivated and native browse legumes as calf supplements in Ethiopia. J. Range Manag. 45, 231 –238. Coultas, C. L., Post, T. J., Jones, J. B. Jr, and Hsieh, Y. P. (1996). Use of velvet bean to improve soil fertility and weed control in corn production in northern Belize. Commun. Soil Sci. Plant Anal. 27(9– 10), 2171–2196. Craswell, E. T., Sajjapongse, A., Howlett, D. J. B., and Dowling, A. J. (1998). Agroforestry in the management of sloping lands in Asia and the Paciﬁc. Agrofor. Syst. 38, 121–137. Crosson, P., and Anderson, J. R. (2002). Technologies for meeting future global demand for food. Discussion Paper 02-02. Resources for the Future, 1616 P Street, NW, Washington D.C. 20036.

544

A. GRAVES, R. MATTHEWS AND K. WALDIE

Dalland, A., Va˚je, P. I., Matthews, R. B., and Singh, B. R. (1993). The potential of alley cropping in improvement of cultivation systems in the high rainfall areas of Zambia. III. Soil chemical and physical properties. Agrofor. Syst. 21, 117 –132. Dasthagir, M. G., Kohli, R. K., Arya, K. S., and Atul (1996). Review of research on tropical agroforestry. In “IUFRO-DNAES International Meeting; Resource Inventory Techniques to Support Agroforestry and Environment”. pp. 306 –309. DAV College, Chandigarh, India. David, S. (1995). What do farmers think? Farmer evaluations of hedgerow intercropping under semiarid conditions. Agrofor. Syst. 32, 15–28. Degrande, A., and Duguma, B. (2000). Adoption potential of rotational hedgerow intercropping in the humid lowlands of Cameroon, p. 8. Agricultural Research and Extension Network, Network Paper No. 103. Overseas Development Institute, London, UK. Delve, R. J., Cadisch, G., Tanner, J. C., Thorpe, W., Thorne, P. J., and Giller, K. E. (2001). Implications of livestock feeding management on soil fertility in the smallholder farming systems of sub-Saharan Africa. Agric. Ecosyst. Environ. 84, 227–243. Deugd, M., Ro¨ling, N., and Smaling, E. M. A. (1998). A new praxeology for integrated nutrient management, facilitating innovation with and by farmers. Agric. Ecosyst. Environ. 71, 269– 283. Dewees, P. A., and Saxena, N. C. (1995). Tree planting and household land and labour allocation: case studies from Kenya and India. In “Tree Management in Farmer Strategies: Responses to Agricultural Intensiﬁcation” (J. E. M. Arnold and P. A. Dewees, Eds.), pp. 198– 241. Oxford University Press, Oxford. Dickson, K. B., and Benneh, G. (1995). In “A New Geography of Ghana”. Longman, Harlow. Diop, A. M. (1999). Sustainable agriculture: new paradigms and old practices? Increased production with management of organic inputs in Senegal. Environ. Develop. Sustainability 1, 285–296. Douthwaite, B., Manyong, V. M., Keatinge, J. D. H., and Chianu, J. (2002). The adoption of alley farming and Mucuna: lessons for research, development and extension. Agrofor. Syst. 56(3), 193–202. Dowling, N. G., Greenﬁeld, S. M., and Fischer, K. S. (1998). Sustainability of Rice in the Global Food System, p. 404. Paciﬁc Basin Study Centre, Davio, California, and International Rice Research Institute, Manila. Drechsel, P., and Reck, B. (1998). Composted shrub-prunings and other organic manures for smallholder farming systems in southern Rwanda. Agrofor. Syst. 39, 1–12. Drechsel, P., Steiner, K. G., and Hagedorn, F. (1996). A review on the potential of improved fallows and green manure in Rwanda. Agrofor Syst. 33(2), 109 –136. Durban Metropolitan, (1999). Durban Metro State of the Environment and Development: 1999 (Accessed 12-Dec-2002 online at http://www.ceroi.net/reports/durban/issues/waste/state.htm). Elbasha, E., Thornton, P. K., and Tarawali, G. (1999). An ex post economic impact assessment of planted forages in West Africa. ILRI Impact Assessment Series, No. 2. International Livestock Research Institute (ILRI), Nairobi. Ellis, F. (1988). In “Peasant Economics: Farm Households and Agrarian Change”. Cambridge University Press, Cambridge. Enyong, L. A., Debrah, S. K., and Bationo, A. (1999). Farmers perceptions and attitudes towards introduced soil fertility enhancing technologies in western Africa. Nutrient Cycling in Agroecosystems 53, 177 –187. Euroconsult, (1989). “Agricultural Compendium for Rural Development in the Tropics and Subtropics”. The Netherlands Ministry of Agriculture and Fisheries, Amsterdam. FAO, (1996). “Food Requirements and Population Growth”. Food and Agriculture Organisation, Rome, Italy. Farrell, R. E., Sandercock, P. J., Pennock, D. J., and van Kessel, C. (1996). Landscape-scale variations in leached nitrate: relationship to denitriﬁcation and natural N15 abundance. Soil Sci. Soc. Am. J. 60, 1410–1415.

LOW EXTERNAL INPUT TECHNOLOGIES

545

Fassler, O. M., and Lascano, C. E. (1995). The effect of mixtures of sun-dried tropical shrub legumes on intake and nitrogen balance by sheep. Trop. Grasslands 29, 92–96. Field, S. P., Momuat, E. O., Ayre-Smith, R., and Bamualim, A. (1992). Developing land rehabilitation technologies compatible with farmer needs and resources. J. Asian Farm. Syst. Assoc. 1(3), 385 –394. Franzel, S. (1999). Socioeconomic factors affecting the adoption potential of improved tree fallows in Africa. Agrofor. Syst. 47, 305–321. Fujita, K., Matsumoto, K., Ofosu-Budu, G. K., and Ogata, S. (1993). Effect of shading on growth and dinitrogen ﬁxation of kudzu and tropical pasture legumes. Soil Sci. Plant Nutr. 39(1), 43–54. Gachengo, C., Palm, C., Adams, E., Giller, K. E., Delve, R. J., and Cadisch, G. (1998). Organic Resources Database. Tropical Soil Biology and Fertility Programme (TSBF), Nairobi and Wye College, University of London. Gachengo, C. N., Palm, C. A., Jama, B., and Othieno, C. (1999). Tithonia and Senna green manures and inorganic fertilizers as phosphorus sources for maize in western Kenya. Agrofor. Syst. 44(1), 21 –36. Galiba, M., Vissoh, P., Dagbenonbakin, G., and Fagbohoun, F. (1998). The reactions and fears of farmers utilizing velvet bean (Mucuna pruriens). Cover crops in West Africa: contributing to sustainable agriculture. In “Proceedings of a conference 1–3 October 1996, Cotonou, Benin” (D. Buckles, A. Eteka, O. Osiname, and G. Galiano, Eds.), pp. 55–64. International Development Research Centre (IDRC), Ottawa, Canada. Gallagher, R. S., Fernandes, E. C. M., and McCallie, E. L. (1999). Weed management through shortterm improved fallows in tropical agroecosystems. Agrofor. Syst. 47, 197–221. Galpin, M., Dorward, P., and Shepherd, D. (2000). Participatory farm management methods for agricultural research and extension: a training manual. Agricultural Extension and Rural Development Department, Reading University, Reading, UK. Ganry, F., Feller, C., Harmand, J. M., Guibert, H., Martius, C., Tiessen, H., and Vlek, P. L. G. (2001). Management of soil organic matter in semiarid Africa for annual cropping systems. Nutrient Cycling in Agroecosystems 61(1– 2), 105–118. Garcia, J. N. M., Gerrits, R. V., and Cramb, R. A. (2002). Adoption and maintenance of contour bunds and hedgerows in a dynamic environment: experience in the Philippine uplands. Mountain Res. Develop. 22(1), 10 –13. Garrity, D. P., and Lai, C. K. (2000). Innovation in shifting cultivation in Asia: indigenous fallow management. Ileia Newslett. 16(3), 5–6. Gashaw, G., Varvikko, T., and Khalili, H. (1991). Smallholder Dairy Production in Northern Shoa Province, Central Highlands of Ethiopia. Assessment of Feed Resources in Relation to the Requirements on a Smallholder Farm (Final Report), p. 25. International Livestock Centre for Africa (ILCA), Addis Ababa. George, T. S., Gregory, P. J., Robinson, J. S., Buresh, R. J., and Jama, B. A. (2001). Tithonia diversifolia: variations in leaf nutrient concentration and implications for biomass transfer. Agrofor. Syst. 52(3), 199–205. Gill, H. S. (1982). The role of hand and mechanical weeding in weed management in the advancing countries. Improving Weed Management (Plant Production and Protection Paper 44). FAO, Rome, Italy. Giller, K. E. (2001). “Nitrogen Fixation in Tropical Cropping Systems”, p. 313. CAB International, Wallingford, UK. Giller, K. E., and Cadisch, G. (1995). Future beneﬁts from biological nitrogen ﬁxation: an ecological approach to agriculture. Plant Soil 174, 255 –277. Giller, E. K., Cadisch, G., Ehaliotis, C., Adams, E., Sakala, W. D., and Mafongoya, P. L. (1997). Building soil nitrogen capital in Africa. In “Replenishing Soil Fertility in Africa”. SSSA Special Publication No. 51 (R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 151 –192. Soil Science Society of America, Madison, Wisconsin.

546

A. GRAVES, R. MATTHEWS AND K. WALDIE

Girish, G. K., and Chandrashekar, C. P. (2000). Farmers practice of nutrient management for sugarcane in GLBC area of Karnataka — a status paper. Cooperative Sugar 31(6), 477 –484. Gladwin, H., Buhr, K. L., Goldman, A., Hiebsch, C., Hildebrand, P. E., Kidder, G., Langham, M., Lee, D., Nkedi-Kizza, P., and Williams, D. (1997). Gender and Soil Fertility in Africa. In “SSSA Special Publication No. 51: Replenishing Soil Fertility in Africa” (R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 219–236. Soil Science Society of America, Wisconsin. Golan, E. H. (1994). Land tenure reform in the Peanut Basin of Senegal. In “Searching for Land Tenure Security in Africa” (J. W. Bruce and S. N. Migot-Adholla, Eds.), pp. 231– 249. Kendall/ Hunt Publishing Company, Dubuque, Iowa, USA. Goyal, S., Mishra, M. M., Hooda, I. S., and Singh, R. (1992). Organic matter– microbial biomass relationship in ﬁeld experiments under tropical conditions: effects of inorganic fertilization and organic ammendments. Soil Biol. Biochem. 24, 1081–1084. Gregory, P. J., Ingram, J. S. I., Andersson, R., Betts, R. A., Brovkin, V., Chase, T. N., Grace, P. R., Gray, A. J., Hamilton, N., Hardy, T. B., Howden, S. M., Jenkins, A., Meybeck, M., Olsson, M., Ortiz-Monasterio, I., Palm, C. A., Payn, T. W., Rummukainen, M., Schulze, R. E., Thiem, M., Valentin, C., and Wilkinson, M. J. (2002). Environmental consequences of alternative practices for intensifying crop production. Agric. Ecosyst. Environ. 88, 279–290. Grist, P., Menz, K., and Nelson, R. (1999). Multipurpose trees as improved fallow: an economic assessment. Int. Tree Crops J. 10(1), 19–36. Harris, D., Joshi, A., Khan, P. A., Gothkar, P., and Sodhi, P. S. (1999). On-farm seed priming in semiarid agriculture: development and evaluation in maize, rice and chickpea in India using participatory methods. Exp. Agric. 35, 15–29. Harsono, A. (1996). Potency and opportunity of soybean farming system on upland in Sumbawa (NTB) and Sikka (NTT) districts. In “Proceedings of a Seminar on Legumes and Tuber Crops Research Results in 1995” (N. Marwoto, A. Saleh, Kasno, and Sunardi, Eds.). Malang, Indonesia. Hartemink, A. E., Buresh, R. J., Jama, B., and Janssen, B. H. (1996). Soil nitrate and water dynamics in Sesbania fallows, weed fallows and maize. Soil Sci. Soc. Am. J. 60, 568–574. Hauser, S. (1990). Water and nutrient dynamics under alley cropping versus monocropping in the humid–subhumid transition zone. Transactions of the 14th International Soil Science Society Congress, Kyoto, Japan, 204– 209. Hellin, J., Welchez, L. A., and Cherrett, I. (1999). The Quezungual system: an indigenous agroforestry system from western Honduras. Agrofor. Syst. 46(3), 229 –237. Hoang Fagerstro¨m, M. H., van Noordwijk, M., Thai, Phien, and Nguyen Cong, Vinh (2001). Innovations within upland rice-based systems in northern Vietnam with Tephrosia candida as fallow species, hedgerow, or mulch: net returns and farmers’ response. Agric. Ecosyst. Environ. 86(1), 23–37. Humphries, S. (1996). In Milk cows, migrants, and land markets: unravelling the complexities of forest to pasture conversion in northern Honduras. University of Guelph, Guelph, Ontario, Canada. ICRAF, (1997). Using the Wild Sunﬂower, Tithonia, in Kenya for Soil Fertility and Crop Yield Improvement. International Centre for Research in Agroforestry, Nairobi, Kenya. IIRR, (1998). The Machobane farming system in Lesotho, Sustainable Agricultural Extension Manual for Eastern and Southern Africa. pp. 171 –176. International Institute of Rural Reconstruction, Africa Regional Ofﬁce, PO Box 66873, Nairobi, Kenya. Jackson, D., Kiff, E., and Chan, M.-K. (1999). Integrated Food Crop Systems Project, Ghana: Development and Promotion of Improved Techniques of Water and Soil Fertility Management for the Sustainable Production of Crops on Land in the Humid Forest Belt. (Progress Against Milestones Review). Report No. 4242. Natural Resources Institute, London. Jadhao, S. L., Giri, D. G., Kharat, B. S., Deshmukh, J. P., and Kubde, K. J. (1999). Effect of integrated nutrient management on productivity of sorghum-wheat cropping system. Ann. Plant Physiol. 13(2), 128–132.

LOW EXTERNAL INPUT TECHNOLOGIES

547

de Jager, A., Nandwa, S. M., and Okoth, P. F. (1998). Monitoring nutrient ﬂows and economic performance in African farming systems (NUTMON) I. Concepts and methodologies. Agric. Ecosyst. Environ. 71, 37– 48. de Jager, A., Onduru, D., van Wijk, M. S., Vlaming, J., and Gachini, G. N. (2001). Assessing sustainability of low-external-input farm management systems with the nutrient monitoring approach: a case study in Kenya. Agric. Syst. 69, 99–118. Jama, B., Buresh, R. J., and Place, F. M. (1998). Sesbania tree fallows on phosphorus-deﬁcient sites: maize yield and ﬁnancial beneﬁt. Agron. J. 90, 717–726. Jama, B., Palm, C. A., Buresh, R. J., Niang, A., Gachengo, C., Nziguheba, G., and Amadalo, B. (2000). Tithonia diversifolia as a green manure for soil fertility improvement in western Kenya: a review. Agrofor. Syst. 49, 201 –221. Janssen, B. H. (1993). Integrated nutrient management: the use of organic matter and mineral fertilizers. In “The Role of Plant Nutrients for Sustainable Food Crop Production in sub-Saharan Africa” (H. Van Reuler and W. Prins, Eds.), pp. 89 –106. VKP, Leidschendam. Jeffery, R., Jeffery, P., and Lyon, A. (1989). Cattle-dung and dung-work: an aspect of the value of women’s work in rural north India. Int. Asienforum 20(1–2), 21–42. Jiri, O., and Waddington, S. R. (1998). Leaf prunings from two species of Tithonia raise maize grain yields in Zimbabwe, but take a lot of labour! pp. 4–5. In “Newsletter of Soil Fertility Network”, Harare, Zimbabwe. Kang, B. (1993). Alley cropping: past achievements and future directions. Agrofor. Syst. 23, 141 –155. Kang, B. T. (1997). Alley cropping — soil productivity and nutrient recycling. For. Ecol. Manag. 91, 75 –82. Kang, B. T., Wilson, G. F., and Sipkens, L. (1981). Alley cropping maize (Zea mays L.) and Leucaena (Leucaena leucocephala LAM) in Southern Nigeria. Plant Soil 63, 165–197. Kang, B. T., Grimme, H., and Lawson, T. L. (1985). Alley cropping sequentially cropped maize and cowpea with leucaena on a sandy soil in southern Nigeria. Plant Soil 85, 267–276. Kanyama-Phiri, G., and Snapp, S. S. (1997). Developing improved organic matter technologies in partnership with smallholder farmers in southern Malawi. In “Africa Crop Science Conference, Pretoria, S. Africa”. Kaya, B., Hildebrand, P. E., and Nair, P. K. R. (2000). Modelling changes in farming systems with the adoption of improved fallows in southern Mali. Agric. Syst. 66, 51–68. Keatinge, J. D. H., Qi, A., Wheeler, T. R., Ellis, R. H., and Summerﬁeld, R. J. (1998). Effects of temperature and photoperiod on phenology as a guide to the selection of annual legume cover and green manure crops for hillside farming systems. Field Crops Res. 57, 139–152. Keatinge, J. D. H., Qi, A., Wheeler, T. R., Musitwa, F., Franks, N. A. P., Kuribuza, D., Ellis, R. H., and Summerﬁeld, R. J. (1999). Potential annual-sown legumes for low-input systems in the East African Highlands of southwestern Uganda. Mountain Res. Develop. 19(4), 345 –353. Kiff, E., Chan, M.-K., and Jackson, D. (1997). Integrated Food Crop Systems Project, Ghana: Development and Promotion of Improved Techniques Of Water and Soil Fertility Management for the Sustainable Production of Crops on Land in the Humid Forest Belt. (Inception Report). Report No. R6789 AO683. Natural Resources Institute, London. Kiff, E., O’Connell, N., Musa, H., and Nelson, D. (1999). Integrated Food Crop Systems Project, Ghana: Development and Promotion of Improved Techniques of Soil and Fertility Management for the Sustainable Production of Crops on Land in the Humid Forest Belt. Report No. R6789 AO683. Natural Resources Institute, London. Kirchhof, G., and Salako, F. K. (2000). Residual tillage and bush-fallow effects on soil properties and maize intercropped with legumes on a tropical Alﬁsol. Soil Use Manag. 16(3), 183–188. Kline, S. J. (1995). In “Conceptual Foundations for Multi-disciplinary Thinking”. Stanford University Press, Stanford, California.

548

A. GRAVES, R. MATTHEWS AND K. WALDIE

Koech, E. K., and Whitbread, R. (2000). Disease incidence and severity on beans in alleys between Leucaena hedgerows in Kenya. Agrofor. Syst. 49(1), 85–101. Kumara Rao, J. V. D. K., Dart, P. J., Matsumato, T., and Day, J. M. (1981). Nitrogen ﬁxation by pigeonpea. In “International Workshop on Pigeonpeas” (Y. I. Nene, Ed.), pp. 190– 199. International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, Andhra Pradesh, India. Kumwenda, J. D. T., Saka, A. R., Snapp, S. S., Ganunga, R. P., and Benson, T. (1997a.). Effects of organic legume residues and inorganic fertilizer nitrogen on maize yield in Malawi. In “Soil Fertility Research for Maize-Based Farming Systems in Malawi and Zimbabwe” (S. R. Waddington, H. K. Murwira, J. D. T. Kumwenda, and D. Hikwa, Eds.), CIMMYT, Mutare. Kumwenda, J. D. T., Waddington, S., Snapp, S. S., Jones, R. B., and Blackie, M. J. (1997b). Soil fertility management in the smallholder maize-based cropping systems of eastern and southern Africa. In “The Emerging Maize Revolution” (C. Eisher, Ed.), pp. 153– 172. Michigan State University, Michigan. Kwesiga, F. R., Franzel, S., Place, F., Phiri, D., and Simwanza, C. P. (1999). Sesbania sesban improved fallows in eastern Zambia: their inception, development, and farmer enthusiasm. Agrofor. Syst. 47, 49–66. Lal, R. (1989a). Agroforestry systems and soil surface management of a tropical alﬁsol. II. Water runoff, soil erosion and nutrient loss. Agrofor Syst. 8, 97–112. Lal, R. (1989b). Agroforestry systems and soil surface management of a tropical alﬁsol. III. Changes in soil chemical properties. Agrofor. Syst. 8, 113–132. Lehmann, J., da Silva, J. P., Trujillo, L., Blanke, M., and Pohlan, J. (2000). Legume cover crops and nutrient cycling in tropical fruit tree production. Acta Hortic. 531, 65–72. Lekasi, J. K., Tanner, J. C., Kimani, S. K., and Harris, P. J. C. (1998). Manure Management in the Kenya Highlands: Practice and Potential, p. 32. HDRA Publications, UK. Leng, R. A. (1990). Factors affecting utilisation of poor quality forages by runinants under tropical conditions. Nutr. Res. Rev. 3, 277 –303. Long, A. J., and Nair, P. K. R. (1999). Trees outside forests: agro-, community, and urban forestry. New Forests 17, 145–174. Low, A. (1988). Farm household economics and the design and impact of biological research in southern Africa. Agricultural Administration and Extension 29, 23–34. Mafongoya, P. L., Barak, P., and Reed, J. D. (1997a). Carbon, nitrogen and phosphorus mineralization of tree leaves and manure. Biol. Fert. Soils 30(4), 298 –305. Mafongoya, P. L., Dzowela, B. H., and Nair, P. K. (1997b). Effect of multi-purpose tree, age of cutting, and drying method on pruning quality. In “Driven by Nature; Plant Litter Quality and Decomposition” (G. Cadisch and K. E. Giller, Eds.), pp. 167–174. CAB International, Wallingford, UK. Mahamoudou, S., and Meritan, R. (1998). Agricultural intensiﬁcation and agroforestry in the Comoros Islands. Cahiers Agric. 7(3), 207–212. Mann, C. (1997). Reseeding the Green Revolution. Science 227(August), 1038–1043. Manu, V. T., Halavatau, S., and Craswell, E. T. (1994). Agroforestry in the food production systems in the South Paciﬁc. In Soil Organic Matter Management for Sustainable Agriculture. “Proceedings of a Conference held at Ubon, Thailand, 24 –26 August 1994” (R. D. B. Lefroy and G. J. Blair, Eds.), pp. 63–68. ACIAR Proceedings No. 56. Maroko, J., Buresh, R. J., and Smithson, P. C. (1998). Soil nitrogen availability as affected by fallow– maize systems on two soils in Kenya. Biol. Fert. Soils 26, 229 –234. Marschner, H. (1995). In “Mineral Nutrition of Higher Plants”. Academic Press Ltd, London. Matthews, R. B. (2000). The role of quantitative methods in integrating biophysical and socioeconomic aspects of soil fertility management. In “Proceedings of a workshop on ‘Biophysical and socio-economic tools for assessing soil fertility’ held at ARS-Lumle, Nepal, 17 –18 October 2000” (J. Ellis-Jones, Ed.). Silsoe Research Institute, UK, Silsoe.

LOW EXTERNAL INPUT TECHNOLOGIES

549

Matthews, R. B., Azam-Ali, S. N., Saffell, R. A., Peacock, J. M., and Williams, J. H. (1991). Plant growth and development in relation to the microclimate of a sorghum/groundnut intercrop. Agric. For. Meteorol. 53, 285–301. Matthews, R. B., Holden, S. T., Lungu, S., and Volk, J. (1992a). The potential of alley cropping in improvement of cultivation systems in the high rainfall areas of Zambia. I. Chitemene and fundikila. Agroforestry Systems 17, 219– 240. Matthews R. B., and Stephens, W. (Eds.), (2002). In “Crop–Soil Simulation Models: applications in Developing Countries.” p. 277. CAB International, Wallingford, UK. Matthews, R. B., Volk, J., Lungu, S., Holden, S. T., and Solberg, K. (1992b). The potential of alley cropping in improvement of cultivation systems in the high rainfall areas of Zambia. II. Maize production. Agrofor. Syst. 17, 241–261. Matthews, R. B., Wassmann, R., Knox, J. W., and Buendia, L. (2000). Using a crop/soil simulation model and GIS techniques to assess methane emissions from rice ﬁelds in Asia. IV. Upscaling of crop management scenarios to national levels. Nutrient Cycling in Agroecosystems 58, 201–217. McDonagh, J., and Hillyer, A. (2000). Soil fertility, legumes and livelihoods in northern Namibia. In DFID Plant Sciences Research Programme Annual Report. 1999/2000 (J. Witcombe and D. Harris, Eds.), pp. 39–47. Centre for Arid Zone Studies, University of Bangor, Bangor, UK. Mekonnen, K., Buresh, R. J., and Jama, B. (1997). Root and inorganic nitrogen distributions in sesbania fallow, natural fallow, and maize ﬁelds. Plant Soil 188, 319 –327. Mogaka, L. M. (1993). An economic evaluation of zero-grazing feeding system for high yielding cows on smallholder farms in Kenya. In “Sustainable Feed Production and Utilisation for Smallholder Livestock Enterprises in sub-Saharan Africa” (J. Ndikumana and W. d. Leeuw, Eds.), AFRNET, Nairobi. Moore, A. W. (1962). The inﬂuence of a legume on soil fertility under a grazed tropical pasture. Empire J. Exp. Agric. 30, 239– 248. Morris, D. R., Weaver, R. W., Smith, G. R., and Rouquette, F. M. (1990). Nitrogen transfer from arrowleaf clover to ryegrass in ﬁeld plantings. Plant and Soil 128, 293 –297. Moser, C. M., and Barrett, C. B. (2002). The system of rice intensiﬁcation in practice: explaining low farmer adoption and high disadoption in Madagascar. Paper presented at the Water-wise Rice Production Workshop, 8 –11 April 2002. International Rice Research Institute, Los Ban˜os, Philippines. Muchagata, M. G. (1997). Forests and people: the role of forest production in frontier farming systems in eastern Amazonia. (Development Studies Occasional Paper 36). School of Development Studies, University of East Anglia, Norwich, UK. 79 pp. Mugendi, D. N., Nair, P. K. R., Mugwe, J. N., MK, O. N., and Woomer, P. L. (1999). Alley cropping of maize with Calliandra and Leucaena in the subhumid highlands of Kenya. Part 1. Soil-fertility changes and maize yield. Agrofor. Syst. 46(1), 39–50. Muriwara, H. K., and Kirchmann, H. (1993). Nitrogen dynamics and maize growth in a Zimbabwean sandy soil under manure fertilization. Commun. Soil Sci. Plant Anal. 24, 2343–2359. Murwira, K. H., Swift, M. J., and Frost, P. G. H. (1995). Manure as a key resource in sustainable agriculture. In International Conference on Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of Sub-Saharan Africa (J. M. Powell, S. Fernandez-Rivera, T. O. Williams, and C. Renard, Eds.), pp. 131–148. ILCA, Addis Ababa. Musa, A. M., Harris, D., Johansen, C., and Kumar, J. (2001). Short duration chickpea to replace fallow after Aman rice: the role of on-farm seed priming in the high Barind Tract of Bangladesh. Exp. Agric. 37(4), 509–521. Muza, L. (1995). Green manuring through legume relay cropping in maize in communal areas of Zimbabwe. In “Second South Eastern Africa Soil Fertility Network Workshop”. CIMMYT, Kadoma, Zimbabwe. Nair, P. (1993). An Introduction to Agroforestry. Kluwer Academic Publishers, Dordrecht, Netherlands.

550

A. GRAVES, R. MATTHEWS AND K. WALDIE

Nandwa, S. M. (1995). Synchronization of nitrogen mineralization with N uptake through maize stover placements and N fertilization under continuous maize monocropping systems in Kenya, p. 344. University of Exeter, Exeter, UK, Ph.D. Thesis. Nandwa, S. M., and Bekunda, M. A. (1998). Research on nutrient ﬂows and balances in East and Southern Africa: state-of-the-art. Agric. Ecosyst. Environ. 71, 5–18. Neill, S. P., and Lee, D. R. (2001). Explaining the adoption and disadoption of sustainable agriculture: the case of cover crops in northern Honduras. Econ. Develop. Cult. Change 49(4), 793–820. Nelson, R. A., and Cramb, R. A. (1998). Economic incentives for farmers in the Philippine uplands to adopt hedgerow intercropping. J. Environ. Manag. 54(2), 83 –100. Nelson, R. A., Grist, P. G., Menz, K. M., Paningbatan, E. P., and Mamicpic, M. A. (1996). Cost-beneﬁt analysis of hedgerow intercropping in the Philippine uplands using SCUAF (Soil Changes Under Agroforestry). SEARCA/SEAMEO Regional Center for Graduate Study and Research in Agriculture, College, Laguna. Nelson, R. A., Cramb, R. A., and Mamicpic, M. A. (1998). Erosion/productivity modelling of maize farming in the Philippine uplands. Part III: Economic analysis of alternative farming methods. Agric. Syst. 58(2), 165– 183. Nguyen The, D., Klinnert, C., Martius, C., Tiessen, H., and Vlek, P. L. G. (2001). Problems with and local solutions for organic matter management in Vietnam. Nutrient Cycling in Agroecosystems 61(1–2), 89–97. Nibbering, J. W. (1991). Crisis and resilience in upland land use in Java. In “Indonesia: Resources, Ecology and Environment” (J. Hardjono, Ed.), pp. 104–132. Oxford University Press, New York. Nyathi, P., and Campbell, B. M. (1993). The acquisition and use of miombo litter by small-scale farmers in Masvingo, Zimbabwe. Agrofor. Syst. 22, 43–48. Oko, B. F. D., Shiyam, J. O., Eneji, A. E., Honna, T., and Yamamoto, S. (2000). Yield of selected food crops under alley-cropping with some hedgerow species in humid tropical southeastern Nigeria. Jpn. J. Tropic. Agric. 44(3), 167–170. Onduru, D. D., Gachini, G. N., de Jager, A., and Diop, J. M. (1999). Participatory research on compost and liquid manure in Kenya. “Managing Africa’s Soils”, Vol. 8, p. 23. Drylands Programme, International Institute for Environment and Development, London, UK. Ong, C. (1994). Alley cropping — ecological pie in the sky? Agrofor. Today 6, 8–10. Ong, C. K., Wilson, J., Deans, J. D., Mulayta, J., Raussen, T., Wajja Musukwe, N., Turner, N. C., and Ward, P. R. (2002). Tree– crop interactions: manipulation of water use and root function. Agric. Water Manag. 53(1–3), 171–186. Oue´draogo, E., Mando, A., and Zombre´, N. P. (2001). Use of compost to improve soil properties and crop productivity under low input agricultural system in West Africa. Agric. Ecosyst. Environ. 84, 259–266. Padwick, G. W. (1983). Fifty years of Experimental Agriculture in the maintenance of soil fertility in tropical Africa: A review. Exp. Agric. 19, 293–310. Pagiola, S., and Holden, S. (2001). Farm household intensiﬁcation decisions and the environment. In “Tradeoffs or Synergies?” (D. R. Lee and C. B. Barrett, Eds.), pp. 73–88. Agricultural Intensiﬁcation, Economic Development, and the Environment. CAB International, Wallingford, UK. Palm, C., Wangari, N., and Gichuru, M. (1996). Selection of organic inputs for soil fertility management. In “Soil Fertility Network for Maize-based Farming Systems in Selected Countries of southern Africa. Research Results and Network Outputs in 1994 and 1995” (S. R. Waddington, Ed.), pp. 19 –24. CIMMYT, Harare, Zimbabwe. Palm, C. A., Myers, R. J. K., and Nandwa, S. M. (1997a). Combined use of organic and inorganic nutrient sources for soil fertility maintenance and replenishment. In “Replenishing Soil Fertility in Africa” (R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 193–217. Soil Science Society of America, Madison, Wisconsin, USA. Palm, C. A., Myers, R. J. K., and Nandwa, S. M. (1997b). Organic–inorganic nutrient interactions in soil fertility replenishment. In “Replenishing Soil Fertility in Africa”

LOW EXTERNAL INPUT TECHNOLOGIES

551

(R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 193–218. Soil Science Society of America, Madison. Palm, C. A., Gachengo, C., Delve, R., Cadisch, G., and Giller, K. E. (2001). Organic inputs for soil fertility management in tropical agroecosystems: application of an organic resource database. Agric. Ecosyst. Environ. 83, 27–42. Pande, R. S. (1997). Fodder and Pasture Development in Nepal. Udatya Research and Development Services, Kathmandu. Paningbatan, E. P. Jr. (1990). Alley cropping for managing soil erosion in sloping lands, Transactions of the 14th International Soil Science Society Congress, Kyoto, Japan. pp. 376 –377. Pantanali, R. (1996). Lesotho: a note on the Machobane System. Report No. TCI Occasional Paper Series No. 7. Food and Agricultural Organisation, Rome, Italy. Patra, B., and Poi, S. C. (1998). Inﬂuence of intercropping on the nodulation in legumes. Environ. Ecol. 16(2), 418–423. Pieri, C. (1995). Long-term soil management experiments in Semi-arid Francophone Africa. In “Soil Management Experimental Basis for Sustainability and Environmental Quality. Advances in Soil Science” (R. Lal and B. A. Stewart, Eds.), pp. 225 –266. CRC, Lewis. Pilbeam, C. J., Tripathi, B. P., Munankarmy, R. C., Gregory, P. J., Ellis Jones, J., Mason, T., and Keatinge, D. (1999a). Productivity and economic beneﬁts of integrated nutrient management in three major cropping systems in the mid-hills of Nepal. Mountain Res. Develop. 19(4), 333–344. Pilbeam, C. J., Tripathi, B. P., Sherchan, D. P., Gregory, P. J., and Gaunt, J. (1999b). Nitrogen balances for the households in the mid-hills of Nepal. Agric. Ecosyst. Environ. 79, 61–72. Place, F., and Dewees, P. (1999). Policies and incentives for the adoption of improved fallows. Agrofor. Syst. 47, 323 –343. Pound, B., Webb, M., Carreno, B., and Flores, A. (1999). Weed Management for Sustainable Agriculture in Forest Margins; Investigacion Adaptiva en Margenes de Bosque. (Joint Final Technical Report). Natural Resources Institute, London, and CIAT, Cali Bolivia. Powell, J. M., and Williams, T. O. (1993). Livestock, nutrient cycling and sustainable agriculture in West African Sahel. International Institute for Environment and Development, London, Gatekeeper Series No. 37. Pradeepmani, D. (1988). Introducing multipurpose trees on small farms in Nepal. In “Proceedings of an international workshop held November 2–5, (1987) in Pattaya, Thailand” (D. Withington, K. G. MacDicken, C. B. Sastry and N. R. Adams, Eds.), pp. 197–203. Winrock International Institute for Agricultural Development, Arlington, Virginia, USA, and International Development Research Centre, Canada. Prasad, P. V. V., Satyanarayana, V., Murthy, V. R. K., and Boote, K. J. (2002). Maximizing yields in rice–groundnut cropping sequence through integrated nutrient management. Field Crops Res. 75(1), 9– 21. Pretty, J. (1995). Regenerating Agriculture: Policies and Practice for Sustainability and Self-reliance, p. 336. Earthscan, London. Pretty, J. (1999). Can sustainable agriculture feed Africa? New evidence on progress, processes and impacts. Environ. Develop. Sustainability 1, 253–274. Pretty, J., and Koohafkan, P. (2002). Land and Agriculture: a compendium of recent sustainable development initiatives in the ﬁeld of agriculture and land management, p. 76. Food and Agriculture Organization of the United Nations, Rome. Pretty, J. N., Morison, J. I. L., and Hine, R. E. (2003). Reducing food poverty by increasing agricultural sustainability in developing countries. Agric. Ecosyst. Environ. 95, 217– 234. Prudencio, C. F. (1993). Ring management of soils and crops in the West African semi-arid tropics: the case of the mossi farming system in Burkina Faso. Agric. Ecosyst. Environ. 47, 237 –264. Quansah, C., Drechsel, P., Yirenkyi, B. B., and Asante-Mensah, S. (2001). Farmers’ perceptions and management of soil organic matter — a case study from West Africa. Nutrient Cycling in Agroecosystems 61, 205–213.

552

A. GRAVES, R. MATTHEWS AND K. WALDIE

Rachmat, R., Stur, W. W., and Blair, G. J. (1992). Cattle feeding systems and limitations to feed supply in South Sulawesi, Indonesia. Agric. Syst. 39(4), 409 –419. Ramasamy, S., ten Berge, H. F. M., and Purushothaman, S. (1997). Yield formation in rice in response to drainage and nitrogen application. Field Crops Res. 51, 65–82. Rao, M. R., and Mathuva, M. N. (2000). Legumes for improving maize yields and income in semi-arid Kenya. Agric. Ecosyst. Environ. 78(2), 123–137. Rao, M. R., Nair, P. K. R., Ong, C. K., and Latt, C. R. (1997). Biophysical interactions in tropical agroforestry systems. Agrofor. Syst. 38(1–3), 3 –50. Reddy, D. D., Rao, A. S., and Rupa, T. R. (2000). Effects of continuous use of cattle manure and fertiliser phosphorus on crop yields and soil organic phosphorus on a vertisol. Bioresour. Technol. 75, 114– 118. Reed, J. D. (1986). Relationships among soluble phenolics, insoluble proanthocyanins and ﬁbre in East African browse species. J. Range Manag. 39, 5–7. Reid, J. D. Jr. (1976). Sharecropping and agricultural uncertainty. Econ. Develop. Cult. Change 24(3), 549–576. Reynolds, L., Di Domenico, C., Atta-Krah, A. N., and Cobbina, J. (1991). Alley farming in southwestern Nigeria: the role of farming systems research in technology development. In “Planned Change in Farming Systems: Progress in On-farm Research” (R. Tripp, Ed.), pp. 85–108. John Wiley & Sons Ltd, London. Rocheleau, D., and Rocheleau, D. E. (1990). Gender complementarity and conﬂict in sustainable forestry development: a multiple user approach “19th IUFRO World Congress”, pp. 432– 448. Canadian IUFRO World Congress Organizing Committee, Montreal. Rodell, M. G. W., Hopley, J. D. H., and Boaltwood, J. N. (1980). Effects of applied nitrogen, kraal compact and maize stover on yields of maize growth on a poor granite soil. Zimbabwe Agric. J. 77, 229 –232. Roose, E., and Ndayizigiye, F. (1997). Agroforestry, water and soil fertility management to ﬁght erosion in tropical mountains of Rwanda. Soil Technol. 11, 109 –119. Roose, E., Kabore, V., and Guenat, C. (1999). Zai practice: a West African traditional rehabilitation system for semiarid degraded lands, a case study in Burkina Faso. Arid Soil Res. Rehabil. 13(4), 343–355. Ruthenberg, H. (1980). Farming Systems in the Tropics, p. 424. Clarendon Press, Oxford. Saka, A. R., Phombeya, H. S. K., Itimu, O. A., Jones, R. B., Minae, S., and Snapp, S. S. (1995). Agroforestry Commodity Team Annual Report. Department of Agricultural Research, Malawi Ministry of Agriculture, Lilongwe. Salako, F. K., Hauser, S., Babalola, O., and Tian, G. (2001). Improvement of the physical fertility of a degraded Alﬁsol with planted and natural fallows under humid tropical conditions. Soil Use Manag. 17(1), 41– 47. Salam, M. A., Noguchi, T., and Koike, M. (2000). Understanding why farmers plant trees in homestead agroforestry in Bangladesh. Agrofor. Syst. 50, 77–93. Sanchez, P. A. (1999). Improved fallows come of age in the tropics. Agrofor. Syst. 47, 3–12. Sanchez, M. D., and Rosales Mendez, M. (1999). Agroforestry systems for intensifying sustainable livestock production in tropical Latin America. “FAO Anim Prod Health Paper”, 143, pp. 1 –13. Sanchez, P. A., Jama, B., Niang, A., and Palm, C. (2001). Soil fertility, small-farm intensiﬁcation, and the environment in Africa. In “Tradeoffs or Synergies? Agricultural Intensiﬁcation, Economic Development, and the Environment” (D. R. Lee and C. B. Barrett, Eds.), pp. 325–344. CAB International, Wallingford, UK. Sanginga, N., Vanlauwe, B., and Danso, S. K. A. (1995). Management of biological N2 ﬁxation in alley cropping systems: estimation and contribution to N balance. Plant Soil 174(1–2), 119– 141. Schaber, J. (1997). “FARMSIM: a dynamic model for the simulation of yields, nutrient cycling and resource ﬂows on Philippine small-scale farming systems”, MSc Thesis. p. 115. Universita¨t Osnabru¨ck, Osnabru¨ck.

LOW EXTERNAL INPUT TECHNOLOGIES

553

Scoones, I., and Toulmin, C. (1998). Soil nutrient balances: what use for policy? Agric. Ecosyst. Environ. 71, 255–267. Seif El Din, A. G., and Raintree, J. B. (1987). Gum hashab and land tenure in western Sudan, Land, trees and tenure. In “Proceedings of an international workshop on tenure issues in agroforestry, Nairobi, May 27 –31, 1985”, pp. 217–224. International Council for Research in Agroforestry (ICRAF), Nairobi, Kenya, and Land Tenure Center (LTC), Madison, Wisconsin, USA Selvarajan, K. N., and Krishnamoorthy, S. (1990). Cost beneﬁt analysis of potato in the Nilgiris district, Tamil Nadu. Artha Vikas 26(1–2), 59– 66. Senesi, N. (1989). Composted materials as organic fertilisers. Sci. Total Environ. 81/82, 521 –542. Shepherd, K. D., and Soule, M. J. (1998). Soil fertility management in west Kenya: a dynamic simulation of productivity, proﬁtability and sustainability at different resource endowment levels. Agric. Ecosyst. Environ. 71, 131–145. Sherchan, D. P., Pilbeam, C. J., and Gregory, P. J. (1999). Response of wheat-rice and maize/millet systems to fertilizer and manure applications in the mid-hills of Nepal. Exp. Agric. 35(1), 1–13. Shumba, E. M., Dhilwayo, H. H., and Mukoko, O. Z. (1990). The potential of maize–cowpea intercropping in low rainfall areas of Zimbabwe. Zimbabwe J. Agric. Res. 28, 33–38. Siaw, D. E. K. A., Kang, B. T., and Okali, D. U. U. (1991). Alley cropping with Leucaena leucocephala (Lam.) de Wit and Acioa barteri (Hook f.). Engl. Agrofor. Syst. 14, 219–231. Singh, R. P., and Desai, G. M. (1991). Farmers’ fertilizer practices in the semi-arid tropics of India. (Progress Report). Report No. 100, p. 54. Economics Group, Resource Management Program, ICRISAT, Patancheru, Andhra Pradesh, India. Singh, R. P., and Jodha, N. S. (1990). Determinants of intercropping in the semi-arid tropics. (Progress Report). Report No. 95, p. 14. Economics Group, Resource Management Program, ICRISAT, Patancheru, Andhra Pradesh, India. Singh, K. N., and Sharma, D. P. (2001). Response of wheat to nitrogen and potassium in saline soils. Exp. Agric. 37(3), 417 –427. Singh, R. P., Ong, C. K., and Saharan, N. (1989). Above and below ground interactions in alley cropping in semi-arid India. Agrofor. Syst. 9, 259–274. Singh, K., Rao, B. R. R., Singh, C. P., Bhattacharya, A. K., Kaul, P. N., and Rajeswara Rao, B. R. (1998). Production potential of aromatic crops in the alleys of Eucalyptus citriodora in semi-arid tropical climate of south India. J. Medicinal Aromatic Plant Sci. 20(3), 749–752. Slingerland, M. A., and Stork, V. E. (2000). Determinants of the practice of Zai and mulching in North Burkina Faso. J. Sustainable Agric. 16(2), 53– 76. Snapp, S. S., Mafongoya, P. L., and Waddington, S. (1998). Organic matter technologies for the integrated nutrient management in smallholder cropping systems of southern Africa. Agric. Ecosyst. Environ. 71, 185 –200. Somda, Z. C., Powell, J. M., Fernadez-Rivera, S., and Reed, J. (1995). Feed factors affecting nutrient excretion by ruminants and fate of nutrients when applied to soil. In “Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of sub-Saharan Africa” (J. M. Powell, S. FernadezRivera, J. Reed, S. Williams, and C. Renard, Eds.), pp. 227–246. ILCA, Addis Abbaba. Stoop, W. A., Uphoff, N., and Kassam, A. (2002). A review of agricultural research issues raised by the system of rice intensiﬁcation (SRI) from Madagascar: opportunities for improving farming systems for resource-poor farmers. Agric. Syst. 71, 249 –274. Stur, W. W., Horne, P. M., Gabunada, F. A. Jr, Phengsavanh, P., and Kerridge, P. C. (2002). Forage options for smallholder crop–animal systems in Southeast Asia: working with farmers to ﬁnd solutions. Agric. Syst. 71(1–2), 75–98. Styger, E., Rakotoarimanana, J. E. M., Rabevohitra, R., and Fernandes, E. C. M. (1999). Indigenous fruit trees of Madagascar: potential components of agroforestry systems to improve human nutrition and restore biological diversity. Agrofor. Syst. 46, 289 –310. Sutihar, D. N. (1984). An econometric analysis of wheat production in Nepal: a case study of Saptari district. Econ. J. Nepal 7(1), 56– 66.

554

A. GRAVES, R. MATTHEWS AND K. WALDIE

Swift, M. J., Seward, P. D., Frost, P. G. H., Qureshi, J. N., and Muchena, F. N. (1994). Long-term experiments in Africa: developing a database for sustainable land use under global change. In “Long-term Experiments in Agricultural and Ecological Sciences.” (R. A. Leigh and A. E. Johnson, Eds.), pp. 229 –251. CAB International, Wallingford, UK. Swinkels, R., and Franzel, S. (1997). Adoption potential of hedgerow intercropping in maize-based cropping systems in the highlands of western Kenya. 2. Economic and farmers’ evaluation. Exp. Agric. 33, 211–223. Swinkels, R. A., Franzel, S., Shepherd, K. D., Ohlsson, E., and Ndufa, J. K. (1997). The economics of short rotation improved fallows: evidence from areas of high population density in western Kenya. Agric. Syst. 55(1), 99–122. Szott, L. T. (1987). “Improving the production of shifting cultivation in the Amazon Basin of Peru through the use of leguminous vegetation”, Ph.D. Thesis. North Carolina State University, Raleigh, North Carolina, p. 168. Szott, L. T., and Palm, C. A. (1996). Nutrient stocks in managed and natural humid tropical fallows. Plant Soil 186, 293–309. Szott, L. T., Palm, C. A., and Buresh, R. J. (1999). Ecosystem fertility and fallow function in the humid and subhumid tropics. Agrofor. Syst. 47, 163–196. Tanner, J. C., Holden, S. J., Winugroho, M., Owen, E., and Gill, M. (1993). Feeding livestock for compost production: a strategy for sustainable upland agriculture on Java. In “International Conference on Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of SubSaharan Africa” (J. M. Powell, S. Fernandez-Rivera, T. O. Williams and C. Renard, Eds.), pp. 115–128. Addis Ababa. Tarawali, G., Manyong, V. M., Carsky, R. J., Vissoh, P. V., Osei-Bonsu, P., and Galiba, M. (1999). Adoption of improved fallows in West Africa: lessons from mucuna and stylo case studies. Agrofor. Syst. 47, 93–122. Thomas, G. V., Rohini, I., Bopaiah, B. M., and Iyer, R. (1991). Beneﬁcial microbes in the nutrition of coconut. J. Plantation Crops 19(2), 127–138. Tian, G., Hauser, S., Koutika, L. S., Ishida, F., Chianu, J. N., Tian, G., Keatinge, D., Carsky, R., and Wendt, J. (2001). Pueraria cover crop fallow systems: beneﬁts and applicability, sustaining soil fertility in West Africa. In “Proceedings of a symposium sponsored by the Soil Science Society of America and the American Society of Agronomy, 5– 9 November 2000”, pp. 137–155. Soil Science Society of America Inc. Madison; USA, Minneapolis, USA. van Tienhoven, N., Icaza, J., and Lagemann, J. (1982). Farming Systems in Jinotega Nicaragua. CATIE Technical Series Technical Report No. 31. CATIE, Turrialba, Costa Rica. Todaro, M. P. (1999). “Economic Development”, Addison-Wesley Publishing, Boston, USA. p. 783. Turner, M. (1995). The sustainability of rangeland to cropland nutrient transfer in semi-arid west Africa: ecological and social dimensions neglected in the debate. In “Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of sub-Saharan Africa”. Volume II: Technical papers. Proceedings of an International Conference held in Addis Ababa, Ethiopia, 22 –26 November (1993) (J. Powell, S. Fernandez-Rivera, T. Williams and C. Renard, Eds.), pp. 435 – 452. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. UNEP, (1999). “Use of Mucuna cover cropping to control Imperata cylindrica”. United Nations Development Programme, New York, NY 10017, Accessed 14-Nov-2002 online at: http://www. undp.org/tcdc/bestprac/agri/cases/benin15.htm. UNPD, (2001). World Population Prospects: The 2000 Revision “Highlights”, Population Division, Department of Economic and Social Affairs, United Nations, New York. p. 22. Uphoff, N. T. (1999). Agroecological implications of the system of rice intensiﬁcation (SRI) in Madagascar. Environ. Develop. Sustainability 1(3–4), 297– 313. Valverde, B. E., Merayo, A., Reeder, R., and Riches, C. R. (1999). Integrated management of itchgrass (Rottboellia cochinchinensis) in maize in seasonally-dry Central America: facts and perspectives.

LOW EXTERNAL INPUT TECHNOLOGIES

555

“Proceedings of International Crop Protection Conference, Brighton, UK, 15–18 November 1999”, pp. 131– 140. British Crop Protection Council, Farnham, UK. Vandermeer, J. (1989). “The Ecology of Intercropping”. University of Cambridge, Cambridge. Vanlauwe, B., Aihou, K., Houngnandan, P., Diels, J., Sanginga, N., Merckx, R., Jensen, E. S., and Recous, S. (2001). Nitrogen management in ‘adequate’ input maize-based agriculture in the derived savanna benchmark zone of Benin Republic. Plant and Soil 228(1), 61–71. Versteeg, M. N., Amadji, F., Eteka, A., Gogan, A., and Koudokpon, V. (1998). Farmers’ adoptability of Mucuna fallowing and agroforestry technologies in the coastal savanna of Benin. Agric. Syst. 56(3), 269 –287. Vosti, S. A., and Witcover, J. (1996). Slash and burn agriculture — household perspectives. Agric. Ecosyst. Environ. 58, 23– 38. Vosti, S. A., Witcover, J., Oliveiara, S., and Faminow, M. (1998). Policy issues in agroforestry: technology adoption and regional integration in the western Brazilian Amazon. Agrofor. Syst. 38, 195 –222. Wakle, P. K., Khandare, V. S., Kalaskar, S. S., Karanjikar, P. N., and Daunde, A. T. (1999). Constraints faced by farmers in adoption of banana cultivation technology. J. Soils Crops 9(2), 182 –184. Waldie, K. J. (1990). Cattle and concrete: some aspects of social organisation among the Fula in and around Kabala, northern Sierra Leone, Ph.D. Thesis, Manchester University, Manchester, UK. (Available online at http://csac.anthropology.ac.uk/CSACMonog/Waldie/). Wandera, J. L., Kiruiro, E. M., Chemitei, V. C., and Bulimu, B. O. (1993). Factors limiting the use of fodders by smallholder dairy farmers in Kenya: experiences from on-farm, farmer-managed, animal feeding trials. In “Sustainable Feed Production and Utilisation for Smallholder Livestock Enterprises in Sub-Saharan Africa” (J. Ndikumana and W. de Leeuw, Eds.), AFRNET, Nairobi, Kenya. Webster, C., and Wilson, P. (1998). “Agriculture in the Tropics”. Blackwell Science, Oxford. Wendt, J. W., Jones, R. B., and Itimu, O. A. (1994). An integrated approach to soil fertility improvement in Malawi, including agroforestry. Soil fertility and climatic constraints in dryland agriculture. In “Proceedings of ACIAR-SACCAR Workshop, Harare, Zimbabwe, 30 August–1 September 1993” (E. T. Craswell and J. Simpson Eds.), pp. 74–79. Australian Centre for International Agricultural Research (ACIAR), Canberra, Australia. Wheeler, T. R., Qui, A., Keatinge, J. D. H., Ellis, R. H., and Summerﬁeld, R. J. (1999). Selecting legume cover crops for hillside environments in Bolivia. Mountain Res. Develop. 19(4), 318 –324. Whittome, M. P. B. (1994). “The Adoption of Alley Farming in Nigeria and Benin: the On-Farm Experience of IITA and ILCA.” Ph.D. Thesis. University of Cambridge, Cambridge, England. Wolf, E. C. (1986). “Beyond the green revolution: new approaches for third world agriculture”, Worldwatch Institute, Washington DC, USA. p. 46. Wortmann, C. S., and Kaizzi, C. K. (1998). Nutrient balances and expected effects of alternative practices in farming systems of Uganda. Agric. Ecosyst. Environ. 71, 115– 129. Yamoah, C., Agboola, A. A., and Wilson, G. F. (1986a). Nutrient contribution and maize performance in alley cropping systems. Agrofor. Syst. 4, 247 –254. Yamoah, C. F., Agboola, A. A., and Mulongoy, K. (1986b). Decomposition, nitrogen release and weed control by prunings of selected alley-cropping shrubs. Agrofor. Syst. 4, 239 –246. Yaseen, T. M., Kothari, S. K., Singh, U. B., Singh, K., Sattar, A., and Roy, S. K. (2001). Bio-economic evaluation of scented rose (Rosa damascena) based intercropping systems in north Indian plains. Journal of Medicinal and Aromatic Plant Sciences 23(2), 69–74. Young, A. (1987). The potential of agroforestry for soil conservation and sustainable land use, (ICRAF Reprint No. 39). In “Collected papers” (J. Kozub, Ed.), p. 16. Economic Development Institute of the World Bank, Washington, DC, USA. Young, A. (1989). “Agroforestry for Soil Conservation”, CAB International, Wallingford, UK. p. 284.

This Page Intentionally Left Blank

AMMONIA EMISSION FROM MINERAL FERTILIZERS AND FERTILIZED CROPS Sven G. Sommer,1 Jan K. Schjoerring2 and O.T. Denmead3 1

Department of Agricultural Engineering, Danish Institute of Agricultural Sciences, Research Centre Bygholm, PO Box 536, DK-8700 Horsens, Denmark 2 Plant Nutrition Laboratory, Department of Agricultural Sciences, The Royal Veterinary and Agricultural University, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmark 3 CSIRO Land and Water, GPO Box 1666, Canberra ACT 2601, Australia

I. Introduction II. Mineral Fertilizer Consumption III. Ammonia Volatilization from Mineral Fertilizers A. Production and Transport of NH3 in the Soil– fertilizer – atmosphere Interface B. Temporal NH3 Loss Pattern C. Hydrolysis of Urea D. Soil Hþ E. Soil CEC F. Solid Phase Processes G. Climate and Inﬁltration H. Microbial Processes (nitriﬁcation/immobilization) IV. Ammonia Emission from Crop Foliage A. Transport of NH3 Between Leaves and the Atmosphere B. Magnitude of NH3 Losses C. Physiological Processes Involved in NH3 Emission from Crops V. Management Strategies A. Techniques for Reduction of NH3 Emission B. Fertilizer Composition C. Flooded Fields (rice paddies) D. Injection of Anhydrous Ammonia E. Crop Emissions as Affected by Fertilizer Application F. Ammonia Emission from Decomposing Plant Material G. Absorption by Crops VI. Measurement Techniques A. Tracer Techniques B. Enclosures C. Micrometerological Methods D. Gradient Diffusion Methods E. Eddy Correlation F. Relaxed Eddy Accumulation or Conditional Sampling G. Lagrangian Dispersion Models 557 Advances in Agronomy, Volume 82 Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved 0065-2113/03 $35.00

558

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD VII. Conclusions and Perspectives Appendix A Acknowledgment References

A thorough understanding of the physical and chemical processes involved in NH3 emission from inorganic N fertilizers and fertilized crops is required if reliable and operational NH3 emission factors and decision support systems for inorganic fertilizers are to be developed, taking into account the actual soil properties, climatic conditions and management factors. For this reason, the present review focuses on processes involved in NH3 volatilization from inorganic nitrogen fertilizers and the exchange of ammonia between crop foliage and the atmosphere. The proportion of nitrogen lost from N fertilizers due to NH3 volatilization may range from < 0 to .50%, depending on fertilizer type, environmental conditions (temperature, wind speed, rain), and soil properties (calcium content, cation exchange capacity, acidity). The risk for high NH3 losses may be reduced by proper management strategies including, e.g., incorporation of the fertilizer into the soil, use of acidic fertilizers on calcareous soils, use of fertilizers with a high content of carbonate-precipitating cations, split applications to rice paddies or application to the soil surface beneath the crop canopy. The latter takes advantage of the relatively low wind speed within well-developed canopies, reducing the rate of vertical NH3 transport and increasing foliar NH3 absorption. Conversely, NH3 is emitted from the leaves when the internal NH3 concentration is higher than that in the ambient atmosphere as may often be the case, particularly during periods with rapid N absorption by the roots or during senescence induced N-remobilization from leaves. Between 1 and 4% of shoot N may be lost in this way. q 2004 Academic Press.

I. INTRODUCTION Agriculture is recognized as a major source of atmospheric ammonia (NH3), contributing 50% of global NH3 emissions (Schlesinger and Hartley, 1992). Recent inventories have shown that mineral fertilizers and plants account for about 20% of the total emission of NH3 in Europe (ECETOC, 1994; Pain et al., 1998; Hutchings et al., 2001) and 23% of global emission of NH3 is derived from fertilizers and ﬁeld-applied manure (Bouwman et al., 2002). 22 Ammonia is a chemically active gas and readily combines with NO2 3 and SO4 þ in acid cloud droplets to form particulates (Asman et al., 1998). The NH4 particles can be transported over long distances before being dry- or wet-deposited, while

AMMONIA EMISSION

559

20

Ammonium sulphate

Other NP-N

Ammonium phosphates

10

Calcium Nitrate Ammonium

15

Nitrogen sollutions

20

NPK-N Anhydrous-NH3

25

Ammonium nitrate

40

30

Urea

60

35 Other straight N

Developed world Developing world World

80

Consumption, Tg N year–1

Million tonnes per year

gaseous NH3 usually is deposited much closer to the source (Asman and van Jaarsveld, 1991). The deposited NH3 may cause acidiﬁcation and eutrophication of natural ecosystems (Schulze et al., 1989). Therefore, the United Nations has included NH3 in the Convention on Long-range Transboundary Air Pollution (CLRTP), and in addition, the EU Commission has set a limit — the NH3 ceiling — to the emission of NH3 from European countries (EEA, 1999). For farmers, the loss of fertilizer nitrogen (N) due to NH3 emission may signiﬁcantly reduce N-fertilizer efﬁciency, contributing to a rather low overall efﬁciency of applied N, i.e., , 50% in the tropics and , 70% in temperate areas (Malhi et al., 2001). In order to prevent potential negative consequences of gaseous NH3 losses, farmers may also apply fertilizer-N in excess of crop requirements, which will increase the loss of N to the environment and production costs. Although the use of fertilizers increased dramatically during the 20th century (Fig. 1), the recent trend is towards decreased use, as awareness of the economic consequences and environmental impacts of N losses is encouraging more efﬁcient N utilization. In the 1980s, emission factors were introduced to calculate NH3 emission from European agriculture (Buijsman et al., 1987). Country-speciﬁc emission factors were introduced in inventories in the 1990s (Misselbrook et al., 2000; Hutchings et al., 2001). Major uncertainties are still associated with the use of NH3 emission factors for inorganic fertilizers, because they in many cases are highly empirical or have been derived from measurements carried out under conditions that deviate considerably from modern management practices associated with handling and applying fertilizers. As an example, the generally used emission factor for urea is 15% for Europe and 25% for the tropics (see e.g., Bouwman et al., 1997), which contrasts with the fact that NH3 emission from urea can be completely avoided if the fertilizer is incorporated into the upper soil layers

5 0

0 1960

1970

1980 Year

1990

2000

Figure 1 Left: Global consumption of mineral fertilizers (IFA Statistics, 2002); Right: Distribution of the fertilizer consumed in late 1990 (Bouwman et al., 1997).

560

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

(Harrison and Webb, 2001), or reduced to well below 10% if applied to a wellestablished crop (Schjoerring and Mattsson, 2001). More reliable NH3 emission estimates could potentially be derived from mathematical models based on the physico-chemical processes controlling NH3 emission from fertilizers and their interactions with soil, canopy and atmospheric variables. Complex, mechanistic models of NH3 emission exist (Avnimelech and Laher, 1977; Fleisher et al., 1987; Kirk and Nye, 1991; Genermont and Cellier, 1997), but there are still difﬁculties in describing the controlling processes and their interactions. The substantial requirement for input data makes these models difﬁcult to apply in decision support systems. The present review focuses on the processes underlying NH3 volatilization from inorganic N fertilizers. The inﬂuence of various soil and climatic parameters on the production of NH3/NHþ 4 during dissolution of the fertilizer is described and related to the temporal variation in NH3 losses and their accumulated values. The existing knowledge on the ability of growing crop plants to reduce NH3 losses from fertilizers lying on the soil surface beneath the canopy is outlined, as is the case for the magnitude and mechanisms of NH3 emission from the crop foliage itself. Finally, a description of measurement techniques for quantifying NH3 losses under ﬁeld conditions is included, since methodological aspects are important for assessing loss rates reported in the literature as well as for planning new experiments aiming at obtaining more and better knowledge on NH3 volatilization. Taken together, the information in the review provides an integrated picture of fertilizer-derived NH3 emissions, facilitating the development of decision support systems intended to limit the volatile loss of fertilizer-N from fertilizers and plants.

II.

MINERAL FERTILIZER CONSUMPTION

Figure 1 shows the consumption of mineral fertilizers up to the end of the 20th century. In the 1950s, mineral fertilizers became cheap and the use of mineral-N increased, reducing dependence on leguminous N ﬁxation and animal manure as sources of plant nutrients. In the developed world, the consumption of N fertilizers increased until the late 1980s, but has declined since then due to stricter environmental legislation and a growing number of set-aside policies in the European Union. The consumption of fertilizers in central and Eastern Europe as well as in the Commonwealth countries (EFMA, 1997) fell signiﬁcantly during the 1990s due to the economic crisis that emerged after the fall of the Berlin Wall. Nitrogen fertilizer consumption has increased steadily in the developing world since 1960, mainly due to the increased use of mineral fertilizers in Asia (IFA Statistics, 2002). Urea constitutes about 38% of the total consumption of nitrogenous fertilizers, which is reported to be about 77 Tg N year21 (Bouwman et al., 1997). Urea is an

AMMONIA EMISSION

561

organic compound (amide) with a very high concentration of N (46%). After application to soil, the urea is hydrolyzed to ammonium; therefore, urea is included in the category of mineral fertilizers. Other straight N fertilizers than urea include ammonium bicarbonate (ABC) and ammonium chloride. Ammonium bicarbonate comprises 90% of the other straight N fertilizers, and is the second most used N-containing mineral fertilizer after urea (Fig. 1). It is widely used in China, where it is produced locally by small- and medium-sized manufacturers (Andreas Pacholsky, 2000; personal communication). Apart from urea and ABC, the usage of straight N fertilizers, in order of consumption, is ammonium nitrate (AN) . anhydrous ammonia (AA) . nitrogen solutions . calcium –ammonium – nitrate (CAN). Nitrogen solutions often consist of urea — ammonium – nitrate solutions (UAN), which contain 28– 32% N (EFMA, 1997). Examples of multi-nutrient N-containing fertilizers are NPK, NP and ammonium sulfate. These fertilizers are either produced by chemical reactions (complex or compound fertilizers) or by mechanical blending of relevant minerals (EFMA, 1997). The nitrogen content of the fertilizers may be very variable (5 – 26%) and the content of ammonium and nitrate may vary among fertilizers.

III. AMMONIA VOLATILIZATION FROM MINERAL FERTILIZERS A. THE

PRODUCTION AND TRANSPORT OF NH 3 IN SOIL – FERTILIZER – ATMOSPHERE INTERFACE

Emission of NH3 from applied fertilizers is consequent on transport of NH3 from the surface of an ammoniacal solution to the atmosphere. The rate of emission is determined by the concentration gradient and resistance to NH3 transport between the surface and the atmosphere as controlled by atmospheric transport processes, the chemical composition of the solution, and transformations of Total Ammoniacal Nitrogen (TAN) (NH3 2 N þ NHþ 4 2 N) in the soil. The air above the surface can be envisaged as a laminar or turbulent-free layer close to the surface and, above this, a turbulent layer. Ammonia gas at the liquid – air interface (x) is transported through the laminar layer by molecular diffusion and then through the turbulent layer to the free atmosphere by turbulent diffusion. The instantaneous rate of NH3 loss (Fg) may be given by the following equation, as shown by Rachhpal-Singh and Nye (1986a) and van der Molen et al. (1990):

562

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

þ NHþ 4 $ NH3 " þH

ð1Þ

Fg ¼ Kb £ ðx 2 NH3;a Þ

ð2Þ

where Kb is a bulk transfer coefﬁcient, and x and NH3,a are, respectively, the concentration of NH3 in the air at the soil – air interface and in the free atmosphere. The transfer coefﬁcient is highly dependent on wind speed, but also depends on atmospheric stability. x is determined by the concentration of TAN and equilibrium processes in the solution (Freney et al., 1983; Sherlock and Goh, 1985): ½NH3;L ¼

TAN 1 þ 10ð0:09018þ2729:92=T 2 pHÞ

x ¼ ½NH3;L £ 101477:7=T21:69

ð3Þ ð4Þ

where NH3,L is the concentration of NH3 in the solution, and x is the equilibrium concentration of NH3 in the atmosphere in immediate contact with the soil solution. The change in NH3,L concentration of a solution of TAN and bicarbonate at different pH and temperatures is shown in Fig. 2. It is seen that below pH 7, the concentration of NH3,L is low. Therefore, the increase in NH3,L with temperature is most noticeable when pH is . 7. Soil buffers and cations and anions added with fertilizers will affect soil pH and, hence, NH3 emissions. Furthermore, NH3

Figure 2 Effect of temperature and pH on NH3 concentrations in a liquid solution of TAN and TIC under conditions with no emission of NH3 and CO2 gases. It is assumed that 100 kg urea is mixed homogeneously in the top 1 cm soil layer and the soil has a water content of 30% (vol/vol); thus after hydrolysis, TAN concentration is ca. 0.1 mol l21 and TIC is 0.05 mol l21.

AMMONIA EMISSION

563

emissions will be reduced through reduction of TAN as a result of NH3 loss, inﬁltration of TAN or fertilizers below the soil surface by rainfall, reaction with the cation exchange complex, immobilization, uptake by plants or oxidation by nitrifying organisms. Transport processes at the gaseous interface and in the air layers above the soil surface can be modeled by use of a bulk transfer coefﬁcient (Eq. (2)). The transfer coefﬁcient (Kb) is parameterized by a series of resistances that are additive and are related to the resistances to transport by diffusion and convection. The resistances comprise (I) an aerodynamic resistance (ra) representing the resistance of the turbulent air layer between observation height and the aerodynamic roughness length of the surface (van der Molen et al., 1990; Genermont and Cellier, 1997), (II) a laminar resistance (rb) representing the resistance to transport across the laminar boundary layer near the surface, which is dominated by molecular diffusion, and (III) an interfacial resistance (rc) representing the resistance to transport within the soil or plant solution layer to the air interface (van der Molen et al., 1990). Kb ¼

1 ra þ rb þ rc

ð5Þ

A similar set of resistances describes NH3 loss from water bodies, including rice paddies (Denmead and Freney, 1992). The resistance ra in the turbulent layer can be calculated from aerodynamic theory using developments by, for example, Paulson (1970) and Monteith and Unsworth (1990): ra ¼

ln½ðz 2 dÞ=z0 2 cg ðz 2 dÞ kup

ð6Þ

where z is the height of measurement within the boundary layer of the emitting ﬁeld, d is the zero plane displacement (the height at which wind speed appears to go to zero), z0 is the roughness length of the surface, cg is a function that accounts for the effects of atmospheric stability on gas transport and is deﬁned in Appendix, k is von Karman’s constant (< 0.4), and up is the friction velocity, a measure of the ﬂux of momentum between the atmosphere and the ground. The friction velocity can be measured directly by eddy covariance or estimated from the proﬁle of horizontal wind speed u: up ¼

ku ln½ðz 2 dÞ=z0 2 cm ðz 2 dÞ

ð7Þ

The stability function cm is given in Appendix. The roughness length and the zero plane displacement vary with surface characteristics. z0 and d are assumed to be linear functions of the canopy height (h) (Monteith and Unsworth, 1990): z0 < 0:1h:

ð8Þ

564

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

d < 0:6h:

ð9Þ

The resistance of the laminar boundary layer rb above the crop canopy or soil surface can be estimated using the empirical relationship of Thom (1972): rb ¼ 6:2u20:67 p

ð10Þ

Van der Molen et al. (1990) estimate the resistance rc within the soil surface layer as rc ¼

0:5Ll Daq KH þ Dg

ð11Þ

where the numerator represents the average distance over which the ammoniacal N has to be transported to reach the soil surface, and Daq and Dg are the soil – liquid and soil – gas diffusion coefﬁcients for ammoniacal N. Sutton et al. (1998) and Nemitz et al. (2001) present formulations for the resistances to NH3 transport in plant leaves and canopies, and Denmead and Freney (1992) do the same for NH3 transport across the viscous sublayer below the air– water interface in water bodies. Soil surface roughness leads to an increase in turbulence, which leads to an increase in friction velocity (Brutsaert, 1982), and accordingly to an increase in the exchange between the soil surface and the atmosphere. Thereby, surface NH3 gas becomes better mixed and dispersed in the near atmosphere, and NH3 emission is increased. Consequently increasing wind speed and temperature will increase NH3 emission from applied fertilizers (Eqs. (1) –(10)), but on days with high global radiation, increasing wind speed will reduce soil and plant surface temperatures and thereby the emission potential of the solution. Thus, a number of studies of NH3 emission from manure applied to ﬁelds have shown that NH3 emission is not always related to wind speed (Bussink et al., 1994; Beauchamp et al., 1978; Sommer et al., 1997). The effect of the internal boundary layers may be illustrated by considering a ﬁeld amended with ammoniacal fertilizers and surrounded by untreated ﬁelds. Downwind from the leading edge, the atmospheric NH3 concentration will gradually increase due to the input of NH3 volatilized from the soil solution, but the effect of the increase on emission rate will depend on the surface boundary conditions. The driving force for NH3 emission is the concentration difference between the soil interface containing TAN and the atmosphere (Eq. (2)). If NH3 is freely available at the surface, as following surface spreading of slurries, the surface concentration will remain more or less unchanged and the NH3 emission will gradually decline downwind. The dynamics of the emission will resemble those for a constant concentration boundary condition at the surface, or a radiation boundary condition if the wind speed is varying in time. If, however, NH3 production is controlled by soil processes, such as the hydrolysis of urea incorporated

AMMONIA EMISSION

565

4.0

F/F1000

3.5 3.0 2.5 2.0 1.5 1.0 0

200

400 600 800 Length of field, m

1000

Figure 3 Ammonia emission rate (F) as a fraction of the emission rate (F1000) at 1000 m calculated using the model developed by Philip (1959).

into the soil, the emission rate will be virtually unaffected by plot size or fetch. Ammonia concentrations in the soil air will simply increase in response to the increase in the concentration of the air above and the dynamics will resemble those for a constant ﬂux condition at the surface. In the case of slurry spreading, increasing the area to which fertilizer is applied at a constant application rate will be expected to reduce the emission rate of NH3 as a percentage of the amount of TAN applied (Genermont and Cellier, 1997). The initial NH3 emission rates after fertilizer application will tend to be larger from small plots than big ﬁelds, and to decline faster because TAN has been reduced due to the larger emission during the ﬁrst few hours. The inﬂuence of ﬁeld size on NH3 loss has been demonstrated by results of simulations for smaller plots (0 – 25 m) calculated by Vlek and Craswell (1981) for a ﬂooded soil using a diffusion model (Bouwmeester and Vlek, 1981). Figure 3 shows the dependence of emission rate on plot size, calculated from Philip’s (1959) analysis for a constant concentration boundary condition at the surface. Figure 3 indicates that the average emission rate from a small treated plot with a fetch of only 1 m would be around 2.5 times that from a large ﬁeld with a fetch of 1000 m, and even for a plot with a fetch of 100 m, the emission rate would still be about 30% higher than that from the large ﬁeld.

B.

TEMPORAL NH 3 LOSS PATTERN

NH3 emission from a soil applied fertilizer is a function of dissolved NH3 in equilibrium with the atmosphere either directly or through soil pores. According to the dissociation of NHþ 4 and Henry’s constant (Eqs. (3) and (4)) the emission is related to the concentration of TAN and Hþ(pH ¼ 2 log(Hþ)). In consequence,

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD Cumulative NH3 volatilization (% of N)

566

18 16

DAP Urea

14 12 10 8 6 4 2 0 0

2

4

6

8

10

12

Days from application Figure 4 Ammonia volatilization from urea and diammonium phosphate (DAP) applied to a sandy loam measured with a wind tunnel (after Sommer and Jensen, 1994).

doubling the TAN concentration will double NH3 emission, whereas doubling the Hþ concentration will halve the emission. Typical patterns of NH3 volatilization over time for urea and diammonium phosphate are shown in Fig. 4. Since urea fertilizer contains no TAN, no NH3 is emitted immediately after urea application to the soil. Subsequently, the urea in contact with the soil will absorb water, and urea will hydrolyze, producing TAN and bicarbonate (HCO2 3 ). The rate of hydrolysis is related to the amount of soil water absorbed and temperature; therefore, the initial lag phase with no NH3 volatilization may vary. The NH3 loss rate declines after 5– 10 days due to a reduction in NH3 concentration as TAN is volatilized, dissolved in increasing volumes of soil water, leached by rain into the soil, absorbed to soil, or transformed to nitrate (Black et al., 1985; Haynes and Williams, 1992). The rate of NH3 volatilization from urea has been described by a logistic equation by Stevens et al. (1989), showing that maximum loss rates may occur within one to 10 days after application, depending on soils and environmental conditions. A sigmoidal model was used by Sommer and Ersbøll (1996) to relate cumulated loss of NH3 from urea to the time from application, showing that for loamy soils half the total loss may be reached 2 –7 days after urea application (Table I). The pattern of NH3 volatilization from urea is affected by whether urea is applied in pellets, in solution, or crushed to a ﬁne powder, because dissolution of and diffusion of urea into the soil solution will be different. From urea applied as a powder or in solution, the emission will occur earlier than after application of prilled or granulated urea. This delay in emission from prilled or granular urea is related to hydrolysis of the applied urea (see Section III.C on hydrolysis of urea). Other ammoniacal fertilizers are composed of NHþ 4 salts of phosphate, sulfate, or nitrate; these salts are readily dissolved in the soil water after application of

Table I Models for Predicting NH3 Volatilization from Surface-Applied Ammoniacal Fertilizer Fertilizer Urea

Equation 2ct

F ¼ a=ð1 þ be

Þ 2 að1 þ bÞ

Nmax ¼ ab=ð1 þ bÞ

F is the NH3 volatilization (% of urea-N or % of TAN), t is the time in days from application of urea and a, b, and c are time-related parameters Nmax is total cumulated NH3 volatilization (% of urea-N) tmax is the time in days where the loss rate is at maximum

Urea

F ¼ Nmax ð1 2 e2bt Þc

NHþ 4 mineral fertilizer

F ¼ Nmax ð1 2 e2bt Þ

Urea

F ¼ 3:67 þ 22:28ðDMÞ

DM ¼ dry matter per unit surface (25 £ 65 cm2)

Urea

F ¼ 12:1 2 0:97CEC þ 0:02CEC 2

CEC ¼ cation exchange capacity (cmol kg21), SMC ¼ soil moisture content (g g21) and T ¼ temperature (8C) TA is total acidity (mmol kg21)

2 0:011SMC þ 0:049T Urea

F ¼ 20:84 2 0:0452TA

References Stevens et al. (1989)

Sommer and Easlo`ll (1996) Sommer and Easlo`ll (1996) Hoult and McGarity (1987) McGarry et al. (1987)

AMMONIA EMISSION

tmax ¼ lnðbÞ=c for b . 0 tmax ¼ 0 for b # 0

Deﬁnition of variables

Stevens et al. (1989)

Data were taken from Stevens et al. (1989), Sommer and Ersbøll (1996), Hoult and McGarity (1987), McGarry et al. (1987), Stevens et al. (1989). In the study of McGarry et al. (1987), a closed chamber was used and emission was measured by absorbing NH3 in a solution of acid, in the other studies emission was measured with dynamic chambers in the laboratory.

567

568

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

the fertilizer, and NH3 volatilization will start within a short time after their application. The NH3 volatilization rate will decline due to the same processes as mentioned previously for the decline in volatilization rate after application of urea. In consequence, the NH3 volatilization pattern shows no lag phase after application and the cumulated loss may therefore be described by a simple exponential equation (Table I).

C.

HYDROLYSIS

OF

UREA

Urea is the diamide of carbonic acid and as such an organic fertilizer. Like other amides, urea in solution is hydrolyzed to NH3 and bicarbonate (HCO2 3 ). No NH3 is lost from urea that has not been transformed and emission of NH3 from urea applied in the ﬁeld is, therefore, closely related to hydrolysis of urea by the enzyme urease: 2 COðNH3 Þ2 þ 2H2 O $ NH3 þ NHþ 4 þ HCO3

ð13Þ (NHþ 4)

Activity, micro mol g–1 soil h –1

and bicarbonate This reaction produces a mixture of NH3, ammonium (HCO2 3 ). The rate of hydrolysis of urea is related to urease activity, the availability of water (Eq. (13)), pH, and temperature (Bremner and Douglas, 1971; Fig. 5). Very dry soils with water content below permanent wilting point will delay or inhibit urea hydrolysis; thus at soil water potential , 21.5 MPa, hydrolysis has been shown to be insigniﬁcant (Al-Kanani et al., 1991). Increasing water content of the soil will increase the rate of hydrolysis (Vlek and Carter, 1983; McInnes et al., 1986a; Reynolds and Wolf, 1987). Air humidity inﬂuences hydrolysis of urea because prilled and granular urea are hygroscopic and will absorb water at

12 10

Idasoil, pH 7.5 Marshallsoil, pH 6.8 Edinasoil, pH 6.1

40 30

8 20

6 4

Soil pH 4.9 Soil pH 7.8 Soil pH 6.4

10

2 0

0 7

8

9 pH

10

0

2

4

6

8

10

Urea, mol N l–1 (soil solution)

Figure 5 Rate of urea hydrolysis related to soil pH (left: Tabatabai and Bremner, 1972) and concentration of urea (right: Nye, 1992).

AMMONIA EMISSION

569

high air humidity. In consequence, NH3 emission may be signiﬁcant from urea applied to a dry soil if air humidity is high (Bouwmeester et al., 1985; Reynolds and Wolf, 1987; Black et al., 1987a). Hydrolysis of urea may be delayed in soils low in pH or after the addition of acidifying anions mixed with the urea (Ouyang et al., 1998); optimum pH for soil urease activity has been reported to range from pH 8 to 9. Figure 5 shows that urease activity varies between soils and that for some soils the rate of hydrolysis is not much affected by pH. This may explain why Rachhpal-Singh and Nye (1986a) found that the rate of hydrolysis only varied insigniﬁcantly when soil pH was in the range from 6.6 to 8.6. In soils where water content is not limiting, urease activity increases with temperature (Vlek and Carter, 1983). Urease activity also increases with increasing urea concentration (Tabatabai and Bremner, 1972, Fig. 5). The hydrolysis of urea applied in pellets is slow compared with that applied in solution or as a ﬁne powder, due to the slow diffusion of urea into the soil where urease is abundant (Vlek and Carter, 1983). Therefore, hydrolysis may reach a maximum within 3 –5 days for pellets and 24 –48 h for urea in solution or as a ﬁne powder (Lyster et al., 1980; Black et al., 1987a; Thomas et al., 1988; Whitehead and Raistrick, 1990). Recous et al. (1988) showed that urea applied in the ﬁeld was hydrolyzed with a half-life of 1.9 days at 3.58C; on the other hand, urea well mixed with the soil was hydrolyzed with a half life of 22, 15 and 6 h when incubated in the laboratory at 4, 10 and 208C. This indicates that hydrolysis rate increases signiﬁcantly with temperature, and that urea mixed with the soil is hydrolyzed rapidly. At urea concentrations below 1 mol N per liter (soil) the rate of hydrolysis initially increases linearly with concentration, then reaches an optimum value, eventually decreasing at higher urea concentrations (Fig. 5). This may be explained as substrate inhibition, which can be described by a Michaelis –Menten equation (Nye, 1992). For the purpose of predicting hydrolysis in soils low in urea and with pH in the range from 6.5 to 8.4, urease activity related to pH has been described using the Michaelis– Menten equation at substrate concentrations up to 0.2 mol urea-N per liter of soil followed by a straight line at urea concentrations above 0.2 mol N per liter (Rachhpal-Singh and Nye, 1986b). It is important to realize that after application of urea fertilizer in the ﬁeld, the maximum concentration of urea may be 10 mol N per liter (soil) near the fertilizer granules (Nye, 1992), and that urea hydrolysis has been observed to occur at concentrations below 1 mol N per liter (soil) in most of the studies cited; therefore, the algorithms given cannot predict hydrolysis near the applied granules. Hydrolysis of urea applied to ﬂooded paddy soils for rice production may be affected by redox potential as well as pH. Thus at low redox potential (anoxic), half-time of the hydrolysis (t1/2) has been shown to be 11.2 –24.8 h compared with 7.4 –13.8 h for oxidized suspensions of acid soils (Lindau et al., 1989). In a suspension of calcareous soil, hydrolysis rate is not affected signiﬁcantly by

570

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

differences in redox potential (Lindau et al., 1989); however, the rate of hydrolysis in calcareous suspension was greater (t1/2 of 1.7 –1.9 h) than in the suspension of acid soil. The effect of organic and inorganic compounds added to urea with the function of delaying hydrolysis of urea and thereby reducing NH3 emission is discussed in Section V on management strategies.

D.

SOIL H1

One of the most important rate controlling factors of NH3 emission from applied fertilizers is the Hþ concentration. At pH values higher than 7, the concentration of NH3,L (see Eqs. (3) and (4)) is at a level that may produce signiﬁcant emissions from the applied fertilizer (Fig. 2). The fertilizers may be grouped as acidic and alkaline with, respectively, a low and high potential for NH3 emission (Table II). Soil alkalinity will also interact with fertilizers and affect the concentration of Hþ. Furthermore, the fertilizer application will create “hot spots,” and soil Hþ concentration will typically show a trend from the fertilizer granule microsite to surrounding unaffected soil (Rachhpal-Singh and Nye 1986a,b). After application of ammonium salts (phosphate, nitrate, sulfate, etc.), the emission of NH3 will produce acids of the anions (Eq. (14)), as shown for ammonium sulphate (AS) (AS ¼ NH4HSO4) in the following example: NH4 HSO4 þ H2 O $ NH3 " þHþ þ HSO2 4

ð14Þ

Each mole of emitted NH3 will increase the concentration of Hþ by one mole. Therefore in acidic soils or poorly buffered soils, the pH will decline due to the NH3 emission and the NH3 emission will gradually decline, because little NH3,L will be available for volatilization, consequently, the total loss of TAN will be low. In calcareous soils, emission will be related to the solubility of the anion (Fenn and Hossner, 1985; Du Preez and Burger, 1988). The solubility of nitrate (NO2 3 ),

Table II Fertilizers Grouped in Relation to the Acidity or Alkalinity They Produce When Dissolved in Soil Water Acid (NH4)2SO4 NH4NO3 NH4Cl (NH4)2HPO4

Moderate acid

Alkaline

(NH4)H2PO4

(NH4)HCO3 (NH4)2CO3 NH3 (NH3)2CO

AMMONIA EMISSION

571

22 for example, is high compared to SO22 4 and hydrogenphosphate (HPO4 ), which 2þ may precipitate with Ca (Fenn and Hossner, 1985). Following precipitation of 22 and SO22 the anions HPO22 4 4 , carbonate (CO3 ) will dissolve and this will increase the pH of the soil solution (Larsen and Gunary, 1962; Fenn et al., 1981). The pH in calcareous soils will be between 7 and 8 and most of the total inorganic 22 carbonate (Total inorganic carbon (TIC) ¼ HCO2 3 þ CO3 þ CO2) will be in 2 the form of HCO3 ; thus the reaction in the soil will be as follows:

2 ðNH4 Þ2 X þ CaCO3 $ NH3 " þ NHþ 4 þ HCO3 þ CaX #

ð15Þ

2 ðNHþ 4 Þ þ HCO3 $ NH3 " þCO2 " þH2 O

ð16Þ

ðNH4 Þ2 X þ CaCO3 $ 2NH3 " þCO2 " þCaX # þH2 O

ð17Þ

As mentioned above, HCO2 3 is an alkaline anion and the precipitation of the 22 or SO acid anions (X ¼ HPO22 4 4 ) and dissolution of CaCO3 will increase pH, thereby increasing the potential for NH3 emission. It is, therefore, a general rule that emission is lower after application of fertilizers containing soluble anions than from fertilizers containing anions that may precipitate. 2 2 Hydrolysis of urea will produce a mixture of NH3, NHþ 4 , HCO3 and CO3 22 and this may increase pH, because NH3 and CO3 are conjugate bases 2 22 (pKa ¼ 9.48 for NH3/NHþ 4 and pKa ¼ 10.4 for HCO3 /CO3 ). Application of ammonium bicarbonate (ABC ¼ (NH4)HCO3) or ammonium carbonate (AC ¼ (NH4)2CO3)) may also increase soil pH due to the alkaline properties of NH3 and CO22 3 . In soil amended with urea or ABC, the pH in the microsites affected by the fertilizers will be . 8; at this pH a high proportion of TAN will be in the NH3 form and most of the TIC will be in the form of HCO21 3 . In consequence, each mole of emitted NH3 will increase the concentration of Hþ by one mole (Eq. (1)), and emission of CO2 will reduce the Hþ concentration by one mole (Eqs. (18), (19), Fig. 6). þ 2 CO22 3 þ H3 O $ HCO3 þ H2 O

ð18Þ

þ HCO2 3 þ H3 O $ CO2 " þH2 O

ð19Þ

Due to the combined NH3 and CO2 emission from these fertilizers, the NH3 emission may be very high after application of urea, ABC, and AB to soils (Roelcke et al., 2002). Alkaline fertilizers such as ABC, urea, and diammoniumphosphate (DAP) generally cause a much higher NH3 emission than neutral or acidic fertilizers such as AS, CAN, or monoammonium phosphate (MAP), because pH is a dominant factor controlling NH3 emission. Inorganic ammonium compounds 2 containing anions that are not producing precipitating calcium salts (NO2 3 , Cl ) will reduce pH and the emission will be signiﬁcantly lower than the emission from fertilizers precipitating with Ca ions (Fenn et al., 1981). The addition of, for example, calcium chloride, potassium chloride, sulfate, or triple superphosphate

572

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

pH of soil solution

10 9 8 7 6 5 0.025

NH3, mol ltr.–1

0.020 0.015

CO2: NH3 =1:1 CO2: NH3 =2:1 CO2: NH3 =0:1

0.010 0.005 0.000 0.00

0.01

0.02

0.03

NH3 emission, mol l –1 Figure 6 Effect of emission of NH3 and CO2 from a liquid solution of TAN (NH3 þ NHþ 4 ), TIC 2 21 22 (CO2 þ HCO2 , TIC 3 þ CO 3 ), and Cl . The initial concentrations are TAN 0.1 mol l 0.05 mol l21, and Cl2 0.06 mol l21, and pH is 6.85. It is assumed that either equal amounts (mol) of TIC and TAN (CO2:NH3 ¼ 1:1), 2 units of TIC to 1 unit of TAN (CO2:NH3 ¼ 2:1) or 1 unit of TAN (CO2:NH3 ¼ 0:1) is emitted.

reduces pH and in consequence will reduce NH3 emissions (Sloan and Anderson, 1995; Christianson et al., 1995; Fan et al., 1996; Ouyang et al., 1998; Goos et al., 1999). The typical pattern of accumulated NH3 emission from fertilizers will generally show NH3 emission rates in the following order ABC . urea . DAP . AS . CAN ¼ MAP due to the alkaline properties of the fertilizers and interaction of precipitation of anions (Du Preez and Burger, 1988; Whitehead and Raistrick, 1990; Sommer and Jensen, 1994; Ouyang et al., 1998; He et al., 1999; Zia et al., 1999; Roelcke et al., 2002). From soils with limited buffer capacity, the NH3 volatilization will increase at increasing application rates of ABC and urea, because pH will be high for a long period (Black et al., 1985; Fenn et al., 1987). However, 2 32 emission of NH3 from acid fertilizers (SO22 4 , PO4 , Cl etc.) will decrease at increasing concentrations of TAN in fertilized soils with limited buffer capacity (Avnimelech and Laher, 1977), because NH3 emission will immediately cause a reduction in pH locally (Eq. (1)). From soils high in

AMMONIA EMISSION

573

buffer capacity, soil pH is of importance irrespective of the concentration of TAN (Avnimelech and Laher, 1977). NH3 volatilization from granulated fertilizers and AA has been correlated with the total acidity and titratable acidity of soils (Izaurralde et al., 1987; Stevens et al., 1989; Sommer and Ersbøll, 1996). On non-calcareous soils, the accumulated NH3 losses (Nmax) from urea and CAN are positively related to soil pH and inversely to total acidity (Fig. 7), i.e., NH3 losses are closely related to both soil pH and Hþ buffering capacity of the soil (Avnimelech and Laher, 1977; Ferguson et al., 1984; Stevens et al., 1989). Therefore, for the purpose of predicting losses

Cumulated NH3 volatilization (% of N)

40 F=6.12pH-19.7, R2=0.70 30

20

10

0 3

4

5

6 pH(H2O)

7

8

9

Cumulated NH3 emission (% of N)

30 25 F= –0.05TA+22.4 R2=0.52

20 15 10 5 0 0

100

200 TA, mmol

300

400

500

kg–1

Figure 7 Cumulative NH3 volatilization related to soil pH and total acidity (TA) of the soil (adapted from Stevens et al., 1989; Sommer and Ersbøll, 1996; Whitehead and Raistrick, 1990).

574

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD F=0.245exp(0.5858 pHmax)R2 = 0.607

Cumulated NH3 volatiliztion (% of N)

60 50 40 30 20 10 0

3

4

5

6

7 pH max

8

9

10

Figure 8 Cumulative NH3 volatilization from urea and mineral fertilizers related to maximum pH after soil amendment (adapted from Whitehead and Raistrick, 1990; Sommer and Ersbøll, 1996).

from urea applied to the ﬁeld, the relation between Nmax and soil pH may be used, as shown in Fig. 7; this relationship explains about 70% of the variation in three different studies. The volatilization from ammoniacal mineral fertilizers may also be predicted on the basis of the relation between NH3 volatilization and pHmax (Fig. 8), but the pHmax of the soil surface after fertilizer application should be measured or estimated. More complicated mechanistic models presented by Avnimelech and Laher (1977); Rachhpal-Singh and Nye (1986a,b); Fleisher et al. (1987); and Kirk and Nye (1991) may be used for predicting the risk of volatilization when applying different types of fertilizers to the ﬁeld; the problem with these models is that the input data for the models may often be missing.

E.

SOIL CEC

TAN in soil will be partitioned between the three phases — liquid, solid, and þ gas phase. The NHþ 4 component in solution will be in equilibrium with NH4 on the solid phase exchange site (Fleisher et al., 1987). At pH , 8, more than 95% of the TAN (Fig. 2) will be of NHþ 4 form and can be exchanged with exchangeable cations (ex-C). Agricultural soils contain the divalent cations Ca2þ and Mg2þ with a high afﬁnity for adsorption; the exchange of NHþ 4 on exchange sites therefore can be deﬁned by use of the activity ratio law (Russel, 1977): þ NHþ 4 þ ex 2 C $ ex 2 NH4 þ C

NHþ Ex 2 NH4 4 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼K 2þ 2þ Ex 2 Ca; Mg Ca þ Mg

ð20Þ ð21Þ

AMMONIA EMISSION

575

It can be seen from Eqs. (20) and (21) that increase in cation exchange capacity (CEC) will increase NHþ 4 adsorbed to the soil (Avnimelech and Laher, 1977; McGarry et al., 1987; Whitehead and Raistrick, 1993). At low cation exchange capacity, adsorption of NHþ 4 will not affect NH3 emission; at CEC ca. 250 mmol kg21 the effect of absorption tends to be insigniﬁcant (O’Toole et al., 1985), whereas when CEC is above this level, NH3 volatilization is reduced 2þ 2þ signiﬁcantly. Further, NHþ 4 does not exchange easily with Ca , thus high Ca concentrations in the soil may reduce the effect of a high soil CEC. This will be the case in calcareous soils, where high concentrations of exchangeable Ca2þ will reduce the fraction of NHþ 4 absorbed. The consequence of the ratio law is that increasing soil water content due to rain will change the equilibrium and the divalent cations in solution will be exchanged with NHþ 4 (Chung and Zasoski, 1994). Conversely, if the solution is concentrated by water being removed due to drying, NHþ 4 will exchange with divalent cations on the CEC. Thus, during a drying event after fertilizer application, the concentration of NHþ 4 in solution will not increase linearly with evaporation of water, and in consequence the exchange process will reduce the rate of NH3 emission expressed relative to initial TAN content per time unit and increase the duration of the period with signiﬁcant emission rates; this effect is most pronounced in soil high in CEC, as can be deduced from the work of Fleisher et al. (1987). This retention to CEC during drying events may, in addition to the solid phase theory (see below), contribute to an explanation of why NH3 emission may be low from a soil that has been dried (Fenn and Kissel, 1976). Furthermore, Fenn and Kissel (1976) have shown that in very dry soils there is a physical adsorption of gaseous NH3 to soil.

F.

SOLID PHASE PROCESSES

Ammonia emission from DAP applied to timed soils tend to be lower than the emission from AS applied to soils high in calcium (Sommer and Ersbøll, 1996); the lower NH3 emission from DAP in Ca-rich soil has been ascribed to precipitation of calcium ammonium phosphate or magnesium ammonium phosphate (struvite) (Larsen and Gunary, 1962; Whitehead and Raistrick, 1990). The conditional stability constants for the solid phases of struvite may be calculated according to principles described in detail by Ringbom (1963) and Stumm and Morgan (1981); the reaction is as follows: 32 Mg2þ þ NHþ 4 þ PO4 $ MgNH4 PO4ðsÞ

ð22Þ

Thus, in the microsites with fertilizers, precipitation of struvite is mainly controlled by Mg2þ, phosphate, and pH, as NHþ 4 is present in large amounts.

576

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

Struvite precipitates and is stabile at pH . 6, and will consequently be present as stable solid phases above this pH level. Thus, struvite can be formed after addition of urea and ABC to most soils and in some soils after addition of acid fertilizers; the amount of struvite precipitation will, as mentioned above, be related to Mg2þ, phosphate, and pH. The precipitation of solid phases of ammonium will increase during a drying event, due to increasing concentrations of the ions in solution. Salts with hydrated water may in absence of “free” water, i.e., in a dry soil, donate a proton to NH3, transforming it into NHþ 4 (Evangelou, 1990). The loss of a proton produces a negative charge on the salt crystal, which facilitates the absorption of NHþ 4 . Thus besides the precipitation reactions proposed by Fenn et al. (1981), this mechanism may explain the reduction in NH3 emission from surface-applied ammoniacal fertilizers applied together with Ca or Mg salts.

G.

CLIMATE

AND INFILTRATION

The effect of climate on NH3 volatilization from urea differs signiﬁcantly from mineral fertilizers, because urea has to be hydrolyzed before volatilization will start, and hydrolysis is related to water availability and temperature. Thus after application of urea to a dry soil there will be no hydrolysis of the urea and no NH3 volatilization. A rain event immediately after surface application will dissolve the urea and, as urea movement lags only slightly behind the waterfront, the urea will inﬁltrate the soil (Fenn and Miyamoto, 1979). In the soil, urea is transformed to TAN, which will be absorbed by soil colloids; therefore, TAN is not as easily transported as urea (Black et al., 1987a). The effect of rain may therefore be variable and related to soil humidity and air dryness prior to the rain and also to the amount of water added during the rain event. More NH3 volatilizes after urea application on a wet soil than on a dry soil, because the humidity will initiate urea hydrolysis (Fenn and Miyamoto, 1979). An example of the interaction between application of urea and climate is shown in Fig. 9; in late March 1993 with no rain after application, ca. 9% of the N was lost, while only about 2% was lost in 1994 with a rain event of 13 mm immediately after application (Schjoerring and Mattsson, 2001). The 13 mm of rain in 1994 transported urea into the soil. The second round of urea application in late April resulted in a loss of 7 –8% of the applied N in both years (Fig. 9). However, due to very dry conditions in 1993, the NH3 emission was delayed for more than 2 weeks until a light shower was received mid-May, accelerating the dissolution of the fertilizer granules and the hydrolysis of the urea. The total amount of NH3-N lost was 13 and 9 kg N ha21 in 1993 and 1994, respectively. The effect of convective and diffusive transport of urea in the soil is important when predicting NH3 emission. Diffusion may reﬂect changes in TAN

NH3 emission, kg NH3-Nha–1 season–1

AMMONIA EMISSION 14

577

Winter wheat applied urea

12

14

1993 1994

12

10

10

8

8

6

6

4

4

2

2

0 rch Ma

0 ril Ap

e y Ma Jun

y Jul

t gus Au

Figure 9 Cumulated NH3 emissions ^ SE from urea applied to growing winter wheat crops. Vertical arrows denote time of urea application. The reason for the low loss in 1994 was that 13 mm of rain fell immediately after the urea was applied so that the urea was transported into the soil. Very dry conditions in 1993 delayed the NH3 emission for more than two weeks until a light shower was received mid-May accelerating the dissolution of the fertiliser granules and the hydrolysis of the urea.

concentration and pH around the fertilizer granules and convection the gross transport. Complex models predicting diffusive and convective transport of incorporated urea have been developed by Rachhpal-Singh and Nye (1986a,b) and by Kirk and Nye (1991). These models indicate that convective transport into the soil as affected by rain and transport to the soil surface during drying conditions are important for the prediction of the magnitude of NH 3 volatilization. Less than 10 mm of rain has little effect on the magnitude of volatilization (McInnes et al., 1986a; Ryden et al., 1987), and at rainfall greater than 10 –16 mm, losses are reduced if urea remains in a non-hydrolyzed form (Black et al., 1987b) and no loss is determined after 20 –25 mm of rain (Fenn and Miyamoto, 1979; Bouwmeester et al., 1985). For ammoniacal fertilizers, one may assume that about 20 mm of rain is sufﬁcient to reduce NH3 volatilization signiﬁcantly. Furthermore, transport of TAN into the soil may also be affected by soil humidity. Model simulations of Nye (1992) showed that emissions tended to increase when the moisture level was at about ﬁeld capacity and also when soil was drier. At high soil water content, diffusion of TAN in the liquid phase increases with increasing soil water, and in dry soil diffusion of NH3 in the vapor phase increases with reduction in water content compensating for the reduced TAN diffusion in the liquid phase. The rate of NH3 volatilization from urea will be affected by temperature because both hydrolysis rate and NH3 transfer from the liquid to the atmosphere increase with increasing temperature (Black et al., 1985; McGarry et al., 1987).

578

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

Thus, the lag phase will be shorter and the initial loss rates higher at high soil temperatures. Total losses may not be affected by changes in temperature, because volatilization will continue for a longer period at low than at high temperature (Fenn and Kissel, 1974; Harper et al., 1983); consequently total loss may be related to the potential of losses given by soil and other variables (McGarry et al., 1987). A low temperature in combination with rain may reduce losses signiﬁcantly, especially because hydrolyses is slow and urea may inﬁltrate the soil during showers (Fenn and Miyamoto, 1979). Thus, Harper et al. (1983) concluded that the amount and distribution of rain after urea application appeared to control the total NH3 volatilization from urea application. Ammonium phosphate or sulfate, which may precipitate with calcium, has a different relation to temperature than AN, which will not precipitate (Fenn and Kissel, 1973). The precipitate-forming fertilizers showed no differences in volatilization due to change in temperature at high (. 275 kg NH4-N ha21) and low (, 66 kg NH4-N ha21) application rates. Volatilization rates increased from 26 to 45% of TAN with a temperature increase from 12 to 328C, when the precipitate-forming fertilizers were applied to a calcareous soil at application rates of 110 kg NH4-N ha21. Volatilization from AN increased from 14 to 45% of applied TAN with a temperature increase from 12 to 328C and was not affected by changes in application rates (Fenn and Kissel, 1974). These patterns may be related to interaction of soil buffer capacity and concentrations of TAN as mentioned above; i.e., in soils at high buffer capacities and low NHþ 4 concentrations, the soil buffer will control emission, and at high NHþ 4 concentration, a fraction of the anions applied will be in solution and control emission of NH3. Wind will affect losses of NH3 from fertilizers; if humidity is high enough to keep the salts in solution, volatilization is expected to increase with wind (Eqs. (1) – (10)). In the few ﬁeld studies using non-interfering micrometeorological techniques it has, however, been shown that wind cannot be identiﬁed as the most important variable and wind has only been shown to affect the volatilization from urea signiﬁcantly during short periods (Harper et al. (1983), or the effect of wind speed has not been signiﬁcant during the study (McInnes et al., 1986a). These experiments may back up the assumption that the interaction between wind and soil temperature may confound the effect of wind.

H.

MICROBIAL PROCESSES (NITRIFICATION/IMMOBILIZATION)

Emission of NH3 from TAN supplied to the soil in form of urea or ammonium salts will compete with depletion of the TAN by microorganisms through immobilization to soil organic N or nitriﬁcation to NO2 3 (Malhi and McGill, 1982; Recous et al., 1992). The rate of TAN transformation by microorganisms

AMMONIA EMISSION

579

depends on a range of factors, including the population density of microorganisms, soil temperature, water and oxygen concentrations of the soil and inhibition due to high NH3 concentrations. This means that it is difﬁcult to generalize the contribution of microbial activity to the emission pattern of NH3, but the following general assessment is based on the initial period of 0– 14 days after application of ammoniacal fertilizers, i.e., in the period with signiﬁcant NH3 emission from the fertilizer-treated soil. Microbial transformation of TAN to organic N (assimilation/immobilization) takes place immediately after TAN is dissolved in soil solution. Rate of assimilation is related to availability of carbon in the soil rhizosphere, TAN concentration, soil water content and temperature (Recous et al., 1988; 1999). Few studies have quantiﬁed the relationship between environment and immobilization; however, it is assumed that immobilization is less responsive to temperature than, for example, mineralization (Murphy et al., 2003). Immobilization has been reported to account for 5% of the TAN applied within the ﬁrst 2– 3 days after spreading and 8– 10% after 14 days at an average soil temperature of 3.58C (Recous et al., 1988) and 3.5% during the initial 7 days at 238C (Sørensen, 2001). In humid sandy tropical soils, the rate of microbial immobilization was 12% of TAN applied during 50 days of incubation of urea or (NH4)2SO4 (Atwell et al., 2002), conﬁrming that the response to temperature is not very signiﬁcant and the assimilation by soil microorganisms is relatively low both in temperate and tropical soils. Furthermore, soil type seems not to affect immobilization, as similar immobilization rates have been reported for different soils in rotation after a wheat crop, probably because immobilization is limited by readily assimilable carbon available for the soil microﬂora (Recous et al., 1992). Nitriﬁcation of TAN mixed with soil is largely affected by soil pH, being negligible at soil pH values lower than ca. 4 and increasing linearly with pH increasing from 4 to 6.2 in a sandy soil (Winter and Eiland, 1996). Increasing TAN (substrate) concentrations increases nitriﬁcation until either salt effects or NH3 toxicity reduces nitriﬁcation (Malhi and McGill, 1982). Thus, at a level 21 (soil), nitriﬁcation has been shown to be higher than 250– 300 mg NHþ 4 -N g inhibited after mixing (NH4)2SO4 in soils having a pH of 5.8– 7 (Malhi and McGill, 1982; Flowers and O’Callaghan, 1983); the inhibition was primarily a salt effect, as the NH3 fraction of TAN is low in soils at pH , 7. At high pH, the concentration of the NH3 component of TAN will increase and reach levels toxic to the microﬂora (Malhi and McGill, 1982). Nitriﬁcation can be expected at “permanent wilting point” and increases with increasing water content (Malhi and McGill, 1982; Flowers and Callaghan, 1983). Near water saturation, nitriﬁcation is absent due to a shortage of oxygen (Malhi and McGill, 1982). The nitriﬁcation rate increases linearly with soil water content (zero order process; Flowers and O’Callaghan (1983) and is ca. 2.8 times higher at 0.033 MPa than at 21.5 MPa, respectively, at ﬁeld capacity and permanent wilting point (Malhi and McGill, 1982).

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

2.5˚C 25˚C

5 4

7.5

3 2

Soil I Soil II Soil III

7.0 Soil pH

Nitrification rate, micog–1g(soil) d–1

580

1

6.5 6.0 5.5

0

5.0 0

10

20 30 40 50 Temperature, ˚C

60

0

100 200 300 Concentration of TAN, microg TAN-N g–1(soil)

Figure 10 Left: Measurements of nitriﬁcation in Alberta soils from an area with a mean annual air temperature of 2.58C (Malhi and McGill, 1982) and North Australian soil with a mean annual air temperature of 258C (Myers, 1975). The ﬁgure is inspired by Malhi and McGill (1982). Right: Change in soil pH after completion of nitriﬁcation of (NH4)2SO2 mixed with the soil and no NH3 emission. Soil I: gray luvisol; Soil II: dark gray chernozemic, and Soil III: black chernozemic (Malhi and McGill, 1982).

Temperature dependence of nitriﬁcation is different in warm or tropical soils and temperate soils (Fig. 10), however 15N pool dilution studies have shown that nitriﬁcation may be very rapid in most climatic regions (Watson et al., 2000; Watson et al., 2002 quoted in Murphy et al., 2003). Daily nitriﬁcation rates of 2.2 – 7.9% for an acid sandy soil and a sandy loam at 238C have been reported by Sørensen (2001) and 4– 6% at 58C by Flowers and O’Callaghan (1983), at 21 (soil). These rates substrate concentrations of 50 – 100 mg NHþ 4 -N g correspond to the nitriﬁcation rates calculated by Malhi and McGill (1982) using a simple ﬁrst-order model with input of substrate concentration, temperature, and water content of the soil. A generally applicable model will need inclusion of the effects of pH and changes in the population of nitriﬁers (Gilmour, 1984; Grant, 1994). Immobilization may cause reduction in NH3 emission, but the effect is considered to be negligible because the process is relatively slow compared with volatilization from applied mineral fertilizers, i.e., daily immobilization rates is 0.5 – 1.3% of the TAN or urea applied compared to NH3 emission rates of 5 –10% immediately after hydrolysis of applied urea or dissolution of ammonium salts (see Fig. 4). The studies also indicate that immobilization of NO2 3 is very slow (i.e., about 5% of the immobilization rate of TAN; Mary et al. (1998)) and may therefore not be accounted for in NH3 emission studies. Contrary to immobilization, nitriﬁcation should be accounted for when interpreting NH3 emission experiments or when developing models to predict NH3 emission.

AMMONIA EMISSION

581

Nitriﬁcation may affect NH3 emission through reduction in TAN and by þ reducing pH, as nitriﬁcation of 1 mol NHþ 4 produces 2 mol H , according to the following equation: 21 þ NHþ 4 þ 2H2 O $ NO3 þ 2H3 O

ð23Þ

The study of Flowers and O’Callaghan (1983) showed that nitriﬁcation of 21 100 mg NHþ (soil) resulted in a reduction of 0.5 pH units and nitriﬁcation 4 -N g þ of 250 mg NH4 -N g21 (soil) resulted in a reduction of 1 pH unit. Nitriﬁcation, therefore, reduces NH3 emission due to reduction in both concentration of TAN in soil solution and a reduction in the NH3 component of TAN. Figure 10 depicts the change in soil pH measured in the study of Malhi and McGill (1982), showing that the change in pH is related to amount of added TAN and differences in soil buffering capacity. The addition of fertilizers may cause a local change in soil pH, the magnitude of the change being related to the concentration of fertilizer, soil pH –buffer capacity, and fertilizer type. In consequence of the change in pH due to application of the fertilizers, the nitriﬁcation rates was in the following order urea . DAP . AS in the study of McInnes and Fillery (1989).

IV. AMMONIA EMISSION FROM CROP FOLIAGE A.

TRANSPORT

OF AND THE

NH 3 BETWEEN LEAVES ATMOSPHERE

In plants, the major source of NH3 is TAN dissolved in the water ﬁlm in the mesophyll cell walls of leaves (the so-called apoplastic solution; Husted and Schjoerring, 1995). The concentration of TAN and Hþ is affected by uptake of N, translocation, and transformation of N, which varies with plant developmental stage, climate, and fertilization. The NH3 ﬂux, FNH3, between a single plant leaf and the atmosphere can be described as: F NH3 ¼ gleaf ðx 2 NH3;a Þ

ð12Þ

where gleaf is the conductance to diffusion of NH3 between the atmosphere and the interior of the leaf, and x is the NH3 concentration of the air in the substomatal cavities and intercellular air spaces within the leaf. Whether a leaf will act as a sink for or a source of atmospheric NH3 depends on the difference in internal and external NH3 concentration. If NH3,a exceeds x, NH3 will be absorbed, while in the opposite case emission will occur. When NH3,a equals x, no net NH3 ﬂux occurs between the leaf and the atmosphere. The internal NH3 concentration at which FNH3 is zero is called the stomatal compensation point for NH3

582

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

(Eq. (12); Farquhar et al., 1980; Husted et al., 1996). As evident from Eq. (12), the rate and direction of NH3 ﬂuxes between plant leaves and the atmosphere at a given atmospheric NH3 concentration are controlled by the conductance to NH3 transfer and the internal NH3 concentration. The leaf resistance (inverse of conductance) to NH3 transfer includes a stomatal and cuticular resistance in parallel and usually also includes a boundary layer resistance in series with the two other terms. The latter is a function of the aerodynamic properties of the leaf, wind speed, and canopy turbulence, and will generally be an order of magnitude smaller than the maximum stomatal resistance. The cuticular resistance (Rcut) is extremely high for NH3, probably in the range of 2000 – 40,000 s m21 (van Hove et al., 1989). This means that hardly any NH3 will pass through the cuticle. However, NH3 can readily be deposited on the cuticular surface due to the presence of a surface water ﬁlm. Since NH3 is highly soluble in water, moist leaf surfaces can act as a storage compartment for atmospheric NH3, making the leaf surface a temporary sink for NH3.

B.

MAGNITUDE

OF

NH 3 LOSSES

USA, 1985

16

4 2

DK,1993

DK,1994

6

DK, 1994

8

DK,1993

10

UK

12

DK, 1993-1994

14

UK DK,1989-1990

Ammonia emission,kg NH3-Nha–1

In general, NH3 is emitted from intensive agricultural ecosystems (Fig. 11), while semi-natural ecosystems act as NH3 sinks (Sutton et al., 1993a; Schjoerring et al., 1998). The wild understory plant, Luzula, and three native grass species showed low NH3 compensation points of 0.5 – 2 nmol mol21 (Hanstein et al., 1999; Hill et al., 2001). Crop plants such as oilseed rape or barley usually show

0 Barley

Wheat

Oilseed rape

Figure 11 Seasonal NH3 emissions from agricultural crops. Data from Harper et al. (1987); Mattsson and Schjoerring (2001); Sutton et al. (1993b); Schjoerring et al. (1993); Yamulki et al. (1996).

AMMONIA EMISSION

583

NH3 compensation points between 2 and 6 nmol mol21 (Husted and Schjoerring, 1995, 1996; Husted et al., 1996; Mattsson et al., 1997). In barley, the NH3 compensation point changed in relation to the developmental stage although the plants were grown under constant N limitation (Husted et al., 1996). In a ﬁeld experiment in the Netherlands, intensively managed ryegrass showed NH3 compensation points varying over the season between 1 and 7 nmol mol21 (van Hove et al., 2002). For the same crop species grown in Scotland, Loubet et al. (2002) monitored NH3 compensation points ranging from 0.02 mg NH3 m23 in periods between fertilizations up to 10 mg NH3 m23 just after fertilizations. Under laboratory conditions, both ryegrass and Bromus erectus showed very high NH3 compensation points up to 18 nmol mol21 particularly when supplied with high levels of NHþ 4 to the growth medium (Mattsson and Schjoerring, 2002). Quantiﬁcation of the NH3 exchange between the atmosphere and the canopy of barley, wheat and oilseed rape over two growing seasons show that the crop foliage is a net source of NH3 to the atmosphere, with emissions ranging between 1 and 5 kg NH3-N ha21 year21 (Fig. 11). Very high NH3 emissions of up to 15 kg NH3-N ha21 per season were reported for winter wheat in the United States by Harper et al. (1987). Harper et al. (1996) and Plantaz (1998) also measured high daily NH3 emissions from grassland in the Netherlands during spring and summer and from these emissions, high NH3 compensation points were derived. In contrast, based on measurements of stomatal NH3 compensation points by apoplastic bioassay, van Hove et al. (2002) concluded that plants in intensively managed grasslands in the Netherlands would not contribute to atmospheric NH3 loadings. For wheat, oilseed rape, and barley, the accumulated NH3 loss over a growing season constituted between 1 and 4% of the applied N, or between 1 and 4% of the total shoot N (Schjoerring and Mattsson, 2001). The loss increased under conditions with a high N concentration in the foliage and was positively correlated with the above-ground crop N content at anthesis, but not with that at ﬁnal maturity. There were no indications that NH3 emissions were larger under conditions unfavorable for nitrogen remobilization from vegetative plant parts (low N harvest index). Nevertheless, a distinct peak in NH3 emission occurred during senescence. NH3 emissions from plant stands, measured under simulated environmental conditions in wind tunnels, ranged between 0.8 and 1.4% of the N content of the shoot, equivalent to 1.1– 2.9 kg NH3-N ha21 (Mannheim et al., 1997). The highest emissions were observed in faba beans, whereas the emissions in winter wheat, spring rape, and white mustard were lower. The total NH3 emissions were not affected by removing a part of the ears (sink reduction), but emissions occurred earlier, as did the plant senescence. This suggests that the NH3 emissions are closely related to senescence (Schjoerring et al., 1993; Mannheim et al., 1997). Emission of NH3 has also been suggested to contribute to the decline in shoot nitrogen content that is often observed in agricultural crops

584

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

during the generative growth stage (Wetselaar and Farquhar, 1980; Schjoerring et al., 1989; Francis et al., 1993). In this context, it is important to emphasize that assessments of NH3 losses based on measurements of changes in 15N-labeled or total above-ground N are indirect and other pathways of nitrogen loss may inﬂuence the results, resulting in overestimation of losses (see below in Section VI on Measurement techniques). It can be concluded that plant communities on agricultural cropland represent a net source of NH3 to the atmosphere. Net emissions range from below 1 up to 15 kg NH3-N ha21 per season, depending on plant species, crop nitrogen economy status, and climatic conditions. Crops will in many areas represent a signiﬁcant input of NH3 to the atmosphere and NH3 losses may become large enough to signiﬁcantly affect crop N budgets.

C.

PHYSIOLOGICAL PROCESSES INVOLVED FROM CROPS

IN

NH 3 EMISSION

In addition to being taken up directly from the soil, NH3/NHþ 4 is produced in plant tissues by a number of different metabolic processes. For example, NHþ 4 is produced via nitrate reduction, through the ﬁxation of atmospheric nitrogen by root nodules, by photorespiration in leaves, and through the phenyl propanoid pathway (Hirel and Lea, 2001). Ammonium may also be released during reassimilation of nitrogen transport compounds (e.g., asparagine, glutamine, arginine, and ureides) and through breakdown of other N compounds during senescence and remobilization. These processes take place in different cell organelles and in different tissues leading to both spatial and temporal variation in tissue NHþ 4 concentrations. It is therefore difﬁcult to predict the proportion of the NHþ 4 present in plant leaves that at any stage will be contributing to the emission of NH3. A further complicating factor is that environmental conditions, particularly nitrogen supply, also affect tissue NHþ 4 levels (see below).

1. NH3 Compensation Point The major route for NH3 exchange between plants and the atmosphere is through the leaf stomates. In order for NH3 to be emitted through the stomates, the concentration inside the leaf (i.e., compensation point) has to be higher than the ambient concentration (see Section IV.A). The leaf apoplast constitutes the interface between the atmosphere and the living leaf tissue, and the NHþ 4 concentration in the liquid phase of the apoplast (the water contained within the cell walls) is therefore a critical parameter in determining the gaseous NH3 concentration inside the leaf. The stomatal NH3 compensation point can be

AMMONIA EMISSION

585

calculated from measurements of the pH and NHþ 4 concentration in the apoplastic solution (Eqs. (3) and (4); Husted and Schjoerring, 1995; Mattsson and Schjoerring, 2002). Apoplastic NHþ 4 concentrations normally range between 0.01 and 2 mM but have been shown to be highly dynamic and closely coupled to the plant N metabolism (Mattsson et al., 1998; Nielsen and Schjoerring, 1998; Mattsson and Schjoerring, 2002).

2.

N Uptake and Translocation

There are different pathways for NHþ 4 to reach the apoplastic solution; either þ through xylem transport of NH4 from the root or through NH3 efﬂux from the mesophyll cells. Xylem transport of NHþ 4 can be quite substantial, particularly if the plants are grown on high levels of nitrogen. The concentration of NHþ 4 in xylem sap increases with increasing supply of both NHþ 4 (Mattsson et al., 1998) þ and NO2 3 (Husted et al., 2000a). These increasing xylem sap NH4 concentrations þ are usually reﬂected in the apoplastic NH4 concentration (Mattsson et al., 1998; Finnemann and Schjoerring, 1999; Husted et al., 2000a). Inside the leaf cells, NHþ 4 can either be assimilated by cytosolic glutamine synthetase (GS1) or taken up into the chloroplasts and assimilated by the chloroplastic form of the enzyme (GS2). Depending on the capacity of these enzymes to assimilate NHþ 4 into glutamine and then by the enzyme glutamate synthase (GOGAT) to convert glutamine into glutamate, NHþ 4 may accumulate in the leaves. In general, high NHþ 4 concentrations in xylem sap and apoplast also result in a high leaf tissue extract NHþ 4 concentration (Mattsson et al., 1998), but there are also examples where leaf tissue NHþ 4 did not rise despite increasing xylem and apoplast þ concentrations (Husted et al., 2000a). In grasses grown with either NO2 3 or NH4 , þ signiﬁcant correlations between leaf tissue NH4 concentration and both NH3 emission and apoplastic NHþ 4 concentration were observed (Mattsson and Schjoerring, 2002). Decreasing the activity of GS, either by using an inhibitor such as MSX or by using mutants or transgenic plants with lower assimilation capacity, usually results in increasing concentrations of NHþ 4 in various plant compartments. In consequence, NH3 emission to the atmosphere increases within a few hours after adding MSX to the growth medium in both barley and oilseed rape, because GS activity is decreased and tissue TAN concentrations increase (Husted and Schjoerring, 1995; Mattsson and Schjoerring, 1996). In barley mutants with reduced activity of chloroplastic GS, higher leaf tissue and apoplastic NHþ 4 concentrations have resulted in higher NH3 emission compared with wild-type plants (Mattsson et al., 1997). Emission of NH3 also seems to increase more with increasing temperatures in the mutants than in the wild-type plants, suggesting a higher sensitivity to photorespiration in GS mutants. A massive release of

586

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

NH3/NHþ 4 takes place during photorespiration, particularly at high temperatures (Leegood et al., 1995), making it extremely important for the plant to have an efﬁcient reassimilation of this NH3. The inﬂuence of photorespiration on apoplastic and leaf tissue NHþ 4 concentrations was investigated in antisense GS2 oilseed rape plants, i.e., plants with reduced activity of the key enzyme responsible for re-assimilation of photorespiratory NHþ 4 wild-type plants (Husted et al., 2002). Despite a 50– 75% lower in vitro leaf GS activity in the antisense plants, there was no tendency for these plants to have higher tissue NHþ 4 concentrations than wild-type plants. Antisense plants exposed to leaf temperatures increasing from 14 to 278C or to a three-fold increase in the O2/CO2 ratio did not show any change in steady state leaf tissue NHþ 4 concentration or in NH3 emission to the atmosphere. These results show that oilseed rape has a large surplus of GS in the leaves, which makes the plants less sensitive to increasing photorespiration than is the case for barley (Mattsson et al., 1997). The inﬂuence of N2 ﬁxation on the NH3 emission potential has been investigated in white clover (Herrmann et al., 2002). Symbiotic N2 ﬁxation and mineral N acquisition were shown to be well balanced with respect to the apoplast and plant tissue NHþ 4 concentrations, leading to equal NH 3 compensation points in plants grown with or without an external NO2 3 supply. Field measurements over pea crops similarly indicate NH3 emissions comparable to those for barley and wheat (Schjoerring and Mattsson, 2001). Senescing plant material undergoes a sequence of biochemical and physiological events including protein degradation, which eventually leads to cell death. The massive release of NH3 in senescing material is thought to be a consequence of amino acid deamination and catabolism of nucleic acids (Brouquisse et al., 2001). Dark-induced senescence of barley plants and detached leaves of oilseed rape both showed increased NH3 emission which was synchronized with chlorophyll degradation and liberation of NHþ 4 in the leaf tissue (Schjoerring et al., 1998).

V. MANAGEMENT STRATEGIES Fertilizers are spread mainly as granules or prills. The fertilizers are granulated and prilled because this reduces inhomogeneity and because these fertilizers can be spread evenly. Furthermore, additives mixed with the fertilizers can produce slow-release fertilizers or reduce transformations after application as, for example, with urea. Fertilizers are either broadcast onto the soil surface or injected into the soil. AA is usually injected or can be dissolved in irrigation water, and urea is sometimes incorporated for the purpose of reducing NH3 emission. In Asia, farmers broadcast nitrogen fertilizer (mainly urea) into the water in rice paddy ﬁelds.

AMMONIA EMISSION

A.

TECHNIQUES

FOR

REDUCTION

OF

587

NH 3 EMISSION

10 5

Placed 2 cm

15

Placed 2.5 cm

20

Mixed 3 cm

25

Mixed 2 cm

30

Placed 1 cm

Broadcast

35 Broadcast

Ammonia emission, NH3 pct. of applied N

Top dressing and partly covering urea may reduce emissions signiﬁcantly (Fig. 12). The effect of placing urea is related to the depth and soil characteristics (Fenn and Kissel, 1976). Placing urea in a soil low in CEC and with low pH buffering capacity may create a zone of high concentrations of dissolved TAN and a high pH due to hydrolysis. In consequence, the concentration of NH3 may be very high in the zone affected by urea. In these soils, shallow placement of urea may have little effect on NH3 emission (Fig. 12) because the NH3 will be transported by diffusion to the surface and be lost (Blaise et al., 1996). Thus there is a high correlation between amount of urea applied and placement depth and emission, so at high application rates urea should be placed at greater depths than at low application rates. To eliminate NH3 emission from urea applied to calcareous soils, the fertilizer may need to be placed as deep as 5 –7.5 cm (Fenn and Miyamoto, 1981; Ismail et al., 1991). Harrowing stubble before urea application may halve NH3 volatilization, because cultivation forms cracks and small hollows where urea prills will be protected from volatilization, and rain events or irrigation may leach urea into the soil (Bacon et al., 1986; Bacon and Freney 1989). Mixing ammoniacal fertilizers with the soil may be a less efﬁcient reduction measure than injection to the same depth because a part of the mixed-in fertilizer will be close to the surface and TAN will be transported either by diffusion or convection upwards and be lost (Nye, 1992). Increasing the application rate may reduce the relative emission of NH3 from urea and ammonium fertilizers applied to calcareous soils (Du Preez and Burger, 1988). On acidic soils the proportion of urea-N lost due to NH3 emission has been

0 Wind tunnel

Dynamic chamber

Figure 12 Ammonia emission from urea broadcast or placed at different depth to a bare soil (Bouwmeester et al., 1985; Nye, 1992).

588

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

0

UAN

Urea

Urea

AS Urea

W. wheat, ear initiation, Z 30

W. wheat seeding, Z 05 W. wheat, tillering, Z 14.22

Trash

Trash - heavy rain

Banana plantation

Trash - dry

10

Disced-scarified

20

Coarse silt Calcareous

30

Fine silt loam

40 Fine silt loam

Ammonia emisson, NH3 pct. of applied N

shown to increase by increasing the application rate (Black et al., 1987; Watson and Kilpatrick, 1991). This discrepancy is due to interaction of fertilizer alkalinity and soil acidity. In acid soils, the pH buffer may prevent high increases in pH after addition of urea at low rates, whereas pH may increase in acid soils amended with urea at high rates because the amount of alkaline hydrolysis products will be higher than microsite acidity of the soil. Similarly, surfaceapplication of urea in bands may increase NH3 emission compared to broadspreading when applied to acid soils. Although, absorption of NHþ 4 may affect transport of TAN in the soil, model simulations of Nye (1992) have shown that soil pH buffer capacity is more important in inﬂuencing NH3 emission. Application of pelleted fertilizers or fertilizers in solution may contribute to local “hot spots” with high salt concentrations and pH different from the surrounding soil. In a hot spot with urea, the pH will be high compared with the pH of the surrounding soil, and in a hot spot with “acid” ammonium salts pH will be low. Immediately after band application the high urea concentrations may reduce urease activity due to substrate inhibition (Fig. 5) and as the uncharged urea may inﬁltrate more easily to greater depths than NHþ 4 this may result in reduced emission of NH3 (Fenn and Miyamoto, 1979, Bouwmeester et al., 1985). Straw has a much higher pH buffer capacity and pH than soil and a 20-times higher urease activity than the surface 10 mm of soil; consequently broadcasting urea to straw spread on the soil may contribute to a high potential of NH3 emission (McInnes et al., 1986b). Similar, applying urea to trash left from harvesting of sugarcane may cause signiﬁcant losses of NH3 (Fig. 13) because the urea is not in

Urea

Figure 13 Ammonia emission from UAN (urea ammonium nitrate), urea and AS (ammonium sulphate) applied to bare soil, sugarcane trash-covered soil, to winter wheat, and banana plantation. (Adapted from McInnes et al., 1986a,b; Bouwmeester et al., 1985; Bacon and Freney, 1989; Freney et al., 1992; Prasertsak et al., 2001).

AMMONIA EMISSION

589

contact with soil and the trash has a high pH. Moving the litter aside on no-till soil and applying urea to the soil surface may reduce losses signiﬁcantly (Touchton and Hargrove, 1982).

B.

FERTILIZER COMPOSITION

Ammonia emission from fertilized soils has been reported to vary between negligible and 40% when measured with the micrometeorological mass balance technique (Fig. 13; Freney et al., 1992; McInnes et al., 1986a). The emission from a speciﬁc fertilizer will vary with climate, soil pH – buffer capacity, and CEC; therefore the average values shown in Figs. 13 and 14 cover considerable variation, as indicated by the relatively high standard variation. Ammonia emission is affected by the choice of fertilizer and rate of application. Fertilizers applied to calcareous or limed soils should preferably be acidic fertilizers with anions not forming calcium precipitates, e.g., AN or NH4Cl. Soils low in Ca may be fertilized with MAP, AS, or AN, which will reduce pH upon dissolution in soil water. Surface application of urea to acidic soils with a high pH – buffer capacity or a high CEC may not cause high losses. Urea should not be surface-applied to calcareous soils because TAN is at risk of being lost due to NH3 emission; therefore, incorporation would be recommended.

Ammonia emission, NH3-N pct. of applied N

40

30

20

10

0 Urea

DAP

CAN

AS

Figure 14 Ammonia emission from ammoniacal fertilizers, i.e., urea, diammoniaum phosphate (DAP), calcium ammonium nitrate (CAN) and ammonium sulfate (AS) broadcast to crops, measured with wind tunnels (Sommer and Jensen, 1994; Velthof et al., 1990; van der Weerden and Jarvis, 1997).

590

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

An example of adapting fertilizer application to soil and crop is banana plantations in a tropical area in Queensland (Australia), where 17% of the applied urea-N may be lost due to emission of NH3 (Prasertsak et al., 2001). The fertilizers cannot be incorporated because this will damage the roots; therefore, Australian banana producers are recommended to switch from urea to AN, with a loss potential lower than that of urea (Prasertsak et al., 2001). On grassland, incorporation of solid fertilizers may also be detrimental to the plants, and on acid soils application of liquid urea ammonium nitrate instead of granular urea may reduce losses because the extent of microsites with high TAN concentrations is reduced. Lightner et al. (1990) showed that applying urea in liquid solution may reduce emission, with 35% compared with granular-applied urea. Application of inorganic salts with fertilizers may reduce NH3 emission signiﬁcantly. The salts used should be soluble, thus CaSO4 is not efﬁcient for reducing NH3 emission. Adding CaCl2 with urea will contribute to reduction in 2 þ NH3 emission, because CO22 3 will precipitate as CaCO3 and Cl and NH4 will reduce pH signiﬁcantly as NH3 is lost. KNO3 or KCl may be used with urea to reduce NH3 emission, the Cl2 may contribute to produce an acid environment, reducing the NH3 emission potential. In addition, Kþ cation may exchange with absorbed Ca2þ, which will precipitate the CO22 produced during hydrolysis 3 (Fenn et al., 1982). Sulfur applied with urea will hydrolyze to SO22 4 and reduce NH3 emission, but the effect may not be very signiﬁcant in all environments, i.e., NH3 emission from sulfur-coated urea was equal to or somewhat lower than from untreated urea (10.1% as compared with 12.6%) in the study of Black et al. (1985), and natural ﬁnely ground sediments of pyrite (FeS) have been shown to reduce emission from urea by 54% (Blaise et al., 1996). Thiosulfate at 10% reduces hydrolysis rates and, under drying conditions, CaCl2 may also reduce hydrolysis (Black et al., 1985); the reduction is due to the low pH of the Cl2 (Fig. 5). In dry conditions, Cl concentration will be high due to slow diffusion and the pH reduction will be signiﬁcant (Sloan and Anderson, 1995). Addition of H2SO4 or H3PO4 with urea to calcareous soils is questionable due to precipitation of the anions with calcium. In calcareous soils, adding acids that precipitate should, therefore, be avoided and either HCl or HNO3 may be used with urea to reduce losses of NH3 (Fenn and Richards, 1986). However, Fenn and Hossner (1985) do not recommend the use of these acids because urea nitric acid is a cold explosive in solid form and HCl may be noxious and injurious to plants. Furthermore, adding pyrite and H2SO4 may contribute an increase in the soil’s demand for lime, which will be an additional cost for the farmer. Urease inhibitors delaying hydrolysis of urea to TAN may be an option for reducing NH3 emission. Slowing down the hydrolysis allows time for urea to inﬁltrate into the soil by diffusion or convection after a rain event, because the uncharged urea is more readily transported in soil than the charged NHþ 4 (Fenn and Miyamoto, 1981). Of the inhibitors tested, the phosphoryl di- and

AMMONIA EMISSION

591

triamides, (N-(n-butyl)thiophosphorictriamide (NBPT) or phenylphosphorodiamidate (PPD) were found to be the most useful, and have proven effective to varying degrees depending on the environment and management (Byrnes and Freney, 1995). The efﬁciency is related to climate and soil conditions; i.e., applying the urea with urease inhibitor to a humid soil during a period with little rain may not reduce NH3 emission, whereas application followed by a heavy rainfall few days after application may prove very efﬁcient. To become efﬁcient, NBPT must be converted to the oxygen analogue (N-(n-butyl)phosphorictriamide, NBPTO); on the other hand, PPD is efﬁcient immediately but is decomposed (departs from neutrality) rapidly (Bremner, 1995). In waterlogged soil or in ﬂoodwater in rice ﬁelds, the oxidation of NBPT applied with urea will be retarded and the inhibition of urea hydrolysis will be inefﬁcient (Bremner, 1995). Therefore, Phongpan et al. (1995) applied urea with PPD and NBPT to rice ﬁelds. It appeared that initially PPD inhibited urease activity, and during this time at least part of the NBPT was converted to NBPTO; then as the activity of PPD declined, NBPTO inhibited the hydrolysis of urea. This combined urease inhibitor treatment reduced NH3 emission from 15 to 3% of the applied N. Hydroquinone is a less effective urease inhibitor than the phosphoramides but may have beneﬁcial physiological effects on the plant (Bremner, 1995). Due its low cost, hydroquinone has been tested and is formulated with urea for use in China. The concentration of a chemical inhibitor required to suppress hydrolysis decreases with increasing granule size of the fertilizer, as seen with nitriﬁcation inhibitors (Singh et al., 1994). Thus, increasing granule size of the urea fertilizer may delay hydrolysis signiﬁcantly because the contact to soil is reduced. Black et al. (1987a) reported that urea concentration was signiﬁcant for 5 days when applying 8 mm granules, compared with 2 days after application of powder or , 4 mm granules. The pH in the granules increased to values between 8 and 9 within 2 days in both the small and large granules. Consequently, urea was hydrolyzed into a microsite with high pH and high TAN concentration for a longer period in the large granule than in small granules, causing the cumulated emission to be large. This was conﬁrmed by the study of Watson and Kilpatrick (1991), who found little difference in NH3 emissions between different sized prills of applied urea.

C.

FLOODED FIELDS (RICE PADDIES)

The fertilizer application to rice paddies is highlighted, because rice is the world’s most important food crop and 75% of the world’s rice is grown in several inches of water held in small dikes, i.e., paddies (Maclean, 1997). Management of the rice paddies may brieﬂy be characterized by ﬂooding of the dry paddy, puddling, transplanting of rice, vegetative and reproductive phases of rice,

592

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

and harvesting. Nitrogen uptake efﬁciencies range from 20 to 60% of applied N (Maclean, 1997). This low efﬁciency reﬂects poor management and the use of urea-N, leading to high NH3 emissions. Thus, reducing NH3 emission may act to optimize N-use efﬁciency. The water in paddies is buffered by bicarbonate (HCO2 3 ), which is the main contributor to the alkalinity of the water. Therefore, consumption of CO2 during daytime and respiration of CO2 during the night will cause large ﬂuctuations in the pH of the water in rice paddies with a high photosynthetic biomass, especially algae (Fillery et al., 1986; Bowmer and Muirhead, 1987). The uptake of CO2 will increase pH during the daytime and respiration of CO2 will reduce pH at night. The diurnal variation in NH3 emission clearly reﬂects the variation in pH (Fillery et al., 1986). Alkalinity may be signiﬁcantly higher in paddies irrigated with ﬂoodwater or alkaline well water compared with paddies irrigated with rainwater (Vlek and Crasswell, 1981; Fillery et al., 1984). The alkalinity and pH of the water will be affected by the applied fertilizers; thus (NH4)2SO4 decreased soil water pH ﬂuctuation initially, but after 2 days pH ﬂuctuations were similar to variations in pH of ﬂoodwater fertilized with urea in the study by Fillery et al. (1984). This indicates that after a few days, the HCO2 3 of the water, including CO2 from respiration, had buffered the pH to the initial values. In contrast, emission of NH3 from (NH4)2SO4 applied to water with low alkalinity would have resulted in a reduction in pH, which eventually would have reduced emission signiﬁcantly (Vlek and Crasswell, 1981). Hydrolyzed urea will contribute to the alkalinity and may increase pH to 10 in weakly buffered water (Mikkelsen et al., 1978), and the NH3 emission may be signiﬁcant. Removing the algae biomass or reducing the photosynthetic activity will reduce the diurnal variation in pH of the water, and the addition of a suitable photosynthetic inhibitor (terbutryne) with fertilizer may reduce the daytime pH of rice ﬂoodwater for up to 6 days and the potential NH3 emission by 43% (Bowmer and Muirhead, 1987). Adding urease inhibitors will delay hydrolysis of urea and reduce the concentration of TAN in the water, thereby reducing emission of NH3 (Fillery et al., 1986). The effect of urease inhibitors is variable and related to water chemistry. Mixing of the water due to wind and changes in buoyancy due to heating and cooling of surface water will vary during the day. Stratiﬁcation of subsurface water may reduce emission in periods with heating of surface water and little mixing caused by wind. Radiation during the daytime will heat the surface water, which will enhance stratiﬁcation, and cooling in the afternoon will lead to enhanced mixing (Leuning et al., 1984). In consequence, mixing may contribute to the high NH3 emission rates in the afternoon. Application of acidic fertilizers will reduce NH3 emission. Fig. 15 shows that emission of NH3 from urea may be higher than emission from (NH4)2SO4 when applied at transplanting. In Fig. 15, data from different studies have been used and comparing results from different studies may bias conclusions due to the

AMMONIA EMISSION

593

Panicle, IRRI

At transplanting, IRRI

Incorporation in water, IRRI

10

At transplanting, NWS

20

Panicle, IRRI

30

Incorporation dry soil, IRRI

40

Incorporation in water, IRRI

50 At transplanting, IRRI

NH3 emission, % of applied N

60

0 Urea

(NH4)2SO4

Figure 15 Ammonia emission from rice paddies, measured with the micrometeorological mass balance technique at either IRRI (Philippines) or in New South Wales (Australia). (From Leuning et al., 1984; De Datta et al., 1989, 1991; Fillery and De Datta,1986; Fillery et al., 1986; Freney et al., 1981).

effect of different environmental conditions (e.g., effect of alkalinity, wind, temperature, etc.), but the experiments were carried out on similar soils and wind ﬂuctuations were similar in the studies of Freney et al. (1981) and De Datta et al. (1991). Paddy rice in the tropics is fertilized with urea, which may be applied to the rice ﬁeld before transplanting rice, to the rice at transplanting, to the ﬂoodwater ca. 4 weeks after sowing and at panicle initiation or booting (De Datta et al., 1989, 1991; Son and Buresh, 1994). In the study of De Datta et al. (1991), 46– 54% of the urea-N applied immediately after ﬂooding was emitted as NH3 due to the low buffer capacity of ﬂoodwater and a high pH after hydrolysis of the urea. Incorporating the ﬁrst application of urea into the dry soil before ﬂooding may reduce NH3 emission to about 10% compared with the emission of 46 – 54% of urea-N after broadspreading of the urea to the rice ﬁeld after ﬂooding (Fig. 15). This soil had a pH of 4.6 and a CEC of 250 mmol kg21 and because TAN was absorbed on soil colloids as NHþ 4 , the potential of NH3 emission from the incorporated urea was low. Incorporation of urea in ﬂooded soils has little effect on NH3 emission from urea but signiﬁcantly reduces the emission from (NH4)2SO4; probably urea incorporated in ﬂooded soils is not adequately in contact with the soil to reduce emission (Fig. 15). In addition to incorporating urea before ﬂooding rice ﬁelds, NH3 emission from paddy ﬁelds may also be reduced by matching the application to crop demand by split applications of urea to the ﬂoodwater. Urea application to rice may be delayed 14 – 16 days after transplanting in the period of vigorous growth and assimilation of the nitrogen applied (Son and Buresh, 1994). The plant uptake

594

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

of N will thereby reduce NH3 emission and contribute to a higher fertilizer efﬁciency of the applied urea-N. Little NH3 is emitted from urea applied to the rice ﬁeld at panicle initiation or booting, because rice is rapidly assimilating the nitrogen, and shading of the water reduces temperature and wind speed (Freney et al., 1981; De Datta et al., 1989). Furthermore, shading of the water reduces pH increase due to respiration of algae.

D.

INJECTION

OF

ANHYDROUS AMMONIA

The emission of NH3 from injected AA will be related to application rate, depth of injection, knife spacing, and soil buffering capacity (Izaurralde et al., 1990). Penetration depth of NH3 increases with decreasing soil water content (Blue and Eno, 1954; McDowell and Smith, 1958). Thus, from AA injected into a dry soil, the NH3 emission was 20% because NH3 retention capacity was low and a part of the injected NH3 could move through the air-ﬁlled pore space to the soil surface (Sommer and Christensen, 1992). On the other hand, the crevice left after injection into a wet soil was left open and the distance of NH3 penetrated into the soil to the open crevice was short and emission was as high as 50% (Sommer and Christensen, 1992). Hopkins et al. (1963) showed that injection of AA may push water in the soil ahead of the gas and thereby reduce gas movement in water-ﬁlled soils, which will result in high TAN concentrations in soil water near the site of injection. Anhydrous ammonia is usually injected into the soil after winter or a rainy season. The injection will take place when soils are moist, as driving on wet soils is either impossible or will compact the soil. Under these conditions, little NH3 will be emitted from AA injected to depths below 10 cm (Baker et al., 1959; Denmead et al., 1977; Sommer and Christensen, 1992), because the furrow will close and soil water will absorb the NH3. Injection depth of AA where little or no NH3 emission will take place may be predicted using TAN-transport models with input of amount of applied AA, soil pH buffer capacity, diffusion as affected by soil water, and CEC (Izaurralde et al., 1990). The model calculations have shown, for a ﬁne sandy loam and a silt loam soil, that emission is low at 15 cm injection and signiﬁcantly higher (8%) at 5 cm injection depth.

E.

CROP NH 3 EMISSIONS AS AFFECTED APPLICATION

BY

FERTILIZER

Volatile NH 3 losses from crops depend on seasonal variations in climatic conditions affecting the growth and nitrogen economy of the crops (Schjoerring and Mattsson, 2001). In general, losses are expected to increase with the N concentration of the foliage and to the extent that this is controlled by other

AMMONIA EMISSION

595

Ammonia emission, g NH3 m–2 day–1

growth factors than fertilizer N availability; there may be no connection between NH3 emission and fertilizer application. As an example from a ﬁeld study by Schjoerring and Mattsson (2001), lower NH3 emissions were measured from wheat plants fertilized according to optimum N-recommendations as compared to plants applied a reduced amount of N-fertilizer (75% of optimum), because the latter plants had a higher shoot N concentration during part of the growing season. In the same study, the accumulated NH3 loss over a growing season was positively correlated with the above-ground crop N content at anthesis, but not with that at ﬁnal maturity, and there were no indications that NH3 emissions were larger under conditions unfavorable for nitrogen remobilization from vegetative plant parts (low N harvest index). Enhanced NH3 emission from leaves under conditions with excessive N absorption by roots has been observed in several laboratory experiments (see e.g., Mattsson and Schjoerring, 2002). A similar relationship seems also to be valid under ﬁeld conditions; thus, as shown in Fig. 16, application of the acid fertilizers DAP and AS to a grass ley or young wheat plants in April resulted in a peak of NH3 emission between 5 and 12 days after fertilizer application. A similar peak was not observed when the same fertilizers were applied in early spring (March), i.e., before signiﬁcant plant uptake of N, where the NH3 emission continuously declined following high initial loss rates. The extra peak in April is assumed to represent NH3 emission from the seedlings, following rapid

0.5

DAP, April DAP, March AS, April AS, March

0.4 0.3 0.2 0.1 0.0 0

5

10

15

20

Days from application of fertilizers Figure 16 Wind tunnel measurements of NH3 emission from DAP and AS applied to grass (10– 15 cm) on 9 March 1992 and to winter wheat (5 cm) on 1 April 1992. In the March experiment, air temperature was 4.38C, soil temperature 5.78C and wind speed in the wind tunnel 4.2 m s21. In the April experiment, air temperature was 5.88C, soil temperature 7.48C, global radiation 10.9 MJ m22 and wind speed in the wind tunnel 4.2 m s21. In the April experiment, an atypical peak in emission was measured from 5 to 13 days after fertilizer application; this peak is attributed to NH3 emission from the weak seedlings not having capacity to transform NHþ 4 to amide/amines due to low global radiation.

596

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

nitrogen uptake in a period with low global radiation and, thus, a low photosynthetic activity.

F.

AMMONIA EMISSION FROM DECOMPOSING PLANT MATERIAL

Decomposing crop residues may constitute a signiﬁcant source of atmospheric NH3. In measurements by Mannheim et al. (1997), NH3 emissions from decomposing sugar beet leaves, potato tops, and ﬁeld-bean straw ranged from 0.9 to 3.7% of the N content. The highest emissions, reaching from 8.6 up to 12.6 kg N ha21, occurred from sugar beet leaves and potato shoots with high water content, whereas the emissions from ﬁeld-bean straw with high dry matter and N content were relatively low (3.1 kg N ha21 or 0.9% of the N content). The NH3 emission from sugar beet leaves was reduced 81% by plowing and 63% by mulching (Mannheim et al., 1997). A marked inﬂuence of moisture on the NH3 volatilization during decomposition was also observed for ryegrass herbage, where 20 –47% of the herbage N was lost over a 70 day period if kept moist, while volatilization was less than 1% of herbage N during drying (Whitehead et al., 1988). The quantity of NH3 lost increased with herbage N concentration and temperature during decomposition (Whitehead et al., 1988). As an example, herbage containing 3% N on a dry matter basis lost 10% of its N through NH3 volatilization over a period of 28 days, whereas no volatilization was detected from herbage containing 0.9% N. The profound increase in NH3 volatilization with moisture and nitrogen content of plant residues, as well as with temperature, is consistent with the stimulating inﬂuence of these parameters on protein degradation and liberation of NH3/NHþ 4 in senescing leaves. An example of the inﬂuence of nitrogen status is shown in Fig. 17, where bulk leaf tissue NHþ 4 concentrations increased more during senescence of high-N leaves compared with low-N leaves, particularly when plants had been growing at relatively low light intensity, resulting in reduced C/N ratio relative to high-light grown plants (Fig. 17). The ratio between þ [NHþ 4 ] and [H ], which reﬂects the NH3 emission potential from leaf litter, also increased dramatically during senescence (Fig. 17). Signiﬁcant amounts of N may be lost from the plant by senescent leaves falling off and decaying on the ground. Husted et al. (2000b), and Nemitz et al. (2000) showed that in oilseed rape decomposing plant residues on the soil surface made a large contribution to NH3 emission. Some of the NH3 emitted from these leaves might be reabsorbed by leaves still attached to the stem (Nemitz et al., 2000). The soil has also been suggested as an NH3 emission source. However, only a small amount of NH3 emission from the soil was found and it was proposed that the soil acts mostly as a sink for NH3 (Neftel et al., 1998).

AMMONIA EMISSION

597

40

Bulk tissueNH4+,μmolg–1 tissuewater

A High light 0N High light 3N High light 6N Low light 0N Low light 3N Low light 6N

30

20

10

0 70000

B

Bulk tissue[NH4+]/[H+]ratio

60000 50000 40000 30000 20000 10000 0 day 0

day 2

day 4

Days of senescence Figure 17 (A) Bulk tissue extract NHþ 4 concentrations and (B) NH3 emission potential as þ expressed by the [NHþ 4 ]/[H ] ratio in leaves of Lolium perenne grown for 4 weeks in 0, 3, or 6 mM NO2 3 under high and low light conditions on the day of leaf excision and after 2 and 4 days of senescence in darkness. C/N values in low-light grown plants were 52, 12, and 10 for plants grown in 0, 3, or 6 mM NO2 3 , respectively. The corresponding values for high-light grown plants were 63, 24, and 23. Values are means ^ SE for three replicates. (M. Mattsson and J. K. Schjoerring, unpublished results).

598

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

G.

ABSORPTION

BY

CROPS

Plants are capable of absorbing atmospheric NH3 when the NH3 concentration in the atmosphere exceeds that in the substomatal cavities. The absorption responds linearly to ambient NH3 concentration over a very broad range even up to around 500 nmol mol21 (van Hove et al., 1987; Whitehead and Lockyer, 1987). From urea applied to a grass pasture, the NH3 volatilization decreased with increasing canopy density (Hoult and McGarity, 1987; Ping et al., 2000). The crop may have changed the microclimate and the canopy may have absorbed NH3 volatilized from the urea, thereby reducing the ﬂux of NH3 from the soil – plant system as shown in Fig. 13 for urea applied to winter wheat at increasing physiological age. The maize canopy reduced NH3 emission from NH3 dissolved in irrigation water (Denmead et al., 1982), due to a reduction in wind speed, reduction of temperature and uptake of NH3 by the maize canopy. In oilseed rape, signiﬁcant amounts of N may be lost before ﬂowering in dropped leaves (Schjoerring et al., 1995), followed by signiﬁcant NH3 emissions from the decaying leaves (Sutton et al., 2000). However, a signiﬁcant part of the emitted NH3 is absorbed by leaves still attached to the plants, resulting in a very high cycling of NH3 within the canopy, much higher than the net exchange with the atmosphere above the canopy (Nemitz et al., 2000).

VI. MEASUREMENT TECHNIQUES The following section gives an overview of the most widely used techniques for measuring NH3 emission from fertilizer applied to the soil. McGinn and Janzen (1998) and Harper (2003) have comprehensively reviewed techniques for measuring NH3 ﬂuxes to and from the soil.

A.

TRACER TECHNIQUES

Ammonia emission has been estimated by crop response to fertilization or by soil mass balance after application of fertilizer that in some studies have been enriched with N15 (see Denmead et al., 1977; Farquhar et al., 1980; No¨mmik, 1966; Moal et al., 1995; Morvan et al., 1997). Crop response has been shown to be a very unreliable estimator due to variation in total N uptake and uptake of soil-N, other loss pathways such as leaching or nitriﬁcation/denitriﬁcation, and differences in the release patterns of N14 and N15 (Schjoerring et al., 1989). Isotope techniques for estimating N losses have also been questioned, since they seem to overestimate the loss, because NH3 can be both absorbed and released; a plant grown with 15N-enriched fertilizer will tend to lose 15NH3 and gain 14NH3,

AMMONIA EMISSION

599

even if the net NH3 ﬂux is zero (Francis et al., 1997). A signiﬁcant transfer of 15N between labeled and unlabeled plants has been shown to occur via the atmosphere in controlled environment studies (Janzen and Gilbertson, 1994). This will result in an overestimation of losses estimated by 15N analyses. Soil mass balances are unreliable due to loss pathways other than NH3 emission and also because of variations in concentration of TAN or organic N in the soil, i.e., for comparison, recovery of bromide (Br2) added to soil in columns was between 78 and 116% (de Jonge et al., 2003), showing the low precision of this technique. Recently, Vandre´ and Kaupenjohann (1998) have described a method whereby the transfer factor of NH3 from a source to a passive sampler on experimental plots is determined by releasing NH3 at a known rate via a cylinder and tubing on standard comparison plots. The transfer factor is then applied to passive sampler measurements of concentration from manure-treated plots to determine NH3 release rate (i.e., ﬂux) from treated plots. A similar approach has been used by Warland and Thurtell (2000) to infer rates of nitrous oxide evolution from soil. Sherlock et al. (1995) showed for bare soils that the emission of NH3 from applied mineral fertilizers can be calculated as the product of wind speed, NH3 gas in equilibrium with NH3 dissolved in the surface soil layers, and a transfer coefﬁcient. The equilibrium concentration technique (JTI method; Svensson, 1994) is a micrometeorological method suitable for measuring NH3 emissions from small plots. It involves sampling close to the soil surface to measure the driving force for volatilization and the aerodynamic resistance to ﬂux. The method has recently been veriﬁed against the IHF method for applications of urea fertilizer and manure to large plots (Misselbrook and Hansen, 2001). The above-mentioned techniques have not been used extensively. The most commonly used methods have been enclosures and the micrometeorological methods described below.

B.

ENCLOSURES

Enclosures are much used in both ﬁeld and laboratory experiments. The enclosures can be chambers placed on the soil surface with no air ﬂow through the head-space, i.e., static chambers, or they may be dynamic chambers with lids through which the air is exchanged by means of ventilators or pumps. These methods are useful when emission measurements are required over well-deﬁned areas (e.g., small plots) and for comparing treatments under identical environmental conditions (Livingston and Hutchinson, 1995). Chambers are popular because of their portability, versatility, relative simplicity, and high sensitivity. They permit process studies and experiments with many treatments in numbers that could not be contemplated with conventional

600

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

micrometeorological approaches because of the large land areas that the latter require. Moreover, the large increase in gas concentration that occurs in the headspace means that chambers can detect ﬂuxes that are 100 times smaller than can be detected by micrometeorological techniques (Denmead, 1994). 1.

Static Chambers

The ﬂux (F) is calculated from the rate of increase in gas concentration in the enclosure just after the system has been closed. F ¼ ðV=AÞdx=dt

ð24Þ

where V is the volume of the head-space, A is the area of soil surface enclosed by the chamber, x is gas concentration and t is time. The increase in gas concentration is often measured by absorbing the NH3 in an acid solution and storing in an open container or absorbing it on to a ﬁlter. However, the gas concentration gradient from the emitting surface to the air beneath the enclosure decreases as concentration in the air increases. Hence, the size of the enclosure and the measurement period must be carefully selected to avoid negative feedback on the rate of diffusion of the gas. Thus, the measured emission may be lower than when using methods which pass air over the soil and fertilizer (Volk, 1959; Denmead, 1979; McGarry et al., 1987). Such chambers can be used when gas emission rate is controlled by soil processes, as is the case for N2O and CH4, but not in situations where air exchange has a large impact on the emission rate, as may be the case for NH3. 2.

Dynamic Chambers

Closed dynamic chambers have been used in the laboratory and in the ﬁeld. Air is drawn through the chamber, and the rate of NH3 emission is determined from the NH3 enrichment of the air stream, using Eq. (25). In laboratory studies, the surface of the soil in the chambers is generally about 0.01– 0.04 m2 and the chambers are closed continuously (Kissel et al., 1977; Bacon et al., 1986; Sommer and Ersbøll, 1996), whereas chambers with lids that automatically close during short intervals of NH3 emission measurements have been used in ﬁeld studies (Kissel et al., 1977). The wind tunnel system described by Lockyer (1984) is an example of a large dynamic, open chamber covering a surface area of about 1 m2. It employs a fan to draw air over the treated area. Emission (F) from the area is calculated from: F ¼ ðNH3;o 2 NH3;i Þv

ð25Þ

where NH3,o and NH3,i are the NH3 concentrations in the outlet and inlet air, respectively, and v is the volume of air ﬂowing through the tunnel over

AMMONIA EMISSION

601

the sampling period. Average NH3 recovery of between 74 and 90% has been found for wind tunnel systems (Sommer et al., 1991; van der Weerden et al., 1996). It is suggested that the NH3 trapping efﬁciency of wind tunnel systems should be checked on a regular basis to avoid errors in measurement. As discussed already, emissions from the small plots covered by chambers might be higher than those from a ﬁeld to which fertilizers have been applied. Ryden and Lockyer (1985) showed that NH3 emission measured with a wind tunnel adjusted to the wind speed at 10 cm height in the open was 5% higher than the emission measured with the micrometeorological mass balance technique. Usually wind speed in small laboratory chambers is made deliberately high and emissions correspondingly represent maximum losses (Kissel et al., 1977; Bacon et al., 1986).

C.

MICROMETEOROLOGICAL METHODS

The concept of this approach is to measure NH3 emissions from large, open experimental areas, which can be plots or entire ﬁelds. Micrometeorological methods have the great advantage of not being intrusive, and of integrating across heterogeneities in the experimental area. They include mass balance methods, gradient diffusion approaches, eddy correlation, and relaxed eddy accumulation techniques, and methods based on Lagrangian dispersion. 1.

Mass Balance Methods

These are probably the most widely used techniques for measuring NH3 emissions from larger plots manured with mineral fertilizers. This technology does not affect emissions from the plots to which fertilizer has been applied, but it should be remembered that emission from a small plot might be higher than from a large ﬁeld when scaling up emissions from the former. As discussed earlier in the chapter, whether or not plot size is important in this respect depends on the boundary conditions at the surface. Intercalibration studies have shown that different mass balance techniques give similar results (Schjoerring et al., 1992; Sommer et al., 1995; Wood et al., 2000; Sherlock et al., 2002). 2.

Integrated Horizontal Flux (IHF) Methods

This mass balance method equates the loss of NH3 from the surface of a treated plot with the difference between the amount of NH3 carried off the plot by the wind and the amount carried on to it (Denmead et al., 1977; Denmead, 1983; Wilson et al., 1983). It calculates the average surface ﬂux density of NH3 in the

602

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

treated area, Fg, from the difference in the horizontal ﬂuxes of NH3 across downwind and upwind boundaries: Fg ¼

1 ðzp u ðxd 2 xu Þdz X z0

ð26Þ

where X is the fetch (the distance traveled by the wind across the plot), u is horizontal wind speed, and xd and xu are the downwind and upwind atmospheric NH3 concentrations. The integration limit zp is the height at which the NH3 concentration is at background level. It should be noted that the integral in Eq. (26) is in terms of instantaneous values of u and the NH3 concentrations. The integral is usually evaluated with mean wind speeds and mean concentrations. This neglects turbulent terms implicit in the transport equation, Eq. (26), and results in an overestimation of the true ﬂux by 5 –15% depending on the geometry (Raupach and Legg, 1984; Leuning et al., 1985; Wilson and Shum 1992; R.L. Desjardins et al., unpublished, 2003). In most studies, circular plots and making the “downwind” measurements at the plot center are used. The wind will always blow towards the center regardless of wind direction and the fetch will always be the same, viz., the plot radius. Radii from 3.5 to 40 m have been employed (Beauchamp et al., 1978; Gordon et al., 1988). A large simpliﬁcation in technique for circular plots follows from developments by Wilson et al. (1982) and Denmead (1983), showing that for a given surface roughness and plot radius, there exists one particular height of measurement (ZINST) where the horizontal ﬂux density is in a ﬁxed ratio to the vertical ﬂux density, regardless of atmospheric stability. The emission can thus be inferred from measurements of wind speed and atmospheric NH3 concentration (x) at a single height above the ground (Wilson et al., 1983). The ﬂux F is calculated from: Fg ¼

u £ x Z

ð27Þ

where u¯ and x¯ are the mean wind speed and mean NH3 concentration, respectively, measured at ZINST. The term Z is the normalized horizontal ﬂux (u¯x¯/F0), which is given in the form of nomograms for different surface roughnesses and plot radii by Wilson et al. (1982). The method offers considerable savings in labor and equipment, with results comparable to those obtained with the IHF method outlined above. This approach has been made even simpler by the development of passive samplers that measure the horizontal ﬂux directly (Leuning et al., 1985; Sherlock et al., 1995). These require no power, no pumps, no anemometers, no data-loggers. A single sampler mounted at ZINST can return the mean NH3 ﬂux from the treated plot over periods from 1 day to several weeks (for details see Freney et al., 1992).

AMMONIA EMISSION

603

3. Theoretical Proﬁle Shape — Philip’s Solution (PTPS) McInnes et al. (1985) have used Philip’s (1959) analysis for a constant ﬂux boundary condition to quantify the relationship between the surface ﬂux and horizontal ﬂuxes at any height. Circular geometry is used. Philip’s analysis predicts the concentration C1 generated at a particular height by a unit surface ﬂux density F1 for a given wind speed. The true emission (F) is calculated from the concentration C measured at that height using the equation: F¼

C £ F1 C1

ð28Þ

Calculation of the ﬂuxes at various heights requires measurements of air temperature and wind speed at two heights, as well as soil temperature and atmospheric stability. An advantage of the method is that it allows ﬂuxes to be calculated from measurements at a height where the error to magnitude of measurement is smallest, but it is based on a constant ﬂux surface condition, which may not always be appropriate. 4.

Perimeter Proﬁle Method

The perimeter proﬁle method is another mass balance method that employs four masts placed perpendicular to each other around the perimeter of an experimental area (Schjoerring et al., 1992). Arrays of ﬂux samplers (Ferm, 1991) are mounted in pairs on masts around the boundary of a circular experimental area. The horizontal ﬂuxes of the inward and outward pointing tubes are determined separately for each of several heights on each mast. The vertical ﬂux of NH3 is then determined by stepwise summation of the difference between the inward and outward facing horizontal ﬂuxes. This technique is laborious but has the advantage that there is no demand for a homogeneous surface around the plot to which fertilizer is applied. Denmead et al. (1998) describe a somewhat similar technique in which air is sampled at several heights along the full length of each boundary. It is designed particularly for situations in which there are scattered point sources, such as grazed pastures where NH3 is emitted from scattered dung and urine patches.

D. GRADIENT DIFFUSION METHODS The vertical transport of gases in the surface layer is described by: Fg ¼ 2Kg ›x=›z

ð29Þ

where Fg is ﬂux of NH3, x¯ is its mean concentration over a sampling period long enough to encompass all the signiﬁcant transporting eddies, and Kg is an eddy or

604

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

turbulent diffusivity for gases. Kg is determined by wind speed, height, aerodynamic roughness and atmospheric stability. Equation (29) leads to practical methods for calculating gas ﬂuxes from knowledge of the ﬂux of a tracer gas above the surface and the corresponding gradients or differences in mean atmospheric concentrations of the tracer and the gas. Aerodynamic methods use momentum as the tracer ﬂux and horizontal wind speed as the tracer “concentration”. Energy balance methods use available energy as the tracer ﬂux and a linear function of temperature and humidity as the “concentration”. Both methods are appropriate for large surface sources with fetches of hundreds of meters, and have been widely used for measuring NH3 ﬂux to and from crops.

1.

Aerodynamic Methods

The basic equations are given in Appendix. If measurements are made at only two heights, the gas ﬂux is calculated from: Fg ¼

k2 ðu2 2 u 1 Þðx1 2 x2 Þ {ln½ðz2 2 dÞ=ðz1 2 dÞ 2 ½c1 ðz2 2 dÞ 2 c1 ðz1 2 dÞ}{ln½ðz2 2 dÞ=ðz1 2 dÞ 2 ½c2 ðz2 2 dÞ 2 c2 ðz1 2 dÞ}

ð30Þ

In Eq. (30), the subscripts denote the two measuring heights, the overbars denote means over a suitable measuring period such as 20 or 30 min, and c1 and c2 are corrections for stability effects, given explicitly in Appendix. Aerodynamic methods have been much used for measuring NH3 ﬂuxes from crops, soils, and water bodies (e.g., Denmead et al., 1978; Denmead, 1983; Harper et al., 1983, 2000; Sutton et al., 1993a,b, 2000; Genermont et al., 1998; Grifﬁth and Galle, 2000).

2.

Energy Balance Methods

These methods are also discussed in Appendix. For concentrations measured at two heights, the gas ﬂux is calculated from: Fg ¼

ðRn 2 G0 2 SÞðx1 2 x2 Þ rcp ½ðT 1 2 T 2 Þ þ ðe1 2 e 2 Þ=g

ð31Þ

where Rn is the net radiation receipt at the surface (incoming short- and longwave radiation minus reﬂected and re-emitted radiation), G0 is the ﬂux density of heat into the soil at its surface, S is a storage term, T and e are air temperature and vapor pressure, and g is the psychrometric constant. Advantages of the method over the aerodynamic approach are that it does not require calculation of z0 or d and is applicable in all stability conditions. As well, the basic measurements can provide a measurement of evaporation rate, which

AMMONIA EMISSION

605

often has an important inﬂuence on NH3 production and emission. Its disadvantage is that it can give erroneous ﬂuxes at night because both Rn and G are then small and difﬁcult to measure, and net radiometers may not function effectively since their domes can become covered with dew. The same problems can arise in rainy periods. Aerodynamic methods are usually more reliable at night. Like the aerodynamic approach, the energy balance method is much used for measuring NH3 ﬂuxes on a ﬁeld basis. Some example applications are provided by Denmead et al. (1982, 2003) and Freney et al. (1992). It should be noted that sensors mounted at different heights above the surface have different fetches, so that problems arise in the use of gradient diffusion methods if the surface ﬂuxes are spatially heterogeneous.

E.

EDDY CORRELATION

This is the preferred micrometeorological method for measuring scalar ﬂuxes, because it is a direct measurement requiring no simplifying assumptions (about similarity between the Ks) and no stability corrections, and it gives the vertical ﬂux at the point of measurement. There are no problems with different footprints for different measurement levels as there are for gradient-diffusion methods. The basic method requires simultaneous measurements of the vertical wind speed w (usually by a sonic anemometer) and the gas concentration x at a sampling frequency rapid enough to catch all the signiﬁcant eddies, say 10 Hz. The instantaneous vertical gas ﬂux density ¼ wx, and the mean ﬂux density is the average of wx over a period long enough to encompass all the effective transporting eddy sizes. There are problems in using the method with some trace gases because it requires fast gas measurement (10 Hz sampling), which is not yet possible for many trace gases, and density effects caused by simultaneous ﬂuxes of sensible and latent heat may necessitate large corrections to the apparent ﬂux (Webb et al., 1980). Recently, however, a fast-response gas analyzer employing a tunable diode laser has been developed for measuring atmospheric NH3 concentrations. Further, calculations by Wesely et al. (1989) indicate that for NH3, the corrections for density effects will usually be negligible, so that eddy correlation measurements of NH3 ﬂuxes are likely to become more common.

F.

RELAXED EDDY ACCUMULATION SAMPLING

OR

CONDITIONAL

The relaxed eddy accumulation method retains many of the advantages of eddy correlation inasmuch as it is a point measurement and requires no stability corrections. In addition, it does away with the need for rapid gas measurement. It still requires rapid measurement of the vertical wind speed w, but substitutes fast

606

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

solenoid valves for a fast-response gas analyzer. The anemometer controls a simple valving system. Air is sampled at a point at a constant rate and is diverted via the valves into “up” and “down” bins depending on the direction of the vertical wind. The contents of the bins are then analyzed for the mean gas concentrations of the ascending and descending eddies, but the analyzer can be of slow response and does not have to be mounted in the ﬁeld, or denuders may be used to trap the NH3 in the air streams (Zhu et al. 2000). The ﬂux of NH3 is calculated from (Businger and Oncley, 1990): þ þ 2 F ¼ bsððNHþ 3 Þ 2 ðNH3 Þ Þ

ð32Þ

where b is a coefﬁcient close to 0.6, sw is the standard deviation of w, and þ 2 and (NHþ are the respective average NH3 concentrations of the (NHþ 3) 3) ascending and descending eddies (Businger and Oncley, 1990).

G.

LAGRANGIAN DISPERSION MODELS

Lagrangian dispersion analyses adopt a coordinate system that travels with the dispersing entity. This is in contrast to the Eulerian analyses that we have described so far in this chapter, except for the trajectory-simulation model of Wilson et al. (1982). Eulerian dispersion uses a ﬁxed coordinate system and considers the passage of scalars at a point ﬁxed in space.

1.

Backward Lagrangian Stochastic Dispersion Model

This model allows predictions of the strength of any surface source from onepoint measurements of wind speed and concentration at any location downstream (Flesch et al., 1995). The plot geometry and the location of the measuring point relative to the plot can be quite arbitrary, but must be known. The model calculates trajectories of air parcels backward in time from the sensor location to the source. It employs computer-simulated gas releases (about 10,000) to relate the surface ﬂux density F0 to mean concentrations developed at speciﬁed heights and distances downwind. Required input information is z0, u¯ and gas concentration xg,z in excess of background at one particular height and distance downwind, plus atmospheric stability. Solutions have the form: F0 ¼ nuz xg;z

ð33Þ

where n is a coefﬁcient calculated for the particular situation by the model. This method differs from the ZINST approach because it caters for any geometry and measurements can be made at any height downwind, but the stability needs to be known. Examples of its use are given by McGinn and Janzen (1998).

AMMONIA EMISSION

2.

607

Inverse Lagrangian Analysis

Soils and plants can be both sources and sinks for NH3, so that measurements of the net NH3 exchange by a plant community will usually be insufﬁcient for full understanding of ecosystem functioning. Additional information will be needed on the strengths and locations of the canopy sources and sinks. Recent research, notably by Raupach (1989a,b,c), has led to the development of a micrometeorological tool known as Inverse Lagrangian Analysis that allows the identiﬁcation of the sites of gas exchange in plant canopies in a non-disturbing, continuous way from relatively simple observations of concentrations and turbulence parameters within and above plant canopies. It is beyond the scope of the present review to examine the analysis in any detail, but examples of its application to NH3 exchange in crops of corn, oilseed rape, and sugarcane can be found in Harper et al. (2000), Nemitz et al. (2000), and Denmead et al. (2003).

VII. CONCLUSIONS AND PERSPECTIVES The emission of NH3 from inorganic N fertilizers varies considerably, depending on soil properties, fertilizer type and atmospheric conditions. Particularly the pH and pH buffer capacity of the soil and dissolving fertilizer salts may have a large impact on the NH3 emission. A simple classiﬁcation tool that can be useful in simple decision support systems may be to classify soils as acid soils, calcareous soils, neutral soils, etc. and fertilizers as, e.g., alkaline, acid fertilizers, and fertilizers at risk of causing precipitates with Ca. The effect of temperature, wind and rain may interact and these interactions should be considered when estimating NH3 emission potentials. Thus, a high wind speed on days with high radiation may on the one hand reduce soil surface temperature, while on the other hand increase atmospheric NH3 transport, resulting in no net effect on NH3 emission. Temperature may signiﬁcantly affect NH3 emission, but the total amount lost will not always increase, because the emission may be higher for a longer period at low than at high temperatures. Consequently, simple models relating emission to air temperature and wind speed may be erroneous, and more complex models will have to be developed, including calculation of surface temperature as affected by global radiation and wind. The effect of rain on NH3 emission from urea depends on timing and intensity — a light shower immediately after application may increase emission and heavy rain will reduce emission. Ammonia emission directly from the foliage of crops may on some, but not all, occasions be related to N fertilizer application. There are indications that

608

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

seedlings rapidly absorbing N via the roots may emit NH3 because internal N assimilation is too slow relative to N uptake. At later growth stages, the leaves may absorb NH3 emitted from fertilizer on the soil surface beneath the canopy, while during grain ﬁlling NH3 may again be emitted from the foliage as N is remobilised from senescing leaves. Biological activity in the soil associated with nitriﬁcation and N-immobilization may affect fertilizer NH3 emissions via effects on pH and NHþ 4 concentration. In models of ammonia losses from fertilized ﬁelds, both canopy –atmosphere NH3 exchange and nitriﬁcation should be included. Incorporation of fertilizers is among the most efﬁcient techniques for reducing NH3 emission from dry soil. Splitting fertilizer applications so that the canopy more efﬁciently will be able to capture NH3 may also contribute to reduced emissions. Additives may inﬂuence loss of NH3 from applied fertilizers but their efﬁciency may be variable, e.g., depending on rain transporting urea into the soil. Furthermore, additives inhibiting urea hydrolysis are costly and additives reducing pH may either be a safety risk to farmers or may increase the demand for lime. The trend towards no-till crop production will increase the area of land covered with trash on which fertilizers are subsequently applied. Ammonia emission will be high from these ﬁelds and techniques should be developed to reduce NH3 emission without increasing soil tillage. Dynamic modeling of NH3 emission in relation to soil properties is challenging due to the many interacting processes. The physico-chemical and biological processes (pH, hydrolysis, cation adsorption, precipitation, nitriﬁcation etc.) affecting the gradient of ammonium and pH between the fertilizer granules and the soil may vary spatially and be difﬁcult to predict. Thus, mixed models using empirical algorithms and mechanistic submodels may prove useful for the purpose of developing reliable emission-models with an acceptable and realistic need for input data. Atmospheric turbulence equations may predict transport of NH3 from a bare soil surface to the atmosphere, but may fail to describe NH3 emission from fertilizers on the soil beneath a plant canopy, i.e., the models may not precisely take into account how the canopy will change the vertical wind speed proﬁles and provide shade. Moreover, crops may absorb NH3, with the assimilation being related to NH3 concentration. Data are needed for validating models of ammonia emission and for the purpose of establishing relationship between emission and the most important emission factors. In the processing of data, systematic biases due to effect of measurement technique and plot size must be taken into account, i.e., NH3 emission from small plots may be too high compared with emissions measured on a large ﬁeld scale. Therefore, results from small dynamic chamber studies should only be used qualitatively when assessing the effects of different soil, fertilizer or climatic factors on emission patterns. This bias should be corrected when scaling up the results to ﬁeld scale or when developing decision support systems and calculating national inventories.

AMMONIA EMISSION

609

APPENDIX A. DERIVATION OF OPERATING FORMULAE FOR MEASURING GAS FLUXES BY AERODYNAMIC AND ENERGY BALANCE METHODS 1.

ATMOSPHERIC STABILITY

This Appendix is prefaced with a short explanation about atmospheric stability. Stability is concerned with the relative effects of buoyancy (positive or negative) and wind shear on transport and mixing in the atmosphere. Atmospheres are described as being neutral, stable, or unstable depending on buoyancy. In neutral conditions, which occur for short periods around sunrise and sunset, there is essentially no temperature gradient and hence no buoyancy. Parcels of air displaced upwards are at the same temperature as their surroundings and are neither accelerated nor decelerated. Transport is subject only to mechanical turbulence. In stable conditions, which usually occur at night, the temperature gradient is positive (air is cooler closer to the ground) and parcels displaced upwards are heavier than their surroundings, so their motion is decelerated. Transport is inhibited. In unstable conditions, the usual daytime condition, the temperature gradient is negative (air is warmer closer to the ground) and parcels displaced upwards are lighter than their surroundings and so their motion is accelerated. Transport is enhanced. A common index of stability is the Monin – Obukhov length, L, which can be calculated from measurements of the temperature and wind speed gradients or from direct measurements of the heat and momentum ﬂuxes (Paulson, 1970).

2.

AERODYNAMIC METHOD

The methodology has been described in a number of reviews (e.g., Thom, 1972; Monteith and Unsworth, 1990; Denmead, 1994), and only summary results are given here. The general formula for calculating the gas ﬂux from wind speed and concentration measurements is ›u ›x ›u ›x k 2 z2 k2 ›z ›z ›lnz ›lnz ¼ ðA1Þ Fg ¼ wm w g wm wg where the symbols are as used previously and wm and wg are functions to account for the effects of atmospheric stability on momentum and gas transport, respectively, and are given explicitly in the references above. If wind speeds and concentrations have been measured at a number of points in the surface boundary layer, their gradients can be calculated and Eq. (A1) evaluated directly. Usually, however, wind speeds and concentrations

610

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

are measured at only two levels in the boundary layer. Then, Eq. (A1) becomes: Fg ¼

k2 ðu2 2 u 1 Þðx1 2 x2 Þ {ln½ðz2 2dÞ=ðz1 2dÞ2½cm ðz2 2dÞ2 cm ðz1 2dÞ}{ln½ðz2 2dÞ=ðz1 2dÞ2½cg ðz2 2dÞ2 cg ðz1 2dÞ}

ðA2Þ

where cm and cg are integrated forms of the stability functions, wm and wg. In neutral conditions:

cm ¼ cg ¼ 0

ðA3Þ

in stable conditions,

cm ¼ cg ¼ 25ðz2dÞ=L

ðA4Þ

and in unstable conditions,

cm ¼ 2ln½ð1þxÞ=2þln½ð1þx2 Þ=222tan21 xþ p=2

ðA5Þ

and

cg ¼ 2ln½ð1þx2 Þ=2

ðA6Þ

where ðA7Þ

x ¼ ð1216z=LÞ1=4

If the friction velocity up is available from eddy correlation measurements, an alternative aerodynamic formulation is: Fg ¼

ku* ðx1 2 x2 Þ ln½ðz2 2dÞ=ðz1 2dÞ2½cg ðz2 2dÞ2 cg ðz1 2dÞ

3.

ðA8Þ

ENERGY BALANCE METHOD

The energy balance at the surface is Rn ¼ H þ lE þ G0 þ S;

ðA9Þ

where Rn is the net radiation receipt at the surface (incoming short- and longwave radiation less reﬂected and re-emitted radiation), H is the ﬂux density of sensible heat from surface to atmosphere, lE is the ﬂux density of latent heat, l being the latent heat of vaporization of water and E the surface evaporation rate, G0 is the ﬂux density of heat into the soil at its surface, and S is a storage term, e.g., the change in heat stored in water bodies or the air or biomass of a plant community, or the solar energy ﬁxed in photosynthesis. Rearranging Eq. (A9) leads to H þ lE ¼ Rn 2 G0 2 S

ðA10Þ

AMMONIA EMISSION

611

Following Eq. (17), ›z H ¼ 2Kh rcp ›T=

ðA11Þ

lE ¼ 2Kg ðrcp =gÞ›e =›z

ðA12Þ

and Fg ¼ 2Kg ›x=›z

ðA13Þ

where most of the symbols are as deﬁned previously, Kh and Kg are the eddy diffusivities for heat and gases or vapors respectively, e is vapor pressure, and g is the psychrometric constant (< 66 Pa K21). Much micrometeorological research has shown that in most ﬁeld situations, Kh ¼ Kg. On that assumption, Fg ¼

ðRn 2 G0 2 SÞ›x=›z ›z þ ð›e =›zÞ=g rcp ½›T=

ðA14Þ

or, for measurements at two levels, Fg ¼

ðRn 2 G0 2 SÞðx1 2 x2 Þ rcp ½ðT 1 2 T 2 Þ þ ðe1 2 e 2 Þ=g

ðA15Þ

Thus, H, lE and Fg can be determined from measurements of the energy ﬂuxes at the surface and differences in temperature, humidity, and gas concentration between two heights above the surface. All the terms in Eq. (A15) can be measured directly with appropriate instrumentation. Advantages of the method over the aerodynamic approach are that it does not require calculation of z0 and d and is applicable in all stability conditions.

ACKNOWLEDGMENT Hydro Agri supported this review ﬁnancially.

REFERENCES Al-Kanani, T., MacKenzie, A. F., and Barthakur, N. N. (1991). Soil water and ammonia volatilization relationships with surface-applied nitrogen fertilizer solutions. Soil Sci. Soc. Am. J. 55, 1761–1766. Asman, W. A. H., and van Jaarsveld, H. A. (1991). A variable resolution transport model applied for NHx in Europe. Atmos Environ. 26A, 445 –464. Asman, W. A. H., Sutton, M. A., and Schjørring, J. K. (1998). Ammonia: emission, atmospheric transport and deposition. New Phytol. 139, 27 –48. Atwell, B. J., Fillery, I. R. P., McInnes, K. J., and Smucker, A. J. M. (2002). The fate of carbon and fertilizer nitrogen when dryland wheat is grown in monoliths of duplex soil. Plant Soil 241, 259 –269.

612

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

Avnimelech, Y., and Laher, M. (1977). Ammonia volatilization from soils: equilibrium considerations. Soil Sci Soc Am J 41, 1080–1084. Bacon, P. E., and Freney, J. R. (1989). Nitrogen loss from different tillage systems and the effect on cereal grain yield. Fert Res. 20, 59–66. Bacon, P. E., Hoult, E. H., and McGarity, J. W. (1986). Ammonia volatilization from fertilizers applied to irrigated wheat soils. Fert. Res. 10, 27–42. Baker, J. H., Peech, M., and Musgrave, R. B. (1959). Determination of application losses of anhydrous ammonia. Agron. J. 51, 361 –362. Beauchamp, E. G., Kidd, G. E., and Thurtell, G. (1978). Ammonia volatilisation from sewage sludge applied in the ﬁeld. J. Environ. Qual. 7, 141–146. Black, A. S., Sherlock, R. R., Smith, N. P., Cameron, K. C., and Goh, K. M. (1985). Effects of form of nitrogen, season, and urea application rate on ammonia volatilization from pastures. N. Z. J. Agric. Res. 28, 469– 474. Black, A. S., Sherlock, R. R., and Smith, N. P. (1987a). Effect of urea granule size on ammonia volatilization from surface-applied urea. Fert. Res. 11, 87 –96. Black, A. S., Sherlock, R. R., and Smith, N. P. (1987b). Effect of timing of simulated rainfall on ammonia volatilization from urea, applied to soil of varying moisture content. J. Soil Sci. 38, 679–687. Blaise, D., Tyagi, P. C., and Khola, O. P. S. (1996). Ammonia volatilisation from urea as affected by the addition of iron pyrites and method of application. Nutr. Cycl. Agroecosyst. 46, 97–101. Blue, W. G., and Eno, C. F. (1954). Distribution and retention of anhydrous ammonia in sandy soils. Soil Sci. Soc. Am. Proc. 18, 420–424. Bouwman, A. F., Lee, D. S., Asman, W. A. H., Dentener, F. J., Van Der Hoek, K. W., and Olivier, J. G. J. (1997). A global high-resolution emission inventory for ammonia. Global Biogeochem. Cycl. 11, 561 –587. Bouwman, A. F., Boumans, L. J. M., and Batjes, N. H. (2002). Estimation of global NH3 volatilization loss from synthetic fertilizers and animal manure applied to arable lands and grasslands. Global Biogeochem. Cycl. 16(2) (DOI 10.1029/2000GB001389). Bouwmeester, J. B., and Vlek, P. L. G. (1981). Wind-tunnel simulation and assessment of ammonia volatilization from ponded water. Agron. J. 73, 546– 552. Bouwmeester, R. J. B., Vlek, P. L. G., and Stumpe, J. M. (1985). Effect of environmental factors on ammonia volatilization from a urea-fertilized soil. Soil Sci. Soc. Am. J. 49, 376–381. Bowmer, K. H., and Muirhead, W. A. (1987). Inhibition of algal photosynthesis to control pH and reduce ammonia volatilization from rice ﬂoodwater. Fert. Res. 13, 13 –29. Bremner, J. M. (1995). Recent research on problems in the use of urea as a nitrogen fertilizer. Fert. Res. 42, 321 –329. Bremner, J. M., and Douglas, L. A. (1971). Decomposition of urea phosphate in soils. Soil Sci. Soc. Am. J. 55, 1761–1766. Brouquisse, R., Evrard, A., Rolin, D., Raymond, P., and Roby, C. (2001). Regulation of protein degradation and protease expression by mannose in maize root tips: Pi sequestration by mannose may hinder the study of its signaling properties. Plant Physiol. 125, 1485–1498. Brutsaert, W. (1982). Evaporation into the Atmosphere. Reidel, Dordrecht, The Netherlands. Buijsman, E., Maas, J. F., and Asman, W. A. H. (1987). Anthropogenic NH3 emission in Europe. Atmos. Environ. 21, 1009–1022. Businger, J. A., and Oncley, S. P. (1990). Flux measurement with conditional sampling. J. Atmos. Ocean Technol. 7, 349–352. Bussink, D. W., Huijsmans, J. F. M., and Ketelaars, J. J. M. H. (1994). Ammonia volatilisation from nitric-acid-treated cattle slurry surface applied to grassland. Neth. J. Agric. Sci. 42, 293–309. Byrnes, B. H., and Freney, J. R. (1995). Recent developments on the use of urease inhibitors in the tropics. Fert. Res. 42, 251–259.

AMMONIA EMISSION

613

Christianson, C. B., Carmona, G., Klein, M. O., and Howard, R. G. (1995). Impact on ammonia volatilisation losses of mixing KCl of high pH with urea. Fert. Res. 40, 89– 92. Chung, J.-B., and Zasoski, R. J. (1994). Ammonium–potassium and ammonium–calcium exchange equilibria in bulk and rhizosphere soil. Soil Sci. Soc. Am. J. 58, 1368– 1375. De Datta, S. K., Trevitt, A. C. F., Freney, J. R., Obcemea, W. N., Real, J. G., and Simpson, J. R. (1989). Measuring nitrogen losses from lowland rice using bulk aerodynamic and nitrogen-15 balance methods. Soil Sci. Soc. Am. J. 53, 1275– 1281. De Datta, S. K., Buresh, R. J., Sampson, M. I., Obcemea, W. N., and Real, J. G. (1991). Direct measurement of ammonia and denitriﬁcation ﬂuxes from urea applied to rice. Soil Sci. Soc. Am. J. 55, 543–548. de Jonge, L. W., Sommer, S. G., and Jacobsen, O. H. (2003). Inﬁltration of and ammonia volatilization from pig and cattle slurry applied to harrowed and stubble soils. J. Environ. Qual. (In preparation). Denmead, O. T. (1979). Chamber systems for measuring nitrous oxide emission from soils in the ﬁeld. Soil Sci. Soc. Am. J. 43, 89–95. Denmead, O. T. (1983). Micrometeorological methods for measuring gaseous losses of nitrogen in the ﬁeld. In “Gaseous Loss of Nitrogen from Plant–Soil Systems” (J. R. Freney and J. R. Simpson, Eds.), pp. 133– 158, Martinus Nijhoff/Dr W Junk, The Hague. Denmead, O. T. (1994). Measuring ﬂuxes of CH4 and N2O between agricultural systems and the atmosphere. In “CH4 and N2O: Global Emissions and Controls from Rice Fields and other Agricultural and Industrial Sources” (K. Minami, A. Mosier, and R. Sass, Eds.), pp. 209–234. National Institute of Agro-Environmental Sciences, Tsukuba, Japan. Denmead, O. T., and Freney, J. R. (1992). Transfer coefﬁcients for water–air exchange of ammonia, carbon dioxide and methane. Ecol. Bull. 42, 31–41. Denmead, O. T., Simpson, J. R., and Freney, J. R. (1977). A direct ﬁeld measurement of ammonia emission after injection of anhydrous ammonia. Soil Sci. Soc. Am. J. 41, 1001–1004. Denmead, O. T., Freney, J. R., and Simpson, J. R. (1982). Dynamics of ammonia volatilisation during furrow irrigation of maize. Soil Sci. Soc. Am. J. 46, 149–155. Denmead, O. T., Harper, L. A., Freney, J. R., Grifﬁth, D. W. T., Leuning, R., and Sharpe, R. R. (1998). A mass balance method for non-intrusive measurements of surface–air gas exchange. Atmos. Environ. 32, 357–369. Denmead, O. T., Raupach, M. R., Leuning, R., Dunin, F. X., and Freney, J. R. (2003). Inverse Lagrangian analysis of heat, vapor and gas exchange in plant canopies. In “Micrometeorological Measurements in Agricultural Systems” (E. J. Sadler, J. M. Ham, B. Tanner, J. M. Baker, and J. L. Hatﬁeld, Eds.), American Society of Agronomy, Madison, (in press). Du Preez, C. C., and Burger, R. du T. (1988). Ammonia losses from ammonium-containing and — forming fertilizers after surface application at different rates on alkaline soils. Fert. Res. 15, 71–78. ECETOC. (1994). Ammonia Emissions to Air in Western Europe, Technical Report No. 62. European Centre for Ecotoxicology and Toxicology of Chemicals, Brussels, Belgium. EEA. (1999). Overview of National Programmes to Reduce Greenhouse Gas Emissions. European Environmental Agency, Final version, April 1999. EFMA. (1997). The Fertilizer Industry of the European Union: The Issues of Today, the Outlook for Tomorrow. European Fertilizer Manufacturers Association, Brussels, Belgium. Evangelou, V. P. (1990). Regulation mechanisms and ﬁeld implications of ammonia bonding with various crystalline salts. Soil Sci. Soc. Am. J. 54, 394 –398. Fan, M. X., MacKenzie, A. F., and Blenkhorn, H. D. (1996). A modiﬁed urea based NP fertilizer: Urea — TSP–MAP combinations. Fert. Res. 45, 217–220. Farquhar, G. D., Firth, P. M., Wetselaar, R., and Weir, B. (1980). On the gaseous exchange of ammonia between leaves and the environment: determination of the ammonia compensation point. Plant Physiol. 66, 710–714.

614

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

Fenn, L. B., and Hossner, L. R. (1985). Ammonia volatilization from ammonium or ammoniumforming nitrogen fertilizers. Adv. Soil Sci. 1, 124 –169. Fenn, L. B., and Kissel, D. E. (1973). Ammonia volatilization from surface applications of ammonium compounds on calcareous soils. I. General theory. Soil Sci. Soc. Am. Proc. 37, 855–859. Fenn, L. B., and Kissel, D. E. (1974). Ammonia volatilization from surface applications of ammonium compounds on calcareous soils. I. Effects of temperature and rate of ammonium nitrogen application. Soil Sci. Soc. Am. Proc. 38, 606 –610. Fenn, L. B., and Kissel, D. E. (1976). The inﬂuence of cation exchange capacity and depth of incorporation on ammonia volatilization from ammonium compounds applied to calcareous soils. Soil Sci. Soc. Am. Proc. 40, 394–398. Fenn, L. B., and Miyamoto, S. (1981). Ammonia loss and associated reactions of urea in calcareous soils. J. Soil Sci. Soc. Am. 45, 537–540. Fenn, L. B., and Richards, J. (1986). Ammonia loss from surface applied urea-acid products. Fert. Res. 9, 265– 275. Fenn, L. B., Taylor, R. M., and Matocha, J. E. (1981). Ammonia losses from surface applied nitrogen fertilizer as controlled by soluble calcium and magnesium: general theory. Soil Sci. Soc. Am. J. 45, 777 –781. Fenn, L. B., Matocha, J. E., and Wu, E. (1982). Substitution of ammonium and potassium for added in reduction of ammonia loss from surface-applied urea. Soil Sci. Soc. Am. J. 46, 771–776. Fenn, L. B., Malstrom, H. L., and Wu, E. (1987). Ammonia losses from surface-applied urea as related to urea application rates, plant residue and calcium chloride addition. Fert. Res. 12, 219–227. Ferguson, R. B., Kissel, D. E., Koelliker, J. K., and Basel, W. (1984). Ammonia volatilization from surface applied urea: effect of hydrogen ion buffering capacity. Soil Sci. Soc. Am. J. 48, 578– 582. Ferm, M. (1991). A sensitive diffusional sampler. Report B-1020, Swedish Environmental Research Institute, Gothenburg. Fillery, I. R. P., and De Datta, S. K. (1986). Ammonia volatilization from nitrogen-sources applied to rice ﬁelds. 1. Methodology, ammonia ﬂuxes, and N-15 loss. Soil Sci. Soc. Am. J. 50, 80–86. Fillery, I. R. P., Simpson, J. R., and De Datta, S. K. (1984). Inﬂuence of ﬁeld environment and fertilizer management of ammonia loss from ﬂooded rice. Soil Sci. Soc. Am. J. 48, 915–920. Fillery, I. R. P., Roger, P. A., and De Datta, S. K. (1986). Ammonia volatilization from nitrogen sources applied to rice ﬁelds. II. Floodwater properties and submerged photosynthetic biomass. Soil Sci. Soc. Am. J. 50, 86 –91. Finnemann, J., and Schjoerring, J. K. (1999). Translocation of NHþ 4 in oilseed rape plants in relation to glutamine synthetase isogene expression and activity. Physiol. Plant 105, 469– 477. Fleisher, Z., Kenig, A., Ravina, I., and Hagin, J. (1987). Model of ammonia volatilization from calcareous soils. Plant Soil 103, 205–212. Flesch, T., Wilson, J. D., and Yee, E. (1995). Backward time Lagrangian stochastic dispersion models, and their application to estimate gaseous emissions. J. Appl. Meteorol. 34, 1320– 1332. Flowers, T. H., and O’Callaghan, J. R. (1983). Nitriﬁcation in soils incubated with pig slurry or ammonium sulphate. Soil Biol. Biochem. 15, 337 –342. Francis, D. D., Schepers, J. S., and Vigil, M. F. (1993). Post-anthesis nitrogen loss from corn. Agron. J. 85, 659 –663. Francis, D. D., Schepers, J. S., and Sims, A. L. (1997). Ammonia exchange from corn foliage during reproductive growth. Agron. J. 89, 941–946. Freney, J. R., Denmead, O. T., Watanabe, I., and Craswell, E. T. (1981). Ammonia and nitrous oxide losses following applications of ammonium sulphate to ﬂooded rice. Aust. J. Agric. Res. 32, 37–45. Freney, J. R., Simpson, J. R., and Denmead, O. T. (1983). Volatilization of ammonia. In “Gaseous Loss of Nitrogen from Plant –Soil Systems” (J. R. Freney and J. R. Simpson, Eds.), pp. 1 –32, Martinus Nijhoff/Dr. W. Junk, The Hague.

AMMONIA EMISSION

615

Freney, J. R., Denmead, O. T., Sod, A. W., Safﬁgna, P. G., Chapman, L. S., Ham, G. J., Hurney, A. P., and Stewart, R. L. (1992). Factors controlling ammonia loss from trash covered sugarcane ﬁelds fertilized with urea. Fert. Res. 31, 341–349. Genermont, S., and Cellier, P. (1997). A mechanistic model for estimating ammonia volatilisation from slurry applied to bare soil. Agric. Forest. Meteorol. 88, 145–167. Genermont, S., Cellier, P., Flura, D., Morvan, T., and Laville, P. (1998). Measuring ammonia ﬂuxes after slurry spreading under actual ﬁeld conditions. Atmos. Environ. 32, 279–284. Gilmour, J. T. (1984). The effects of soil properties on nitriﬁcation and nitriﬁcation inhibition. Soil Sci. Soc. Am. J. 48, 1262–1266. Goos, R. J., and Cruz, A. D. (1999). Effect of ammonium sulfate pretreatment on ammonia volatilization after urea fertilization. Commun. Soil Sci. Plant Anal. 30, 1325–1336. Gordon, R., Leclerc, M. Y., Schuepp, P., and Brunke, R. (1988). Field estimates of ammonia volatilisation from swine manure by a simple micrometeorological technique. Can. J. Soil Sci. 68, 369 –380. Grant, R. F. (1994). Simulation of ecological controls on nitriﬁcation. Soil Biol. Biochem. 26, 305 –315. Grifﬁth, D. W. T., and Galle, B. (2000). Flux measurements of NH3, N2O and CO2 using dual beam FTIR spectroscopy and the ﬂux-gradient technique. Atmos. Environ. 34, 1087–1098. Hanstein, S., Mattsson, M., Jaeger, H.-J., and Schjoerring, J. K. (1999). Uptake and utilization of atmospheric ammonia in three native Poaceae species: Leaf conductances, composition of apoplastic solution and interactions with root nitrogen supply. New Phytol. 141, 71–83. Harper, L. A. (2003). Ammonia: measurement issues. In “Micrometeorological Measurements in Agricultural Systems” (E. J. Sadler, J. M. Ham, B. Tanner, J. M. Baker, and J. L. Hatﬁeld, Eds.), American Society of Agronomy, Madison, (in press). Harper, L. A., Catchpoole, V. R., Davis, R., and Weir, K. L. (1983). Ammonia volatilization: soil, plant and microclimate effects on diurnal and seasonal ﬂuctuations. Agron. J. 75, 212–218. Harper, L. A., Sharpe, R. R., Langdale, G. W., and Giddens, J. E. (1987). Nitrogen cycling in a wheat crop: soil plant and aerial nitrogen transport. Agron. J. 79, 965–973. Harper, L. A., Bussink, D. W., van der Meer, H. G., and Corre, W. J. (1996). Ammonia transport in a temperate grassland. 1. Seasonal transport in relation to soil fertility and crop management. Agron. J. 88, 614 –621. Harper, L. A., Denmead, O. T., and Sharpe, R. R. (2000). Identifying sources and sinks of scalars in a corn canopy with inverse Lagrangian dispersion analysis. II. Ammonia. Agric. For. Meteorol. 104, 75–83. Harrison, R., and Webb, J. (2001). A review of the effect of N fertilizer type on gaseous emission. Adv. Agron. 73, 65 –108. Haynes, R. J., and Williams, P. H. (1992). Changes in soil solution composition and pH in urineaffected areas of pasture. J. Soil Sci. 43, 323–334. He, Z. L., Alva, A. K., Calvert, D. V., and Banks, D. J. (1999). Ammonia volatilization from different fertilizer sources and effects of temperature and soil pH. Soil Sci. 164, 751–758. Herrmann, B., Mattsson, M., Fuhrer, J., and Schjoerring, J. K. (2002). Leaf –atmosphere NH3 exchange of white clover in relation to mineral N nutrition and symbiotic N2 ﬁxation. J. Exp. Bot. 53, 139–146. Hill, P. W., Raven, J. A., Loubet, B., Fowler, D., and Sutton, M. A. (2001). Comparison of gas exchange and bioassay determinations of the ammonia compensation point in Luzula sylvatica (Huds.) Gaud. Plant. Physiol. 125, 476– 487. Hirel, B., and Lea, P. J. (2001). Ammonia assimilation. In “Plant Nitrogen” (P. J. Lea and J.-F. Morot-Gaudry, Eds.), pp. 79 –99. Springer-Verlag, Berlin. Hopkins, D. F., Wells, G. L., and Butler, B. J. (1963). High-pressure injection of anhydrous ammonia, Trans. ASAE, Paper no. 62–645 presented at the winter meeting of the American Society of Agricultural Engineers in Chicago, December 1962, pp. 343–352.

616

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

Hoult, E. H., and McGarity, J. W. (1987). The inﬂuence of sward mass, defoliation and watering on ammonia volatilization losses from an Italian ryegrass sward topdressed with urea. Fert. Res. 13, 199–207. Husted, S., and Schjoerring, J. K. (1995). Apoplastic pH and ammonium concentration in leaves of Brassica napus L. Plant Physiol. 109, 1453–1460. Husted, S., and Schjoerring, J. K. (1996). Ammonia ﬂuxes between Brassica napus plants and the atmosphere in response to changes in leaf temperature, light intensity and air humidity: interactions þ with stomatal conductance and apoplastic NHþ 4 and H concentrations. Plant Physiol. 112, 67 –74. Husted, S., Mattsson, M., and Schjoerring, J. K. (1996). Ammonia compensation points in two cultivars of Hordeum vulgare L. during vegetative and generative growth. Plant Cell Environ. 19, 1299–1306. Husted, S., Hebbern, C., Mattsson, M., and Schjoerring, J. K. (2000a). A critical experimental evaluation of methods used for analysing ammonium in plant tissue, xylem sap and apoplastic ﬂuid. Physiol. Plant. 109, 167 –179. Husted, S., Schjoerring, J. K., Nielsen, K. H., Nemitz, E., and Sutton, M. A. (2000b). Stomatal compensation points for ammonia in oilseed rape plants under ﬁeld conditions. Agric. For. Meteorol. 105, 371–383. Husted, S., Mattsson, M., Mo¨llers, C., Wallbraun, M., and Schjoerring, J. K. (2002). Photorespiratory NHþ 4 production in leaves of wildtype and GS2-antisense oilseed rape. Plant Physiol. 130, 989–998. Hutchings, N. J., Sommer, S. G., Andersen, J. M., and Asman, W. A. H. (2001). A detailed ammonia emission inventory for Denmark. Atmos. Environ. 35, 1959– 1968. IFA Statistics (2002). International Fertilizer Industry Association: http://www.fertilizer.org/ifa/ statistics/STATSIND/tablen.asp. Ismail, K. M., Wheaton, F. W., Douglass, L. W., and Potts, W. (1991). Modeling ammonia volatilization from loamy sand soil treated with liquid urea. Trans. ASAE 34, 756–763. Izaurralde, R. C., Kissel, D. E., and Cabrera, M. L. (1987). Titratable acidity to estimate ammonia retention. Soil Sci. Soc. Am. J. 51, 1050–1054. Izaurralde, R. C., Kissel, D. E., and Cabrera, M. L. (1990). Simulation model of banded ammonia in soils. Soil Sci. Soc. Am. J. 54, 917–922. Janzen, H. H., and Gilbertson, C. (1994). Exchange of N-15 among plants in controlled environment studies. Can J. Soil Sci. 74, 109–110. Kirk, G. J. D., and Nye, P. H. (1991). A model of ammonia volatilization from applied urea. V. The effects of steady-state drainage and evaporation. J. Soil Sci. 42, 103 –113. Kissel, D. E., Brewer, H. L., and Arkin, G. F. (1977). Design and test of a ﬁeld sampler for ammonia volatilization. Soil Sci. Soc. Am. J. 41, 1133–1138. Larsen, S., and Gunary, D. (1962). Ammonia loss from ammoniacal fertilisers applied to calcareous soils. J. Sci. Food Agric. 11, 566 –572. Leegood, R. C., Lea, P. J., Adcock, M. D., and Ha¨usler, R. E. (1995). The regulation and control of photorespiration. J. Exp. Bot. 46, 1397– 1414. Leuning, R., Denmead, O. T., Simpson, J. R., and Freney, J. R. (1984). Processes of ammonia loss from shallow ﬂoodwater. Atmos. Environ. 18, 1583–1592. Leuning, R., Freney, J. R., Denmead, O. T., and Simpson, J. R. (1985). A sampler for measuring atmospheric ammonia ﬂux. Atmos. Environ. 19, 1117–1124. Lightner, J. W., Mengel, D. B., and Rhykerd, C. L. (1990). Ammonia volatilization from nitrogen fertilizer surface applied to orchardgrass sod. Soil Sci. Soc. Am. J. 54, 1478– 1482. Lindau, C. W., Reddy, K. R., Lu, W. F., Khind, C. S., Pardue, J. H., and Patrick, W. H. Jr. (1989). Effects of redox potential on urea hydrolysis and nitriﬁcation in soil suspensions. Soil Sci. 148, 184– 190. Livingston, G. P., and Hutchinson, G. L. (1995). Enclosure-based measurements of trace gas exchange: application and sources of error. In “Biogenic Trace Gases: Measuring Emissions from Soil and Water” (P. A. Matson and R. C. Harriss, Eds.), pp. 14–51. Blackwell Science, Oxford.

AMMONIA EMISSION

617

Lockyer, D. R. (1984). A system for the measurement in the ﬁeld of losses of ammonia through volatilization. J. Sci. Food Agric. 35, 837–848. Loubet, B., Milford, C., Hill, P. W., Tang, Y. S., Cellier, P., and Sutton, M. A. (2002). Seasonal variability of apoplastic NHþ 4 and pH in an intensively managed grassland. Plant Soil 238, 97 –110. Lyster, S., Morgan, M. A., and O’Toole, P. (1980). Ammonia volatilization from soils fertilized with urea and ammonium nitrate. J. Life Sci. R. Dublin Soc. 1, 167–175. Maclean, J. (1997). “Rice Almanac”, 2nd edn. IRRI, Los Ban˜os, Philippines. Malhi, S. S., and McGill, K. S. (1982). Nitriﬁcation in three Alberta soils: effect of temperature, moisture and substrate concentration. Soil Biol. Biochem. 14, 393 –399. Malhi, S. S., Grant, C. A., Johnston, A. M., and Gill, K. S. (2001). Nitrogen fertilization management for no-till cereal production in the Canadian Great Plains: a review. Soil Tillage Res. 60, 101 –122. Mannheim, T., Braschkat, J., and Marschner, H. (1997). Ammonia emissions from senescing plants and during decomposition of crop residues. Z. Pﬂanzenernaehr. Bodenkd. 160, 125 –132. Mary, B., Recous, S., and Robin, D. (1998). A model for calculating nitrogen ﬂuxes in soil using 15N tracing. Soil Biol. Biochem. 30, 1963–1979. Mattsson, M., and Schjoerring, J. K. (1996). Ammonia emission from young barley plants: inﬂuence of N source, light/dark cycles, and inhibition of glutamine synthetase. J. Exp. Bot. 47, 477–484. Mattsson, M., and Schjoerring, J. K. (2002). Dynamic and steady state responses of inorganic nitrogen pools and NH3 exchange in leaves of Lolium perenne and Bromus erectus to changes in root N supply. Plant Physiol. 128, 742 –750. Mattsson, M., Ha¨usler, R. E., Leegood, R. C., Lea, P., and Schjoerring, J. K. (1997). Leaf–atmosphere ammonia exchange in barley mutants with reduced activities of glutamine synthetase. Plant Physiol. 114, 1307–1312. Mattsson, M., Husted, S., and Schjoerring, J. K. (1998). Inﬂuence of nitrogen nutrition and metabolism on ammonia emission from plant leaves. Nutr. Cycl. Agroecosyst. 51, 35– 40. McDowell, L. L., and Smith, G. E. (1958). The retention and reaction of anhydrous ammonia on different soil types. Soil Sci. Soc. Am. Proc. 22, 38–42. McGarry, S. J., O’Toole, P., and Morgan, M. A. (1987). Effects of soil temperature and moisture content on ammonia volatilization from urea-treated pasture and tillage soils. Ir. J. Agric. Res. 26, 173 –182. McGinn, S. M., and Janzen, H. H. (1998). Ammonia sources in agriculture and their measurement. Can. J. Soil Sci. 78, 139 –148. McInnes, K. J., and Fillery, I. R. P. (1989). Modeling and ﬁeld measurements of the effect of nitrogen source on nitriﬁcation. Soil. Sci. Soc. Am. J. 53, 1264–1269. McInnes, K. J., Kissel, D. E., and Kanemasu, E. T. (1985). Estimating ammonia ﬂux: a comparison between the integrated horizontal ﬂux method and theoretical solutions of the diffusion proﬁle. Agron. J. 77, 884 –889. McInnes, K. J., Ferguson, R. B., Kissel, D. S., and Kanemasu, E. T. (1986a). Field measurements of ammonia loss from surface applications of urea solution to bare soil. Agron. J. 78, 192 –196. McInnes, K. J., Ferguson, R. B., Kissel, D. S., and Kanemasu, E. T. (1986b). Ammonia loss from applications of urea-ammonium nitrate solution to straw residue. Soil Sci. Soc. Am. J. 50, 969–974. Mikkelsen, D. S., De Datta, S. K., and Obcemea, W. N. (1978). Ammonia volatilization losses from ﬂooded rice soils. Soil Sci. Soc. Am. J. 42, 752–730. Misselbrook, T. H., and Hansen, M. N. (2001). Field evaluation of the equilibrium concentration technique (JTI method) for measuring ammonia emission from land spread manure or fertiliser. Atmos. Environ. 35, 3761–3768. Misselbrook, T. H., van der Weerden, T. J., Pain, B. F., Jarvis, S. C., Chambers, B. J., Smith, K. A., Phillips, V. R., and Demmers, T. G. M. (2000). Ammonia emission factors for UK agriculture. Atmos. Environ. 34, 871 –880.

618

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

Moal, J.-F., Martinez, J., Guiziou, F., and Coste, C.-M. (1995). Ammonia volatilization following surface-applied pig and cattle slurry in France. J. Agric. Sci. 125, 245–252. van der Molen, J., van Faassen, H. G., Leclerc, M. Y., Vriesema, R., and Chardon, W. J. (1990). Ammonia volatilization from arable land after application of cattle slurry: derivation of a transfer model. Neth. J. Agric. Sci. 38, 239–254. Monteith, J. L., and Unsworth, M. H. (1990). “Principles of Environmental Physics”. Edward Arnold, London. Morvan, T., Leterme, Ph., Arsene, G. G., and Mary, B. (1997). Nitrogen transformations after the spreading of a pig slurry on bare soil and ryegrass using 15N-labelled ammonium. Eur. J. Agron. 7, 181– 188. Murphy, D. V., Recous, S., Stockdale, E. A., Fillery, I. R. P., Jensen, L. S., Hatch, D. J., and Goulding, K. W. T. (2003). Gross nitrogen ﬂuxes in soil: theory, measurement and application of 15N pool dilution techniques. Adv. Agron. 79, 119. Myers, R. J. K. (1975). Temperature effects on ammoniﬁcation and nitriﬁcation in a tropical soil. Soil Biol. Biochem. 7, 83–87. Neftel, A., Blatter, A., Gut, A., Hogger, D., Meixner, F., Ammann, C., and Nathaus, F. J. (1998). NH3 soil and soil surface gas measurements in a triticale wheat ﬁeld. Atmos. Environ. 32, 499–505. Nemitz, E., Sutton, M. A., Gut, A., San Jose, R., Husted, S., and Schjoerring, J. K. (2000). Sources and sinks of ammonia within an oilseed rape canopy. Agr. For. Meteorol. 105, 385 –404. Nemitz, E., Milford, C., and Sutton, M. A. (2001). A two-layer canopy compensation point model for describing bi-directional biosphere–atmosphere exchange of ammonia. Q. J. Roy. Meteor. Soc. 127, 815 –833. Nielsen, K. H., and Schjoerring, J. K. (1998). Regulation of apoplastic ammonium concentration in leaves of oilseed rape. Plant Physiol. 118, 1361–1368. No¨mmik, H. (1966). Use of microplot technique for studying gaseous loss of ammonia from added nitrogen materials under ﬁeld conditions. Acta Agric. Scand. 16, 147–154. Nye, P. H. (1992). Towards the quantitative control of crop production and quality. II. The scientiﬁc basis for guiding fertilizer and management practice, particularly in poorer countries. J. Plant Nutr. 15, 1151–1173. O’Toole, P., McGarry, S. J., and Morgan, M. A. (1985). Ammonia volatilization from urea-treated pasture and tillage soils: effects of soil properties. J. Soil Sci. 36, 613– 620. Ouyang, D. S., Mackenzie, A. F., and Fan, M. X. (1998). Ammonia volatilization from urea amended with triple superphosphate and potassium chloride. Soil Sci. Soc. Am. J. 62, 1443– 1447. Pain, B. F., van der Weerden, T. J., Chambers, B. J., Phillips, V. R., and Jarvis, S. C. (1998). A new inventory for ammonia emissions from UK agriculture. Atmos. Environ. 32, 309–313. Paulson, C. A. (1970). The mathematical representation of wind speed and temperature proﬁles in the unstable atmospheric surface layer. J. Appl. Meteorol. 9, 857–861. Philip, J. R. (1959). The theory of local advection. Int. J. Meteorol. 16, 535–547. Phongpan, S., Freney, J. R., Keerthisinghe, D. G., and Chaiwanakupt, P. (1995). Use of phenylphosphorodiamidate and N-(n-butyl)thiophosphorictriamide to reduce ammonia loss and increase grain-yield following application of urea to ﬂooded rice. Fert. Res 41, 59 –66. Ping, J., Bremer, E., and Janzen, H. H. (2000). Foliar uptake of volatilized ammonia from surfaceapplied urea by spring wheat. Commun. Soil Sci. Plant Anal. 31, 165 –172. Plantaz, M. A. G. H. (1998). “Surface/atmosphere exchange of ammonia over grazed pasture”. Dissertation No. 2468. University of Wageningen, The Netherlands. Prasertsak, P., Freney, J. R., Safﬁgna, P. G., Denmead, O. T., and Prove, B. G. (2001). Fate of urea nitrogen applied to a banana crop in the wet tropics of Queensland. Nutr. Cycl. Agroecosyst. 59, 65–73. Rachhpal-Singh, and Nye, P. H. (1986a). A model of ammonia volatilization from applied urea. I. Development of the model. J. Soil Sci., 37, 1–9.

AMMONIA EMISSION

619

Rachhpal-Singh, and Nye, P. H. (1986b). A model of ammonia volatilization from applied urea. II. Experimental testing. J. Soil Sci. 37, 37, 9–29. Raupach, M. R. (1989a). A practical Lagrangian method for relating scalar concentrations to source distributions in vegetation canopies. Q.J. R. Meteorol. Soc. 115, 609–632. Raupach, M. R. (1989b). Stand overstorey processes. Phil. Trans. R. Soc. Lond. B 324, 175 –190. Raupach, M. R. (1989c). Applying Lagrangian ﬂuid mechanics to infer scalar source distributions from concentration proﬁles in plant canopies. Agric. For. Meteorol. 47, 85 –108. Raupach, M. R., and Legg, B. J. (1984). The uses and limitations of ﬂux–gradient relationships in micrometeorology. Agric. Water Manage. 8, 119–131. Recous, S., Fresneau, C., Faurie, G., and Mary, B. (1988). The fate of labelled 15N urea and ammonium nitrate applied to a winter wheat crop. Plant Soil 112, 205 –214. Recous, S., Machet, J. M., and Mary, B. (1992). The partitioning of fertilizer-N between soil and crop: comparison of ammonium and nitrate applications. Plant Soil 144, 101– 111. Recous, S., Aita, C., Mary, B., and Faurie, B. (1999). In situ changes in gross N transformations in bare soil after addition of straw. Soil Biol. Biochem. 31, 119–133. Reynolds, C. M., and Wolf, D. C. (1987). Effect of soil moisture and air relative humidity on ammonia volatilization from surface-applied urea. Soil Sci. 143, 144–152. Ringbom, A. (1963). “Complexation in Analytical Chemistry”. Interscience Publishers, Chichester. Roelcke, M., Li, S. X., Tian, X. H., Gao, Y. J., and Richter, J. (2002). In situ comparisons of ammonia volatilisation from N fertilizers in Chinese loess soils. Nutr. Cycl. Agroecosyst. 62, 73– 88. Russel, E. W. (1977). “Soil Conditions and Plant Growth”, 10th edn. Longman, London, New York. Ryden, J. C., and Lockyer, D. R. (1985). Evaluation of a system of wind tunnels for ﬁeld studies of ammonia loss from grassland through volatilisation. J. Sci. Food. Agric. 36, 781 –788. Ryden, J. C., Whitehead, D. C., Lockyer, D. R., Thompsen, R. B., Skinner, J. H., and Garwood, E. A. (1987). NH3 emission from grassland and livestock production systems in the UK. Environ. Pollut. 48, 173–184. Schjoerring, J. K., and Mattsson, M. (2001). Quantiﬁcation of ammonia exchange between agricultural cropland and the atmosphere: measurements over two complete growth cycles of wheat, oilseed rape, barley and pea. Plant Soil 228, 105– 115. Schjoerring, J. K., Nielsen, N. E., Jensen, H. E., and Gottschau, A. (1989). Nitrogen losses from ﬁeldgrown spring barley plants as affected by rate of nitrogen application. Plant Soil 116, 167–175. Schjoerring, J. K., Sommer, S. G., and Ferm, M. (1992). A simple passive sampler for measuring ammonia emission in the ﬁeld. Water Air Soil Pollut. 62, 13–24. Schjoerring, J. K., Kyllingsbaek, A., Mortensen, J. V., and Byskov-Nielsen, S. (1993). Field investigations of ammonia exchange between barley plants and the atmosphere. I. Concentration proﬁles and ﬂux densities of ammonia. Plant Cell Environ. 16, 161–167. Schjoerring, J. K., Bock, J. H. G., Gammelvind, L., Jensen, C. R., and Mogensen, V. O. (1995). Nitrogen incorporation and remobilization in different shoot components of ﬁeld-grown winter oil-seed rape as affected by rate of nitrogen application and irrigation. Plant Soil 177, 255–264. Schjoerring, J. K., Husted, S., and Poulsen, M. (1998). Soil– plant– atmosphere NH3 exchange associated with Calluna vulgaris and Deschampsia ﬂexuosa. Atmos. Environ. 32, 507– 512. Schlesinger, W. H., and Hartley, A. E. (1992). A global budget for atmospheric NH3. Biogeochemistry 15, 191–211. Schulze, E.-D., De Vries, W., Hauhs, M., Rose´n, K., Rasmussen, L., Tamm, C.-O., and Nilsson, J. (1989). Critical loads for nitrogen deposition on forest ecosystems. Water Air Soil Pollut. 48, 451 –456. Sherlock, R. R., and Goh, K. M. (1985). Dynamics of ammonia volatilization from simulated urine patches and aqueous urea applied to pasture. I. Field experiments. Fert. Res. 5, 181 –195. Sherlock, R. R., Freney, J. R., Bacon, P. E., and van der Weerden, T. J. (1995). Estimating ammonia volatilization from unsaturated urea fertilized and urine affected soils by an indirect method. Fert. Res. 40, 197–205.

620

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

Sherlock, R. R., Sommer, S. G., Khan, R. Z., Wesley Wood, C., Guertal, E. A., Freney, J. R., Dawson, C. O., and Cameron, K. C. (2002). Ammonia, methane, and nitrous oxide emission from pig slurry applied to a pasture in New Zealand. J. Environ. Qual. 31, 1491–1501. Singh, Y., Malhi, S. S., Nyborg, M., and Beauchamp, E. G. (1994). Large granules, nests or bands: methods of increasing efﬁciency of fall-applied urea for small cereal-grains in North America. Fert. Res. 38, 61–87. Sloan, J. J., and Anderson, W. B. (1995). Calcium chloride and ammonium thiosulfate as ammonia volatilisation inhibitors for urea fertilizers. Commun. Soil Sci. Plant Anal. 26, 2425–2447. Sommer, S. G., and Christensen, B. T. (1992). Ammonia volatilization after injection of anhydrous ammonia into arable soils of different moisture levels. Plant Soil 142, 143– 146. Sommer, S. G., and Ersbøll, S. K. (1996). Effect of air ﬂow rate, lime amendments and chemical soil properties on the volatilization of ammonia from fertilizers applied to sandy soils. Biol. Fert. Soils 21, 53 –60. Sommer, S. G., and Jensen, C. (1994). Ammonia volatilization from urea and ammoniacal fertilizers surface-applied to winter wheat and grassland. Fert. Res. 37, 85– 92. Sommer, S. G., Olesen, J. E., and Christensen, B. T. (1991). Effects of temperature, wind speed and air humidity on ammonia volatilization from surface applied cattle slurry. J. Agric. Sci. 117, 91–100. Sommer, S. G., Mikkelsen, H., and Mellqvist, J. (1995). Evaluation of meteorological techniques for measurements of ammonia loss from pig slurry. Agric. For. Meteorol. 74, 169 –179. Sommer, S. G., Friis, E., Bach, A., and Schjørring, J. K. (1997). Ammonia volatilization from pig slurry applied with trail hoses or broadspread to winter wheat: effects of crop developmental stage, microclimate and leaf ammonia absorption. J. Environ. Qual. 26, 1153–1160. Son, T. T., and Buresh, R. J. (1994). Urea management for lowland transplanted rice as affected by application of farmyard manure. Fert. Res. 39, 199 –203. Sørensen, P. (2001). Short-term nitrogen transformations in soil amended with animal manure. Soil Biol. Biochem. 33, 1211–1216. Stevens, R. J., Laughlin, R. J., and Kilpatrick, D. J. (1989). Soil properties related to the dynamics of ammonia volatilization from urea applied to the surface of acidic soils. Fert. Res. 20, 1–9. Stumm, W., and Morgan, J. J. (1981). Aquatic Chemistry: an Introduction Emphasizing Chemical Equilibria in Natural Waters. John Wiley and Sons, New York. Sutton, M. A., Fowler, D., and Moncrieff, J. B. (1993a). The exchange of atmospheric ammonia with vegetated surfaces. I. Unfertilized vegetation. Q. J. R. Meteorol. Soc. 119, 1023– 1045. Sutton, M. A., Fowler, D., and Moncrieff, J. B. (1993b). The exchange of atmospheric ammonia with vegetated surfaces. II. Fertilized vegetation. Q. J. R. Meteorol. Soc. 119, 1047–1070. Sutton, M. A., Burkhardt, J. K., Guerin, D., Nemitz, E., and Fowler, D. (1998). Development of resistance models to describe measurements of bi-directional ammonia surface-atmosphere exchange. Atmos. Environ. 32, 473–480. Sutton, M. A., Nemitz, E., Milford, C., Fowler, D., Moreno, J., San Jose, R., Wyers, G. P., Otjes, R. P., Harrison, R., Husted, S., and Schjoerring, J. K. (2000). Micrometeorological measurements of net ammonia ﬂuxes over oilseed rape during two vegetation periods. Agr. Forest Meteorol. 105, 351–369. Svensson, L. (1994). A new dynamic chamber technique for measuring ammonia emissions from land-spread manure and fertilizers. Acta Agric. Scand. B 44, 35–46. Tabatabai, M. A., and Bremner, J. M. (1972). Assay of urease activity in soils. Soil Biol. Biochem. 4, 479–487. Thom, A. S. (1972). Momentum, mass and heat exchange of vegetation. Q. J. R. Meteorol. Soc. 98, 124–134. Thomas, R. J., Logan, K. A. B., Ironside, A. D., and Bolton, G. R. (1988). Transformation and fate of sheep urine-N applied to an upland U.K. pasture at different times during the growing season. Plant Soil 107, 173–181.

AMMONIA EMISSION

621

Touchton, J. T., and Hargrove, W. L. (1982). Nitrogen-source and methods of application for notillage corn production. Agron. J. 74, 823–826. Van Hove, L. W. A., Adema, E. H., Vredenberg, W. J., and Pieters, G. A. (1989). A study of the adsorption of NH3 and SO2 on leaf surfaces. Atmos. Environ. 23, 1479–1486. Van Hove, L. W. A., Heeres, P., and Bossen, M. E. (2002). The annual variation in stomatal ammonia compensation point of rye grass (Lolium perenne L.) leaves in an intensively managed grassland. Atmos. Environ. 36, 2965–2977. Van Hove, L. W. A., Koops, A. J., Adema, E. H., Vredenberg, W. J., and Pieters, G. A. (1987). Analysis of the uptake of atmospheric ammonia by leaves of Phaseolus vulgaris L. Atmos. Environ. 21, 1759–1763. Vandre´, R., and Kaupenjohann, M. (1998). In situ measurement of ammonia emissions from organic fertilizers in plot experiments. Soil Sci. Soci. Am. J. 62, 467–473. Velthof, G. L., Oenema, O., Postmus, J., and Prins, W. H. (1990). In situ ﬁeld measurements of ammonia volatilization from urea and calcium ammonium nitrate applied to grassland. Meststoffen 1/2, 41 –45. Vlek, P. L. G., and Carter, M. F. (1983). The effect of soil environment and fertilizer modiﬁcations on the rate of urea hydrolysis. Soil Sci. 136, 56–63. Vlek, P. L. G., and Crasswell, E. T. (1981). Ammonia volatilization from ﬂooded soils. Fert. Res. 2, 227 –245. Volk, G. M. (1959). Volatile loss of ammonia following surface application of urea to turf or bare soils. Agron. J. 51, 746–749. Warland, J. S., and Thurtell, G. W. (2000). A micrometeorological method for in situ denitriﬁcation measurements using acetylene inhibition. Agric. For. Meteorol. 103, 387 –391. Watson, C. J., and Kilpatrick, D. J. (1991). The effect of urea pellet size and rate of application on ammonia volatilization and soil nitrogen dynamics. Fert. Res. 28, 163– 172. Watson, C. J., Travers, G., Kilpatrick, D. J., Laidlaw, A. S., and O’Riordan, E. (2000). Overestimation of gross N transformation rates in grassland soils due to non-uniform exploitation of applied and native pools. Soil Biol. Biochem. 32, 2019–2030. Webb, E. K., Pearman, G. I., and Leuning, R. (1980). Corrections to ﬂux measurements for density effects due to heat and water vapour transfer. Q. J. R. Meteorol. Soc. 106, 85–100. van der Weerden, T. J., and Jarvis, S. C. (1997). Ammonia emission factors for N fertilizers applied to two contrasting grassland soils. Environ. Pollut. 95, 205–211. van der Weerden, T. J., van der Moal, J. F., Martinez, J., Pain, B. F., and Guiziou, F. (1996). Evaluation of the wind-tunnel method for measurement of ammonia volatilization from land. J. Agric. Eng. Res. 64, 11– 13. Wesely, M. L., Lenschow, D. H., and Denmead, O. T. (1989). Flux measurement techniques. In “Global Tropospheric Chemistry: Chemical Fluxes in the Global Atmosphere” (D. H. Lenschow and B. B. Hicks, Eds.), pp. 31– 46. National Center for Atmospheric Research, Boulder, CO. Wetselaar, R., and Farquhar, G. D. (1980). Nitrogen losses from tops of plants. Adv. Agron. 33, 263 –302. Whitehead, D. C., and Lockyer, D. R. (1987). The inﬂuence of the concentration of gaseous ammonia on its uptake by the leaves of Italian ryegrass, with and without an adequate supply of nitrogen to the roots. J. Exp. Bot. 38, 818–827. Whitehead, D. C., Lockyer, D. R., and Raistrick, N. (1988). The volatilization of ammonia from perennial ryegrass during decomposition, drying and induced senescence. Annals Bot. 61, 567 –571. Whitehead, D. C., and Raistrick, N. (1990). Ammonia volatilization from ﬁve nitrogen compounds used as fertilizers following surface application to soils. J. Soil Sci. 41, 387– 394. Whitehead, D. C., and Raistrick, N. (1993). The volatilization of ammonia from cattle urine applied to soils as inﬂuenced by soil properties. Plant Soil 148, 43– 51.

622

S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD

Wilson, J. D., and Shum, W. K. N. (1992). A re-examination of the integrated horizontal ﬂux method for estimating volatilization from circular plots. Agric. For. Meteorol. 57, 281 –295. Wilson, J. D., Thurtell, G. W., Kidd, G. E., and Beauchamp, E. G. (1982). Estimation of the rate of gaseous mass-transfer from a surface source plot to the atmosphere. Atmos. Environ. 16, 1861–1867. Wilson, J. D., Catchpoole, V. R., Denmead, O. T., and Thurtell, G. W. (1983). Veriﬁcation of a simple micrometeorological method for estimating the rate of gaseous mass transfer from the ground to the atmosphere. Agric. Meteorol. 29, 183–189. Winter, F. P., and Eiland, F. (1996). The effect of soil pH on nitriﬁcation in coarse sandy soil. In “Progress in Nitrogen Cycling Studies” (O. van Cleemput, Ed.), pp. 655–658. Kluwer Academic Publishers, The Netherlands. Wood, C. W., Marshall, S. B., and Cabrera, M. L. (2000). Improved method for ﬁeld scale measurement of ammonia volatilization. Commun. Soil Sci. Plant Anal. 31, 581–590. Yamulki, S., Harrison, R. M., and Goulding, K. W. T. (1996). Ammonia surface-exchnage above an agricultural ﬁeld in Southeast England. Atmos. Environ. 30, 109–118. Zhu, T., Pattey, E., and Desjardins, R. L. (2000). Relaxed eddy-accumulation for measuring ammonia volatilization. Environ. Sci. Technol. 34, 199 –203. Zia, M. S., Aslam, M., Rahmatullah, Arshad, M., and Ahmed, T. (1999). Ammonia volatilisation from nitrogen fertilizers with and without gypsum. Soil Use Manage. 15, 133 –135.

Index A 1:1– 2:1 mixed-layer clays 139 2:1 layer silicate clays 138– 9, 143, 144 –5 abiotic catalysis, humic substances formation 406 –10 accelerated fallows 500, 501 acidity, volcanic soils 160–3 activation tagging, rice 78 –9 adhesion processes 9, 10, 31, 42, 44 –5 adsorption, humic substances 397 –400 aerodynamic gas ﬂux measurement 609–10 AFM see atomic force microscopy Africa, black pepper 343, 345 aggregates, soil stabilization 423– 6 agricultural productivity 115–16, 500–1 see also soil productivity; yields agricultural sustainability 454 –6 agroforestry 479 –84, 503 see also trees agronomy, black pepper 288–95 algorithms, advanced 19 allele mining, rice 96 –7 allelic series, rice 95 alley cropping 479– 84, 505–6, 517, 524 allophanes, Andisols 119–22, 125, 135– 6, 143 –4, 149, 151 aluminum Al3þ availability 120 humus complexes 141, 143, 146, 149 oxides in soil 400 –3 plant pathogens suppression 162–3 toxicity in Andisols 160 –2 volcanic soil genesis 147–50 ammonia emission 557–622 atmospheric 558–60 compensation point 584 –5 crop foliage 581– 6 crops affected by fertilizers 594 –6 injected anhydrous ammonia 594 leaves-atmosphere transport 581 –2 magnitude of losses 582–4 management strategies 586 –98 measurement techniques 560, 598–607, 609 –11 physiological processes 584 –6 soil pH factor 570–4

ammonia volatilization climate interaction 576 –8 microbial processes 578–81 mineral fertilizers 561 –81 soil CEC 574 –5 soils–fertilizer– atmosphere interface 561–5 solid phase processes 575–6 TAN/pH interaction 565 –8 urea hydrolysis 568–70 ammonium carbonate 561 ammonium chloride 561 ammonium nitrate 561 anaerobic chemical weathering 418–19 andic horizons 125, 126 Andisols see also volcanic soils acidity 160– 3 allophanic and nonallophanic 119–22, 135–6, 143 –4, 149, 151 aluminum toxicity 160–2 distribution 117 fertility 163–7 nitrogen dynamics 151–5 organic matter accumulation 145 –7 phosphorus dynamics 155–7 properties 151 Soil Taxonomy 124, 125–7 andosolization 118 Andosols, WRB scheme 124–5, 127 anhydrous ammonia 561 animal manure 496 –9, 525 biomass 507– 8 labour requirement 521 nutrient content 509– 10, 511 anticarcinogens 359–60 antiinﬂammatories 357–8 antimicrobial agents 361–2 antioxidants 360–1 Arabidopsis genome 66 –8 insertional mutants 73–8 mutants 93 Asia, black pepper trade 274 –7, 344 assays, DNA microarrays 183 –270 atmosphere, TAN, soil-fertilizer interface 561–5

623

624

INDEX

atmospheric ammonia 558–60 atmospheric stability 563, 609 atomic force microscopy (AFM) 11 –27, 44– 5 advanced algorithms 19 artifacts in force measurements 17– 18 bond chemistry and energies 19–25 colloidal tip 37 hysteresis 15–16 jumps 15 linkage protocol 38 optical lever collection system 11–12 probes (cantilever/tip) 11 –18, 28– 30 relevance to biological–inorganic interface 25–7 spring constant determination 16–17 tip shape 16 whole cell technique 35–9 attachment strategies, DNA microarrays 191–7 Australia, black pepper trade 344 azotobactin goethite interaction 31, 32, 33, 36 mineral solution 35 oxide interaction 32, 34 B bacteria black pepper 361 –2 interaction with minerals 412, 414 soil formation role 419–20 bacterium–mineral interface forces 2–4, 35–43 Escherichia coli–muscovite 39 –41 Shewanella oneidensis –goethite/diaspore 41–3 whole cell microscopy technique 35–9 biodegradation, xenobiotics 436–8 biodiversity mechanisms 451 bioinformatics, rice 72–3 biological control nematode infestation 317 –19 Phytophthera 314 VAM 314–15, 317 biological evaluation, rice 71– 2 biological nitrogen ﬁxation (BNF) 513–16 biomass banks 490, 492, 512 labour requirement 521 manuring 496– 7 nutrients 511 –12 quality 490, 508– 11

quantity 490, 491, 504–8 transfer techniques 490–3 biomineralization, metals 445 –6 biomolecule–mineral interface forces 27– 35 ligand linkage schemes 28–30 siderophores and oxide surfaces 30 –5 biotechnology agricultural sustainability 454– 6 black pepper 282–5, 370 rice 57 –9 biotin–avidin interaction 27 black pepper (Piper nigrum) 271 –389 see also Indian black pepper economy; Indonesian black pepper economy agronomy 288– 95 biotechnological research 282 –5, 370 botany 279– 85 breeding 281–5 chemistry 285 –8, 322 color 287–8 DTPA extraction 302, 303–5 essential oils 285–7, 335 –9 fertilization 291– 5 future development 369–72 global economy 339– 47 harvesting 322– 3 history of use 273–5 manuring 291, 294 nutrients availability 278–9, 288 –9, 290 –1, 371– 2 buffer power concept 296– 305, 372 deﬁciency symptoms 293, 295 zinc buffer power 301 –5 pests and their control 311–21 Foot Rot 278, 282, 284, 304, 311–15, 349– 50 insect infestation 320 –1 nematode infestation 315–19 pepper weevil 321 Pollu beetle 308, 320–1 pharmacopoeia 355–65 plantation establishment 306 –11 processing 322–8, 330 –9 cleaning and grading 325–7, 330 –1 drying 324–5 packaging and storing 327–8 production and trade 274–5, 276 –9 products 365, 368 –9 propagation 281 –5, 306– 8 root absorbing power 297– 8

INDEX soils 289–90, 310–11 value added products 332–3, 334, 338, 365 –9, 371 yields 277–8, 339 –40 BNF see biological nitrogen ﬁxation bonds chemistry 19–25 polymers 9, 10 botany, black pepper 279–85 Brazil, black pepper supply 341 breeding, black pepper 281 –5 bridging polymers 8–11, 43 Br½nsted acidity 433–5 buffer power concept, zinc 301–5 bulk density, volcanic soils 164–5 burrowing nematodes 316–17 bush pepper, India 281, 351–2 by-products, black pepper 368 –9

C calcium–ammonium –nitrate 561 candidate gene approach, rice 96–7 carbon global ion cycling 448 –9 mineral colloids inﬂuence 426–31 nitrogen ratio 509 carbonic acid 115, 129, 130, 144 catalytic power, primary minerals 406– 9 catalytic transformations, organic pollutants 431 –6 cation exchange capacity (CEC) 574 –5 central nervous system (CNS) 357 –8 CFM see chemical force microscopy CGAs see community genome arrays chavicine 287 chemical control nematodes 317, 319–20 Phytophthera 314 Pollu beetle 320–1 chemical force microscopy (CFM) 28, 45 see alsoatomic force microscopy chemical weathering anaerobic conditions 418– 19 volcanic soils 115, 127–34 chemical-induced mutants, rice 83– 8 chemistry black pepper 285–8, 322 bonds 19– 25 volcanic soils 145–50

625

China, black pepper 274, 344 classiﬁcation, volcanic soils 123–7 clays 1:1–2:1 mixed-layer 139 2:1 layer silicates 138–9, 143, 144 –5 adsorption and pH 398– 400 climate ammonia volatilization impact 576 –8 change and ecosystem effects 449–50 volcanic soil genesis 122– 3 clinical applications, black pepper 363 cluster analysis 228 –30 CNS see central nervous system CO2 weathering cycle 131 –3 colloids formation and transformation 141 –5 soil environmental conditions 142 volcanic soil constituents 135–41 color, black pepper 287 –8 community genome arrays (CGAs) 253–6 companion cropping, black pepper 309 competition, cover crops 485 complexation reactions, metals 441 –2 composting 493–6, 525 conservation, volcanic soils 166–7 consumer products, black pepper 365 consumption, mineral fertilizers 559, 560–1 contact printing, DNA microarrays 199–200 contamination, DNA microarrays 204 continuous cultivation 500–1 coordination, nonhumic organics 396 costs 524–5 covalent bonding, DNA microarrays 192 –4 cover crops 484–90, 506 crop foliage, ammonia emission 581 –6 crop improvement, rice 95–100 cropping see alley cropping; companion cropping; cover crops; intercropping crops see also agricultural productivity; black pepper; rice; yields ammonia absorption 598 ammonia emission 557–622 cover crops 484 –90, 506 high-value 528–9 legumes 476–7, 513–16, 525 cross-species inference, rice 97–8 cryoground pepper 335 crystalline soil components 134–5 Cuba 529–30

626

INDEX

cultivation black pepper 306 –11 history, rice 56– 9 cultural traditions 533 D DAP see diammoniumphosphate data acquisition, DNA microarray image processing 214 –15 data analysis in microarrays 219–31 see also microarrays data transformation 225 differentially expressed genes 225–7 experimental design 222–4, 231–6 genes used 221 –2 normalization 219–25 statistical methods 227 –31 systemic variation sources 219– 21 data collection, forces 11 –27 data mining 186 data processing, forces 18– 19 decomposing plant material 596 deferoxamine (DFO) 31 Deinococcus radiodurans 238 –9 deletion mutations 86–7 demand, black pepper 278, 343 –5, 369 dendrimeric linker coating 195 deskinning, black pepper 333–4 developing countries, rice 98–100 DFO see deferoxamine diammoniumphosphate (DAP) 571, 575 diaspore 32 Shewanella oneidensis forces 41–3 diseases 284, 311–21 insect infestation 320 –1 nematode infestation 315 –19 Phytophthera 278, 282, 284, 304, 311 –15 distillation, black pepper 286, 335– 6 diversiﬁcation, intercropping 476– 7 DLVO theory 40– 1 DNA microarrays 183–270 data analysis 219 –31 environmental studies 236–45 fabrication 190–205 genomic expression monitoring 231–42 hybridization/detection 184 –6, 205–14 image processing 214 –19 types/advantages 187– 9 DNA pool screening, rice 81 drugs, black pepper 355–7

drying, black pepper 324 –5 DTPA extraction, black pepper 302, 303 –5 duration, cover crops 484–5 dynamic force spectroscopy 25–7 E ecology, soil contamination risk 456–9 economic growth, black pepper impact 273, 346– 7 economics, LEIA 523 –6 ecosystems ammonia emission consequences 558–60 climate change impact 449–50 soil, central role 453–4 soil interactions impact 447–59 eddies, ammonia emission measurement 605–6 electrostatic force 5–7, 40–1 electrostatic interaction, DNA microarrays 192, 193 electroultraﬁltration 303 emissions ammonia 557–622 calculation 559–60 EMS see ethyl methanesulfonate encapsulation, black pepper 338 enclosures, ammonia emission measurement 599– 601 energies, force measurements 19–25 energy balance method, gas ﬂux measurement 610– 11 entrapment tagging, rice 79–80 environment, atmospheric ammonia 558 –60 environmental microarray studies 236–45 enzymes interactions, soil mineral colloids 415 –18 soil immobilization 438 –40 Escherichia coli–muscovite forces 39–41 essential oils, black pepper 285–7, 335–9 ethyl methanesulfonate (EMS) 84 Europe, black pepper 343 exotoxicology 453 –4 experimental design, DNA microarrays 222– 4, 231–6 F fabrication of microarrays 186, 190 –205 arraying technology 197 –201 contamination 204 density 201–3

INDEX nucleic acids, attachment 191–7 printing 199 –205 reproducibility 203 storage time 203– 4 fallows 517 –18 accelerated 500, 501 evaluation 537 improved 499–504, 517, 525– 6 farmers perception 538 requirements 483 –4 FARMSIM model 535–6 FAs see fulvic acids ferrihydrite 140 fertilization, black pepper 291–5, 371 fertilizers see also composting; manuring alkaline 571–2 ammonia volatilization 561–81 ammonium 560 animal manure 496– 7, 498 composition 589–91 consumption 559, 560– 1 costs 524 –5 crop ammonia emissions 557–622 FGAs see functional gene arrays ﬂavoured pepper products 366, 367 ﬂexible polymers 8–9 ﬂooded ﬁelds 591 –4 ﬂuorosis 159 Foot Rot (Phytophthera) 278, 282, 284, 304, 311 –15, 349 –50 force–distance curves 12 –15, 18– 19, 20 forces between mineral and biological surfaces 1 –54 see also atomic force microscopy bacterium– mineral interface 35– 43 biomolecule–mineral interface 27–35 bond rupture 23 –4 chemical force microscopy 28, 45 data collection 11 –27 data processing and statistics 18 –19 dynamic force spectroscopy 25– 7 intermolecular 4–11 spectroscopy 25–7 summary 10 formulae, gas ﬂux measurement 609–11 forward genetics, rice 73 –89 forward screening, rice 80–1 fulvic acids (FAs) 379 fumigation, black pepper 332

627

functional gene arrays (FGAs) 243–50 functional validation, rice genes 89–95 fungi interaction, minerals 412, 421 nematode control 318 –19 Phytophthera 278, 282, 284, 304, 311 –15 fungicides, Phytophthera 314 futures market, black pepper 353–5 G gas ﬂuxes formulae 609–11 gel coating, DNA microarrays 195–6 gender issues 533 gene chips see microarrays genes arrays, rice 71 differential expression identiﬁcation 225–7, 231 –42 environmental studies 242 –50 expression analysis 89–91 functional validation, rice 89–95 microarray data analysis 225–34 probes, FGAs 243– 4 replacement, rice 94 –5 silencing, rice 92 tagging, rice 74–80 genetic diversity, Oryza spp. 59– 62 genetic pathways and regulation, rice 97 genetic stocks, rice 70 genome sequencing 185 genomics, rice 55–111 geomedicine 452– 3 germplasm, black pepper 284 –5 global food production 474, 539 global market, black pepper 345, 355 global pepper economy 339–47 global population 473–4 GLR see growth-light regime glucuronidase reporter gene (GUS) 79 goethite azotobactin interaction 31, 32, 33, 36 Shewanella oneidensis forces 41–3 gradient diffusion methods 603 –5 grading, black pepper 325–7, 330 –1 grain legumes 476 –7, 525 grass family genetics 62–4 gravitational attraction 3 green manures 484 green pepper 332 –3, 365 –6, 367–8 Green Revolution 56, 101, 474

628

INDEX

ground pepper 334, 366 growth-light regime (GLR) 281 GUS see glucuronidase reporter gene

hydrophobic force 8 hysteresis 15–16 I

H halloysite 137–8, 144 harvesting, black pepper 322– 3 HAs see humic acids health see also pharmacopoeia black pepper impact 362 ecotoxicology 453–4 hepatic toxicity 358–9 hedgerows, alley cropping 480 HEIA see high external input agriculture hepatic toxicity, black pepper 358–9 heterologous biassays, rice 92–3 high external input agriculture (HEIA) 474 –5, 526–7 high-troughput technologies, rice 70– 1 high-value crops 528–9 hisingerite 141 holistic approach, LEIA 531, 541 Honduras Mucuna system 486 –7, 530– 1 Qezungual System 503–4 household models, LEIA technologies 534–5, 536–7 human health 362, 453 –4 see also pharmacopoeia human nutrition 451–4 humic acids (HAs) 397, 408 –9 humic substances binding mechanisms 397–400 formation 406 –10 humiﬁcation, Maillard reaction 410, 411 humus 141, 143 hybridization in microarrays 184–6, 205 –14 detection 210 –14 labeling 206 –10, 212 –13 practical issues 211– 14 probe design and synthesis 205– 6 probe DNA retention 211–12 probe–target hybridization 210 spatial resolution and cross-talk 213 hydration force 7–8 hydrogen bonding, nonhumic organics 396 hydrolysis, urea 568– 70 hydrophobic bonding, nonhumic organics 397

image processing, microarrays 214–19 imogolite 119, 120, 121, 125, 137, 143 –4 see also allophanes improved fallows 499–504, 517, 525 –6 Indian black pepper economy 347–55 bush pepper 351–2 Foot Rot 313, 314 futures market 353–5 harvesting 322 manuring 291, 294 marketing 352–3 nematodes 315 Piper species 279– 82 plantations 306 –10 production 349 –52 soils 289–90 supply 340 –1 trade 275 –6 Indonesian black pepper economy plantations 310 –11, 312 processing 328–30 supply 341 –2 trade 276 –7 industrial processing, black pepper 330–9 inﬁltration, ammonia volatilization 576 –8 INM see integrated nutrient management insect infestation, black pepper 320–1 insecticides 321, 364 –5 insertion sequence database, rice 81 –3 insertional mutants, rice 73 –83 integrated nutrient management (INM) 526–30 interactions biodiversity mechanisms 451 ecosystem impact 447 –59 enzymes, soil minerals 415–18 humic– enzyme complexes 415–18 metal transformation 440–7 mineral colloids 394–400 miocroorganisms, soil minerals 410– 14 organic pollutants 431–40 soil mineral colloids 410–26 intercropping 476–9, 517, 525 interface forces see intermolecular forces intermolecular forces 1–54 atomic force microscopy 11–27 bacterium–mineral interface 2–4, 35 –43

INDEX biomolecule–mineral interface 27–35 electrostatic 5–7, 40 –1 solvation 7– 8 steric 8–11 van der Waals 3–5, 40 –1 international collaboration, rice 98–100 International Rice Genome Sequencing Project (IRGSP) 58, 65 International Society of Soil Science (ISSS) 393 ion cycling, ecosystem interaction 448 –9 ion exchange, nonhumic organics 395–6 ion– dipole interaction 396 IR29 rice 90 IR64 rice 85 IRGSP see International Rice Genome Sequencing Project iron oxides 403–6 iron–humus complexes 141, 146 irradiation-induced mutants, rice 83–8 J Japan, black pepper 344 K kaolinite-smectite 139 Kerala, India, black pepper economy 275 –6, 349 –51, 352 –3 “King of spices” see black pepper L labeling of microarrays 206–10, 212–13 labour requirement 520–3, 540 biomass 493, 521 composting 494 –5 improved fallows 502– 3 Machobane farming system 478– 9 manuring 498–9, 521 rice intensiﬁcation 522–3 weed control 521– 2 land alley cropping 482–3 fragmentation 520 Qezungual System 503 requirement 516 –18, 540 tenure 483, 488, 518 Langrangian dispersion models 606– 7

629

Latin America, black pepper 344 legumes biological nitrogen ﬁxation 513– 16 grain legumes 476– 7, 525 LEIA see low external input agriculture Lennard–Jones potential 13–14, 15 Leptothrix, Mn oxide 420 Lesotho, intercropping 478–9 Leucaena, alley cropping 480 Lewis acidity 433–5 ligands linkage schemes 28–30, 44–5 siderophores and oxide surfaces 30–5 light conditions 597 light-directed synthesis, photolithography 197–9 light-growth regime 281 linkages, ligands 28–30, 44–5 livelihood improvement technologies 473 –555 loading, molecular bonds 22–5 low external input agriculture (LEIA) 473– 555 alley cropping 479 –84 animal manure 496 –9 biomass transfers 490–3 composting 493–9 cover crops 484 –90 green manures 484 improved fallows 499 –504 integrated nutrient management 526– 30 intercropping 476–9 modelling 534– 8 socio-economic issues 516 –26, 538 soil fertility management 504–16 systems perspective 530 –4 Lupinus sp., nitrogen ﬁxation 129 –30 M Machobane farming system 478 –9, 531 macro-colinearity, rice 63–4 Maillard reaction 410, 411 maize, cover crop 486– 7, 489 Malaysia, black pepper 276, 277, 342 mammalian cells 37 management, volcanic soils 166–7 management of ammonia emission 586– 98 decomposing plant material 596 –7 fertilizer application 594–6 fertilizer composition 589–91 reduction techniques 587 –9 rice paddies 591 –4

630

INDEX

manuring see also animal manure; composting biomass 496 –7 black pepper 291, 294 green manures 484 labour requirement 498– 9, 521 Machobane farming system 479 marketing, black pepper 352–3 measurement of ammonia emission 560, 598–607, 609–11 eddy correlation and accumulation 605–6 enclosures 599– 601 gradient diffusion methods 603–5 Langrangian dispersion models 606–7 micrometeorological methods 601–3 tracer techniques 598 –9 medicines, black pepper 355–7, 363 melanization 118 MeNPOC 198 metals biomineralization 445–6 oxides, organic substances inﬂuence 400–6 precipitation effects 444– 5 reactions, soil 441–4 transformation, biological processes 440–7 micro-colinearity, rice 64 microarrays 183– 270 advantages 188 –9 data analysis 219 –31 environmental studies 236–45 fabrication 186, 190– 205 image processing 214 –19 substrates 190–1 surface modiﬁcation 191–7 types 187–8, 258 microbial cells 38 microbial processes, ammonia volatilization 578–81 microchips see microarrays micrometeorological methods 601–3 micronutrients, black pepper 294 microorganisms interactions, soil minerals 410–14, 421 –6 metal sorption 445 –6 organic pollutants 431 –40 micropropagation, black pepper 284 microspotting, DNA microarrays 199–200 mineral colloid interactions 394–400 humic–enzyme complexes 415 –18 mineral fertilizers see fertilizers mineral surfaces, humic substances 397–8

mineral–bacterium interface forces 35 –43 mineral–biomolecule interface forces 27– 35 mineralogy volcanic rocks 127, 128 volcanic soils 134 –45 model genetic system, rice 59– 64 modelling, LEIA technologies 534–8 molecular bond chemistry 19–25 MPT see multipurpose tree techniques Mt. St. Helens volcano 129, 132, 157 Mucuna, cover crop 485 –9, 530 multipurpose tree (MPT) techniques 501 –2 muscovite–Escherichia coli forces 39–41 mutagenic effects, black pepper 359–60 mutants chemical- and irradiation-induced 83 –8 deletion and point mutation stocks 85–8 mutagens 84 –5 N natural genetic variation, rice 88–9 nematicides 319–20 nematodes, black pepper 315 –19 neural network analysis 230–1 nitriﬁcation 578–81 nitrocellulose coating, DNA microarrays 196–7 nitrogen see also total ammoniacal nitrogen accumulation and mineralization 151–3 animal manure 509–10, 511 biological ﬁxation 129–30, 513–16 biomass 505, 508 black pepper 291–2 carbon/nitrogen ratio 509 fertilizer efﬁciency 154–5 global ion cycling 448– 9 immobilization 578– 81 leaching in Andisols 153–4 mineral colloids inﬂuence 426 –31 uptake and translocation 585 –6 volcanic soil genesis 151–5 non-contact ink-jet printing 200–1 non–porous microarray substrates 190–1 nonallophanic Andisols 119– 22 noncrystalline material, volcanic soils 115, 118, 127, 134 nonhumic organics, binding mechanisms 394, 395– 7 North America, black pepper trade 275, 344 nucleic acid attachment, microarrays 191–7

INDEX NUTMON project 529 nutrient buffer power concept 296–305, 372 measurement 298 –301 soil solution 296–7 zinc buffer power 301–5 nutrients see also manuring; nitrogen; organic matter; phosphorus; potassium; zinc animal manure 509– 10 biomass transfers 490 –3, 505 black pepper 278–9, 288–9, 290 –1, 371– 2 costs 524 deﬁciency symptoms (black pepper) 293, 295 integrated management 526–30 mining 511 –13 organic 527–8 pepper soils 371–2 root concentration 298–301 subsistence agriculture 528 O odor, black pepper 287 oleoresin 281, 285, 326 –7, 329 –30, 336 –8, 369 oligonucleotide microarrays 183–270 oligonucleotide probes 206 OM see organic matter opaline silica 139–40 open reading frame (ORF) arrays 229 –30, 256 –7 organic ligands 27–30, 44–5 organic matter (OM) 507, 508–10, 511 see also biomass; soil organic matter particulate matter 421 –6 volcanic soils 115, 118, 145 –7 organic pollutants, formation 431– 40 organically grown black pepper 278, 291, 294, 370 Oryza see also rice biotechnology 57 –9 gene and allele reservoir 62 genomes 60 improvement program species 60–1 O. glaberrima 59–60 O. indica 60–1, 65, 68–9 O. japonica 60–1, 68–9 O. nivara 61 O. sativa 59–60

631

oxides aluminum 400– 3 azotobactin interaction 32, 34 iron 403 –6 manganese 420 organic substances inﬂuence 400– 6 redox reactions 440 –1 surfaces 30–5 P packaging, black pepper 327–8 paddy rice 56, 591–4 particulate organic matter (POM) 421–6 see also organic matter pepper see also black pepper ancillary by-products 368– 9 bush pepper 281, 351–2 cryoground 335 dehydrated green 332–3 green 365–6, 367 –9 ground 334, 366 harvesting schedule 322 pharmacopoeia 355– 65 value added products 332–3, 334, 338, 365–8 white 329, 333– 5, 368 pepper oil 329, 335 –9 pepper weevil 321 pests see also diseases black pepper 308, 311–21 pH adsorption dependence 398 clays 398 –400 soils, ammonia emission s 570 –4 TAN interaction 565–8 pharmacopoeia, black pepper 355 –65 phenotyping, rice 71–2 phosphorus biomass 505, 508, 512, 521 black pepper 292 global ion cycling 448–9 mineral colloids inﬂuence 426–31 volcanic soil genesis 155– 7 photolithography, light-directed synthesis 197–9 phylogenetic oligonucleotide arrays 250– 3 physical properties, volcanic soils 163 –7 physiological processes, ammonia emissions 584–6

632

INDEX

Phytophthera (Foot Rot) 278, 282, 284, 304, 311–15, 349–50 phytosanitation 313 pin conﬁguration, DNA microarrays 202–3 Piper nigrum see black pepper Piperaceae 279 piperine 281, 285, 287, 326 –7, 337, 356–7 plant pathogens, suppression by aluminum 162–3 plantations, black pepper 306 –11 point mutations 87–8 Pokkali rice 90 pollination, black pepper 280 Pollu beetle, black pepper 308, 320–1 pollution 436–40, 452 –3 polygenetic soil proﬁle 119 polymers, bridging 8 –11, 43 POM See particulate organic matter porosity, volcanic soils 164 –5 porous substrates, microarrays 190– 1 potassium, black pepper 292 poverty, pathways out of 537, 538 precipitation, metals transformation 444 –5 prices, black pepper 340 –1, 348 principal component analysis (PCA) 228 printing technologies, microarrays 197–201 probe–target hybridization 189, 210 probes, AFM 11–18, 28–30 processing, black pepper 322–8, 330–9 productivity, volcanic soils 115 –16, 150, 160– 3 propagation, black pepper 281 –5, 306– 8 protein proﬁles, rice 91 protonation, nonhumic organics 396 pruning, black pepper 308 Pyrococcus furiosus 239, 241–2 Q Qezungual System, Honduras 503–4 quality control microarray fabrication 203 microarray image processing 215– 19 quantitative measurements, microarrays 247–9, 255–6 R rainfed rice 56 redox reactions, metal oxides, soils 440–1 reduction techniques, ammonia emissions 587– 9 reductionist approach, LEIA 531, 541

reproducibility, microarray quality 203 research see also biotechnology; rice genomics black pepper 370–2 rice 57 –9 retrotransposon tagging, rice 77–8 reverse genetics, rice 73– 89 rhizosphere interactions, environmental 451 –2 metals, transformation 440–4 rice paddies 56, 591 –4 System for Rice Intensiﬁcation 512–13, 522– 3 Rice Biotechnology Program 57–8 rice genomics 55 –111 see also Oryza allelic series 95 bioinformatics 72– 3 biological evaluation 71 –2 candidate gene approach and allele mining 96– 7 chemical- and irradiation-induced mutants 83– 8 crop improvement 95–100 cross-species inference 97 –8 cultivation history 56–9 features and composition 66 –8 functional validation 89 –95 gene discovery ingredients 69–73 gene expression analysis 89 –91 gene replacement 94–5 gene silencing 92 genetic diversity 59– 62 genetic stocks 70 heterologous bioassays 92–4 high-throughput technologies 70 –1 insertional mutants 73 –83 international collaboration 98–100 IR64 variety 85 model genetic system 59–64 natural genetic variation 88–9 Oryza indica/japonica comparison 68–9 pathways and genetic regulation 97 research 57 –9 salient features 66–8 sequencing 64–5 single genetic system 62–4 risk assessment, soil contamination 456– 7 RNA, microarray assay 183 –270 Rockefeller Foundation 57–8

INDEX roots absorbing power (black pepper) 297 –8 exudates 442–4 fungus infection 312 nutrient concentration 298–301 root knot nematodes 315–16 rupture forces, molecular bonds 23–4 Russia, pepper imports 348 S SBH see sequencing by hybridization scatter plots, microarray data 227 self-organizing maps (SOMs) 230–1 sensitivity, DNA microarrays 245, 246– 7, 255 –6 sequencing by hybridization (SBH) 185 share-cropping 518–20 Shewanella oneidensis DNA microarray analysis 236–8 goethite/diaspore forces 41–3 shifting cultivation 499–500 siderophores 30–5, 45 silanization, DNA microarrays 195 silica opaline 139–40 volcanic materials 127 silicates, 2:1 layer 138–9, 143, 144 –5 silicon nutrition in volcanic soils 160 soil solution activity 133 –4 similarity measurement, microarray data 227– 8 SOC see soil organic carbon fraction socio-economic issues of LEIA 516 –26, 538 economics 523 –6 labour 520– 3 land 516 –20 soil environmental conditions, colloid formation 142 soil erosion, alley cropping 482 soil fertility 504–16 animal manure 498 biological nitrogen ﬁxation 513–16 biomass quality 508 –11 biomass quantity 504 –8 biomass transfers 492 composting 494 nutrient mining 511–13 pepper soils 371 volcanic soils 157–8, 163 –7 soil genesis, volcanic materials 118– 23

633

soil mineral colloids binding mechanisms 394–400 biogeochemical cycling 426– 31 interactions 410 –26 surface properties 394 –5 soil minerals humic substances, formation 406 –10 precipitation 421–3 weathering 418– 20 soil moisture, Andisols 126 soil organic carbon fraction (SOC) 504–8 soil organic matter (SOM) 115, 118, 145– 7, 426–31, 504, 505, 506 soil productivity, volcanic soils 150, 160 –3 soil properties, alley cropping 481–2 soil solution 296– 7, 441–2 soil structure, volcanic soils 163 –4 Soil Taxonomy, Andisols 124, 125 –7 soil texture, volcanic soils 163–4 soils aggregation dynamics 423 –6 ammonia emissions, pH factor 570–4 atmosphere, TAN 561 –5 black pepper 289– 90, 310–11 CEC 574– 5 contamination risk 456–9 ecosystem role 453–4 formation, microorganisms 419–20 metals, reactions 441 –4 restoration 457–9 structure stabilization 421–6 TAN, microbial processes 578–81 volcanic 113–82 solvation force 7–8 SOM see soil organic matter speciﬁcity, oligonucleotide microarrays 244–6, 251–2 spices see black pepper spot quality, DNA microarrays 202, 214 –19 SRI see System for Rice Intensiﬁcation Sri Lanka, black pepper trade 276–7, 307, 342 stabilization, soil structure 421–6 statistical methods, microarray data 227–31 statistics, forces 18 –19 steric force 8–11 sterilization, black pepper 330 –2 storage times, DNA microarrays 203–4 structural force 7–8 struvite, precipitation 575–6 subsistence agriculture 473 –555 see also low external input agriculture

634

INDEX

substrates, microarrays 190–1 sulphur global ion cycling 448–9 mineral colloids inﬂuence 426 –31 supply, black pepper 339 –43 surface modiﬁcation, microarrays 191–7 sustainable livelihoods approach 531 –2, 533, 534 System for Rice Intensiﬁcation (SRI) 512–13, 522–3, 530–1 T T-DNA tagging, rice 74–5, 78 –9 TAN see total ammoniacal nitrogen tephra, chemical weathering 129 –33 Thailand, black pepper 342 Tithonia diversifolia 512, 517, 521 biomass transfers 491–2 topography, black pepper plantations 307–8 total ammoniacal nitrogen (TAN) climate interaction 576–8 microbial processes 578– 81 pH interaction 565 –8 soil–fertilizer–atmosphere interface 561–5 toxicity, aluminum 160 –2 toxicological effects, black pepper 363–4 tracer techniques, ammonia emission measurement 598–9 transcription proﬁling, rice 89–90 transposon tagging, rice 75–7 trees alley cropping 481 biomass production 505 –6 cover crops 489 multipurpose tree techniques 501–2 U urea climate interaction 576–8 consumption 560– 1 hydrolysis 568 –70 V value added pepper products 332, 334, 338, 365–9, 371 VAM see vesicular arbuscular mycorrhizae van der Waals force 3–5, 40– 1

vesicular arbuscular mycorrhizae (VAM) 314– 15, 317 volatilization see ammonia volatilization volcanic rocks (ejecta) 114, 116, 124, 129 –33 glass 127–9, 133–4 mineralogy 128 weathering 127–9, 132, 133–4 volcanic soils 113 –82 see also Andisols; Andosols acidity 160–3 aluminum dynamics 147–50 bulk density and porosity 164–5 charge characteristics 150– 1 chemical characteristics 145 –50 chemical weathering 115, 127–34 classiﬁcation 123– 7 colloid formation and transformation 141– 5 colloidal constituents 135–41 distribution 114–15, 116–18 fertility 157–8, 163–7 genesis 118–23 management and conservation 166 –7 mineral nutrients 157–9 mineralogical characteristics 134– 45 nitrogen dynamics 151 –5 noncrystalline material 115, 118, 127, 134 organic matter accumulation 145–7 phosphorus dynamics 155–7 physical properties 163–7 productivity 115 –16, 150, 160–3 properties 115–16 silicon nutrition 160 texture and structure 163 –4 water retention and availability 165–6 volcanism 114, 116 –17 W water retention, volcanic soils 165– 6 weathering see also chemical weathering CO2 cycle 131 –3 soil minerals 418–20 temperature and moisture conditions 130 volcanic soil genesis 119–20, 122 –3, 127 weed control 539 alley cropping 481 cover crops 485–6, 489–90 labour requirement 521 –2 weevils, black pepper 321

INDEX Western Ghats, India, black pepper 281 –2, 290 wheat streak mosaic virus (WSMV) 93–4 white pepper 329, 333–5, 368 whole cell force microscopy technique 35–9 Whole-genome Open Reading Frame arrays 256 –7 wilt, black pepper 312, 313 Working Group MO (1990) 393 World Reference Base for Soil Resources (WRB), Andosols 124–5, 127 WSMV see wheat streak mosaic virus X xenobiotics, bioavailability 432, 436–40

635 Y

‘yellows’, black pepper 317 yields 539 alley cropping 480, 524 animal manuring 497 black pepper 277– 8, 339–40 intercropping 477 organic matter fertilization 510 rice 56–7 Z zinc buffer power in black pepper 301 –5 deﬁciency in black pepper 294, 295

This Page Intentionally Left Blank