VOLUME46
$3
Advisory Board Martin Alexander
Eugene J. Kamprath
Cornell University
North Carolina State University
Kenneth J. Frey
Larry P. Wilding
Iowa State University
Texas A&M University
Prepared in cooperation with the American Society of Agronomy Monographs Committee S. H. Anderson L. P. Bush R. N. Carrow
M. A. Tabatabai, Chairman G. L. Horst R. J. Luxmoore R. H. Miller
G. A. Peterson C. W. Stuber S. R. Yates
Edited by
Donald L. Sparks Department of Plant and Soil Sciences University of Delaware Newark, Delaware
ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers San Diego New York Boston London Sydney Tokyo Toronto
This book is printed on acid-free paper. @
Copyright 0 1991 BY ACADEMIC PRESS, INC. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.
Academic Press, Inc. San Diego, California 92101 United Kingdom Edition published by ACADEMlC PRESS LIMITED 24-28 Oval Road, London NWI 7DX
Library of Congress Catalog Card Number: 50-5598
ISBN 0-12-000746-0 (alk. paper)
PRINTED IN THE UNITED STATES OF AMERICA 91
92 93 94
9 8 7
6 5
4
3 2
1
Contents .......................................... CONTRIBUTORS PREFACE ................................................ INFLUENCE OF SOILMACROPOROSITY ENVIRONMENTAL QUALITY
ix xi
ON
J. Bouma I. 11. 111. IV. V. VI.
Introduction ............................................ Macroporosity: Morphological and Physical Characterization ..... Effects of Macropores on Solute Movement ................... Methods to Study Macropore Flow .......................... Field Studies on Macropore Flow and Environmental Quality ..... Conclusions ............................................ References .............................................
1 2 13 25 32 33 34
DNA MARKERS IN PLANTIMPROVEMENT Andrew H. Paterson, Steven D. Tanksley, and Mark E. Sorrells I. Introduction: Agricultural Genetics and DNA Markers .......... 11. How and W h y are Genetic Maps Made? ..................... 111. Using Genetic Markers to Study and Improve Agricultural Productivity ............................................ IV. Describing Individual Quantitative Trait Loci ................. V. Cumulative Effects of Many Quantitative Trait Loci on Phenotype ofanIndividua1 .......................................... VI. Improving the Efficiency of Breeding Programs ................ VII. Marker-Facilitated Study of Complex Populations .............. VIII. Marker-Facilitated Study of Polyploids ....................... IX. Utilization of Exotic Germplasm in Crop Improvement ......... X. Revealing Evolutionary Relationships among Crop Species and Their Wild Relatives-Utility in Comparative Genetic Mapping ....... XI. Cloning Genes from Map Position? ......................... XII. Challenges for the Future? ................................. XIII. Conclusion ............................................. References ............................................. V
40 42
48 54 61 64 67 69 71
75 76 80 82 82
vi
CONTENTS
SOILSCIENCEAPPLICATIONS OF NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY William F. Bleam
I . Introduction and Overview ................................ I1. A Layman’s View of Modern NMR Spectroscopy .............. I11. NMR Spectroscopy as an Experimenu The Design of NMR Studies ........................................... IV . Current Developments and Their Future Implications ........... V . Conclusions ............................................ VI . Appendices ............................................. Symbols ................................................ References .............................................
91 93 112 138
146 150 151
POTENTIAL OF BROWN.MIDRIB. LOW-LIGNINMUTANTSFOR IMPROVINGFORAGEQUALITY J . H. Cherney. D.J . R . Cherney. D . E . Akin. and J . D. Axtell I . Introduction ............................................ 158
I1. History ................................................ 111. Genetics ............................................... IV . Lignin Biosynthesis ...................................... V . Forage Quality - Plant Response ............................ VI . Phenolic- Carbohydrate Complexes ......................... VII . Plant Morphology and Anatomy ............................ VIII . Forage Quality-Animal Response ........................... IX . Biotechnology Potential ................................... X . Summary and Conclusions ................................. References .............................................
159 160 161 163 173 176 184
190 191 192
MEASUREMENT OF SURFACE CHARGE OF INORGANICGEOLOGIC MATERIALS: TECHNIQUES AND THEIR CONSEQUENCES Anne Lewis-Russ I. Introduction ............................................ I1. Surface Charge Terminology ............................... I11. Surface Charge Development ............................... IV . Measurement of Surface Charge ............................ V . Separating Permanent and Variable Charge ................... VI . Measurement Problems ................................... VII . Applications and Predictions for Composite Materials ........... VIII . Summary ............................................... References .............................................
199 201 204 209 219 222 231 236 2 38
CONTENTS
vii
GENETIC IMPROVEMENTOF MAIZE YIELDS W . A . Russell
I . Introduction ............................................ I1. Genetic Gains in Grain Yield ..............................
I11 . Genetic Gain: Stress versus Nonstress Environments ........... IV Response to Increase in Plant Density and Nitrogen Fertility ..... V . Changes in Other Plant Traits .............................. VI. Genetic Gains via Recurrent Selection in Populations ........... VII . Improvement of Inbred Lines .............................. VIII . Future Trends ........................................... References .............................................
.
INDEX
...........................................................
245 249 260 262 270 282 290 292 294 299
This Page Intentionally Left Blank
Contributors Numbers in parentheses indicate the pages on which the authors’ contributionsbegin.
D. E. AKIN (1 57), RichardB. RussellAgriculturalResearch Center,Agricultural Research Service, U.S. Department of Agriculture, Athens, Georgia 3061 3 J. D. AXTELL ( 1 57), Department of Agronomy, Purdw University, West Lafyette, Indiana 47907 WILLIAM F. BLEAM (9 l), Department of Soil Science, Universityof Wisconsin, Madison, Madison, Wisconsin 53 706 J . BOUMA (l),Department of SoilScience and Geology,Agricultural University, Wageningen, The Netherland D. J . R. CHERNEY ( 1 57), Department ofAnimal Science, Cornell University, Ithaca, New York 14853 J . H. CHERNEY (1 57), Department of Soil, Crop, and Atmospheric Sciences, Cornell University,Ithaca, New York 14853 ANNE LEWIS-RUSS ( 1 99), Department of Chemistryand Geochemistry,Colorado School of Mines, Golden, Colorado 80401 ANDREW H . PATERSON (39), TexasA &M UniversityDepartment of Soil &Crop Sciences College Station, Texas 77843 andDepartment ofplant andsoil Sciences, University of Delaware, Newark, Delaware I 9 71 7 W . A. RUSSELL (245), Department OfAgronomy,Iowa State University,Ames, Iowa 5001I MARK E. SORRELLS (39), Department of Plant Breeding and Biometry, Cornell University,Ithaca, New York 14853 STEVEN D. TANKSLEY (39), Department of Plant Breeding and Biometry, Cornell University,Ithaca, New York 14853
This Page Intentionally Left Blank
Preface The field of Agronomy has changed greatly since the publication of Volume 1 of Advances in Agronomy in 1949. Yet with all these innovations, many challenges remain. In order to continue the tradition of this series which features cutting edge reviews on topics in the plant and soil sciences, leading scholars from throughout the world will survey advances in classical areas, such as plant physiology and nutrition; plant genetics and breeding; soil physics, chemistry, microbiology, genesis, and mineralogy. In addition, exciting and innovative studies in the broad areas of plant biotechnology, terrestrial ecosystems, and environmental concerns will be included. Because novel technologies and approaches are needed in these areas, topics at the forefront of agronomic research will be reviewed, including bioengineering to develop plants that are resistant to adverse environments, herbicides, diseases, and insects; molecular markers for plant genetics and breeding; improved methods for manipulation of rhizosphere microorganisms to enhance plant growth; sustainable agriculture; development and use of alternative crops; characterization and enhancement of biodegradation for organic pollutants; ways to minimize and predict inorganic and organic pollutant transport in soils; and amelioration of soil acidity and salinity. Volume 45 ofAdvancesin Agronomy presents state-of-the-artadvances in several of the above areas. Plant productivity and improvement are discussed in three chapters. One discusses the development of maize inbred lines and hybrids and their tremendous effect on grain yields throughout the world. The tools of molecular biology also offer exciting opportunities for improving plants; the use of DNA markers for enhancingplants is discussed in a second chapter. Forages are extremely important; however, their lignin content often limits their usefulness in ruminants. Using brown-midrib mutants to modify lignin quality and quantity, a new developmentwhich is discussed in a third chapter on plant improvement, has great potential to improve forage quality. Such plant improvementswill undoubtedly provide for a sustainable agriculture. Three chapters cover advances in soil science. One discusses the importance of macropores in soils and how they affect the movement of pollutants. Characterization of macropore using different techniques is included. Spectroscopic techniques offer immense opportunities for characterizing chemical reactions at soil surfaces. A chapter on nuclear magnetic resonance xi
xii
PREFACE
(NMR) spectroscopyto soils covers this topic. A third chapter on advance in soil science provides a critical review of surface measurements for practitioners working with soils and minerals. Such measurements are important in understanding a number of chemical processes in soils. I appreciate the fine contributions from the authors and the many suggestions of the Advisory Board. DONALD L. SPARKS
INFLUENCEOF SOILMACROPOROSITY ON ENVIRONMENTAL QUALITY J. Bouma Department of Soil Science and Geology Agricultural University Wageningen, The Netherlands
I. Introduction 11. Macroporosity: Morphological and Physical Characterization A. Types of Pores B. Pore and Pore-size Distributions 111. Effects of Macropores on Solute Movement A. Basic.Introduction to Water Flow Based on Morphometric Pore Analysis B. Predicting K,, of Soils with Macropores C. Breakthrough Curves: Displacement of Liquid D. Saturated and Unsaturated Flow E. Bypass Flow IV. Methods to Study Macropore Flow A. Measuring K at or near Saturation B. Measuring Bypass Flow C. Pedotransferhnctions D. Modeling V. Field Studies on Macropore Flow and Environmental Quality A. Introduction B. Some Environmental Effects of Macropore Flow Case Studies VI. Conclusions References
I. INTRODUCTION Environmental research requires more attention than is currently being provided to processes of water and solute movement under undisturbed conditions in the field. Environmental laws and regulations increasingly define specific limits and levels of contaminants in soil and groundwater as a function of application regimes and land management procedures. To allow for quantitative predictions of leaching and adsorption rates of 1 Advances in A p v ~ m y V, d 46 Copyrighr 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
2
J. BOUMA
contaminants, basic transport and transformation processes of solutes are being characterizedby mathematicaldescriptionsand simulation modeling. Modeling has been quite successful for many homogeneous flow systems both in the field and in the laboratory. Many field conditions have been encountered, however, where rates of movement of contaminants in soil were much faster than could be explained by considering basic transport and adsorption parameters of the soil material. These conditionshave been well documented and are often attributed to the occurrence of macropores (Beven and Germann, 1982; White, 1985; Van Genuchten el al., 1990). This chapter will not repeat what has already been reviewed elsewhere. So far, most emphasis in reviews has been placed on mathematicaldescriptions of transport and transformationprocesses. Here, attention will be focused on field characterization of flow processes using relatively simple morphological and physical techniques that are accessible to agronomists. Special emphasis will be placed on: (1) a descriptive characterization of macropore patterns in soil, using different morphometric techniques; (2) relating morphology to physical flow processes; (3) description of field techniques to characterize macropore flow; and (4) relating macropore flow to environmental quality by analyzing some field studies.
11. MACROPOROSITY: MORPHOLOGICAL AND PHYSICAL CHARACTERIZATION
A. TYPES OF PORES Macropores can be of different types. According to Brewer ( 1964), their dimensions are larger than those that would correspond with a simple packing of the elementary soil particles. Macropores are larger than 75pm. Macropore continuity is very important for many soil functions and this factor, preferably expressed in terms of well defined dynamic soil -physical terms such as hydraulic conductivity, should receive emphasis in research rather than arbitrarily defined size limits (Bouma, 1981). Occurrence of macropores in different soil materials can be expressed as follows: soil materials with natural aggregates (peds) are called pedal (Fig. lA), and those without peds apedal (Fig. 1B). In both cases, the primary particles constitute the bulk volume of the soil, which is called the groundmass. This is, therefore, the material within the simplest peds, or within apedal soil material. The structure of the groundmass can be described and measured by noting size, shape, and arrangement of the solid particles and the voids. Generally, clayey soils (which
INFLUENCE OF SOIL MACROPOROSITY basic structure
3
basic structure planar voids dal vugh
A fragment of apedal roil marerial
1 crn
pedal soil material c _ _
1 crn
Fig. 1. Simplified structure models in a p e d and pedal soil materials. (See text.)
swell upon wetting and shrink upon drying) have peds, and their basic structures contain more clay (plasma) than basic structures of apedal sandy soil materials. (Fig. 1). The following major types of voids are distinguished: (1) simple packing voids that are voids due to random packing of single grains (Fig. 2); (2)
Fig. 2. Simple packing voids between sand grains.
4
J. BOUMA
Fig. 3. Compound packing voids between natural aggregates (peds).
Fig. 4. Voids that are significantlylarger than simple packing voids but without continuity (vughs).
INFLUENCE O F SOIL MACROPOROSITY
5
Fig. 5. Tubular void, formed by a large root (channel).
compound packing voids that are voids that result from the packing of compound individuals, such as peds, which do not accommodateeach other (Fig. 3); (3) vughs are voids that are significantly larger than simple packing voids, and they appear as discrete entities at the magnification at which they are recognized (Fig. 4); (4) channelsthat are voids that are significantlylarger than simple packing voids and generally have a cylindrical elongated shape (Fig. 5); and (5) planes that are voids that are planar accordingto the ratios of their principle axes (Fig. 6). By virtue of their shape and length, voids constitute an obvious deviation from the normal packing ofprimary soil particles. These pores are schematically shown in Fig. 1 for pedal and apedal soil materials.
B.
PORE AND PORE-SIZE
DISTRIBUTIONS
1 . Morphometric Techniques
Many morphometric techniques are available to measure (macro)poresize distributions. Techniques to obtain samples range from undisturbed plastic-impregnated soil samples that are used to make thin sections of soils to be studied under the microscope to field casts of macropore patterns obtained with plaster of Paris or different plastics (Bouma et al., 1977, 1982; Fitzpatrick et al., 1985). (Fig. 7). Techniques to obtain pore-size distributions range from relatively simple point counts (Van der Plas and Tobi,
6
J. BOUMA
Fig. 6. Planar voids formed by physical shrinkage processes in a clay soil.
1965),to detailed measurements with image analyzers, such as the Quantimet (Bouma et al., 1977; Murphy et al., 1977a,b).Two-dimensional images obtained have to be transformed into data representing three-dimensional volumes of soil. Again, point counts can be used which yield two-dimensional area percentages that are representative for three-dimensional volumes (Brewer, 1964). Stereology techniques have been applied recently to achieve the same objective (Ringrose-Voase and Bullock, 1984; Mackie, 1987). Emerging new technologies are computed tomography (CT) and nuclear magnetic resonance (NMR) techniques that allow, in principle, generation of nondestructive three-dimensional images of macropores (Warner et al., 1989).Detailed discussions of these techniques is beyond the scope of this article. 2. Soil Physical Techniques
Often, reference is made in the soil science literature to pore-sizedistributions that are based on moisture retention data, using the concept of “equivalent” sizes derived from a capillary model (Fig. 8). This model relates the height of capillary rise (h) to pore radius:
h = 2rr/pgr (1) where CT = surface tension of the water (N/m), p = density of the water (kg/ m3),g = gravitational constant (m/s2),and r = radius of capillary tube (m). The relationship expressed in Eq. (1) can be pictured as a continuous
INFLUENCE OF SOIL MACROPOROSITY
7
Fig. 7. Cast of channelsusing plaster of Paris.A represents a vertically continuouschannel. B represents a channel that connects with a mole burrow yielding very high infiltration rates.
graph, relating capillary diameter (2r) to corresponding negative pressure (Fig. 8). The negative pressure below the meniscus in the water (in Pascals) can be expressed in terms of the equivalent height of a column of water (in centimeters) that can b2 "pulled" from a cup of water by the capillary tube (since 1 mbar = 100 Pa 1 cm water). Figure 8 illustrates that fine pores can exercise a larger pull than large pores. For example, a cylindrical pore diameter of 100pm corresponds to a relatively low capillary rise of 28 cm water (pressure below meniscus = -28 cm water) and a diameter of 30 pm with a relatively high rise of 103 cm (pressure = - 103 cm water). These -'I
J. BOUMA
8
pressure potential h (cm) -1 I
l ’ ’ ’ ’ l ‘ ‘ ’ ’ l ’ ’ ’ ‘ l ‘ ’ ‘ l l ~ ~ i
0
50
100
150
200pm
diameter of tubular pore
Fig. 8. The relationshipbetween tubular pore size and corresponding pressure head.
numbers also imply, of course, that it takes a larger force (more energy) to remove water from a small pore than from a large one. It is physically correct to define such conditions in terms of a negative pressure. However, the minus sign may be inconvenient and then conditions are often described in terms of “tensions” or “suctions.” In other words, a moisture pressure of - 20 cm water is equivalent to a suction or tension of 20 cm water. The negative notation is preferred and w ill generally be used in this article. To represent the porosity of a certain soil material as a bundle of capillaries with a characteristic size range is, of course, unrealistic because real pores in the soil have a much more complex configuration. This representation can nevertheless be helpful for visualizing the energy condition of water in soil and to explain flow phenomena. “Equivalent pore-size distributions” are derived by determining the volume of water between two pressure heads in the soil and by relating the two heads to “equivalent” pore sizes according to Eq. (1) and Fig. 8. Volumes of water between pressure heads are derived from moisture retention curves (earlier called pF curves) which constitute standard soil physical data. An example is taken from a study of Spaans et al. ( 1989)focusing on structure
+
INFLUENCE OF SOIL MACROPOROSITY
9
0 -120
-90
I
-
I I
-
I
t h
-60
-30
0
1
I
I
0
FOREST
\
'.
PASTURE
I
\ \
0.
0.42
0.48
0.54
\ 0.60
0.66
Fig. 9. Moisture retention curves for forest and pasture soils in Humoxic Tropohumult from Costa Rica. (O), Forest; (a),pasture.
degradation in a Humoxic Tropohumult following agricultural use of tropical forest. Moisture retention curves for forest and pasture are shown in Fig. 9. Using Figs. 8 and 9, we conclude, for example, that pores between equivalent sizes of 30 pm and 100 pm occupy volumes Of 4% and 2%, respectively, for the forest and the pasture topsoils. Thin section images ofthe two topsoils are shown in Fig. 10. Quantimet counts were made to ascertain macroporosity and total porosity was determined by standard soil - physical methods. Counts in thin section cannot extend to pores finer than 30 pm because dimensions are distorted by the thickness of the sections. Submicroscopic techniques are available to study fine soil pores, but these will not be discussed here, as emphasis is focused on macropores. The maximum size to be measured in thin section is approximately 5000 pm. In summary, physical measurements provide total pore volume. Micro-
10
J. BOUMA
Fig. 10. Thin-section images of the two surface soils represented in Fig. 9. Pasture (upper image); forest (lower image).
INFLUENCE O F SOIL MACROPOROSITY
11
morphology allows a characterization of pores in the range between 30 pm and 5000 pm. Equivalent pore sizes, discussed earlier in this section, can be determined only for sizes below 150 pm because moisture retention curves cannot be measured reliably at negative pressure heads higher (!) than -20 cm. The height of the sample introduces a high relative error in the calculations. For example, if we take a sample with a height of 2 cm (which is already bound to be unrepresentative for the soil) there is a large error when calculating equivalent pore sizes. Assuming the presence of a hanging water column below the sample, a pressure (h) of - 3 cm in the center of the sample (pore diameter d = 970 pm) corresponds to a pressure of - 4 cm on top (d = 730 pm) and -2 cm at the bottom (d = 1450 pm). Ranges in pressure under these wet conditions and associated water losses from the sample are therefore veiy difficult to “translate” in equivalent pore sizes. For all practical purposes it is not realistic to measure equivalent pore sizes larger than approximately 150 pm (pressure = -20 cm). For a topsoil of a Tropohumult soil, pore volumes determined by the physical and morphological techniques are shown in Fig. 1 1 . The strong decrease in macropores following use as pasture is clear. Macroporescan also be classified in size classes, as is also shown in Fig. I 1. Note that a lower size limit of 30 pm was used here, rather than the more usual 75 pm or 100 pm. The examplejust discussed indicates use ofphysical and morphologicaldata together. Each technique provides specific data that cannot be generated by other techniques. One can also ask the question whether data can be combined. To illustrate this aspect, the work by Bouma ( 1969), who studied intra- and interaggregate porosities in packed aggregates is cited (Fig. 12). Here, a point count in thin sections yielded a porosity of 39%which was composed of packing pores between aggregates. The internal porosity of the aggregates was determined in a separate experiment with the use of the apolar liquid kerosene. The result was 36.6 f 0 . 4 % .The core as a whole had 39%interaggregate pores (point-count) and therefore 100 - 39 = 6lY0 aggregates. In these aggregates, 36.6Yo of the volume was occupied by pores and 63.4%by solids. These values can be modified to apply to the entire 100cm3 cylinder, as follows: 61/100 X 36.6%= 22.3% pores and 61/100 X 63.4%= 38.7%solids. Total pore space within the 100 cm3core would thus be 22.3 39 = 61.3%. For comparison,the porosity was also determined by direct physical measurement, using standard procedures. The result was a porosity of 63.6%. Agreement between the values derived by the two procedures is considered to be quite good, indicating the complementarity of morphological and physical methods.
+
forest
5-15 60
CKI
20-30 cm
-
% 40
-
20
-
0
--
20
-
40
-
% 60
pasture
Fig. 11. Pore volumes obtained by micromorphological and physical techniques for the two surface soils represented in the previous two figures. ( Macroporosity I ] (30- 5000) pm); , porosity (<30 and > 5000 pm).
(m),
Fig. 12. Part of a polished block view into 100 cm3 cores which were fded with small aggregates (1 -2 mm) of a sandy loam soil. The image was used for a point count to obtain interaggregate porosity, to be combined with porosities measured with physical methods.
INFLUENCE OF SOIL MACROPOROSITY
13
111. EFFECTS OF MACROPORES ON SOLUTE MOVEMENT A.
BASICINTRODUCTION TO WATER FLOW BASEDON MORPHOMETRIC POREANALYSIS
Physical equations have been developed for certain types of pores to relate pore sizes to flow rates at a given hydraulic- head gradient. For a cylindrical pore of radius r, we find (Childs, 1969): Q/t =-WPP grad H 8rl For a plane slit of width d, and unit length,
where Q/t = volume of liquid (m3) flowing through the tube per second (m3/s),p = density of water (kg/m3),g = gravitational constant (m/s2),rl= viscosity (N/m), and grad H = hydraulic gradient (m/m). These equations are graphically expressed in Fig. 13, demonstrating the significant effect of pore size on flow rates. For example, these graphs show that a tubular (cylindrical) pore with a diameter of 100pm will conduct about 2 cm3/day at a gradient of 1 cm/cm (or 4 X 2 = 8 cm3/day at a gradient of 4 cm/cm), since v (flux density in cm/day) is equal to K ( hydraulic conductivity in cm/day) at a gradient of 1 cm/cm. A plane slit with a width of l00pm (and unit length) will conduct 700 cm3/day. A plane slit with a length of 4 cm will conduct 8400 cm3/day if the gradient is 3 cm/cm (12 X 700 cm3/day). The determination of hydraulic conductivity would be a very simple matter if soil pores were continuous cylindrical tubes or planar slits with known dimensions. Real soil pores are, of course, very irregular in shape and schematization of soil pores in terms oftubes and plane slits represents a very drastic approach. Nevertheless, such representations may be useful to illustrate flow phenomena and the approach will be used here with that purpose in mind. As an introduction we consider horizontal sections through abstract models of “soil” (Fig. 14). Figure 14A shows a surface of 100 cm2occupied by square blocks (“peds”), mutually separated by a distance d. The blocks are impermeable and water moves only along the plane. Figure 14B shows the same surface with one tubular pore (channel). Water can only move through this pore. The Gtvalues (hydraulic conductivity at saturation) were calculated for different sizes of peds and pores, using Eqs. (2) and (3) (see Fig. 13).
J. BOUMA
14
flow rate (cm3/day) a t grad. t i = 1 cm/cm 1 0 8
10'
10 10 10'
10 3 10 10
1 10 - 1 10 -
I
10 - 3 10.4
10
r
10
I
I
1
100 1000 lOOOOpm pore diameter or width
Fig. 13. Flow rates through tubular or planar voids as a function of pore sue at a hydraulic gradient of 1 m/m.
A
a planar-void structure model
a tubular-void structure model channel
\ lOcm \
I
10 crnJ
L
l
O
c
r
n
-= planar voids Fig. 14. Horizontal sections through a planar-void and a tubular model of soil structure. The groundmassaround planar voids and channels is supposed to be impermeable. For further explanation, see text.
INFLUENCE OF SOIL MACROPOROSITY
15
Table I Hydraulic Conductivities and Porosities of Planar and Tubular Pore Models of Soil Structure (Fig. 13,14) Planar void model Size of blocks 1 cm2
Width (d) of planar voids inpm 10
50 100
4 cm2
10 50 100 1000
Porosity in% 0.2 1.o 2.0 0.1 0.5 1.o 10
Kin cm/day 1.5 180 I440 0.7 90 720 719,712
Tubular pore model Diameter of channel (2r) in pm 100
200 500 lo00 4000
Porosity in%
Kin cm/day
0.8 X 0.3 X 0.2 x 0.8 X 0.13
0.02 0.34 14.6 211.2 54,130.5
10 10 10 10
Porosities were also calculated for these models. Results are reported in Table I. Some obvious conclusions can be drawn: 1. Few small pores can conduct large quantities of water. This is due to the fact that the capacity for transmitting liquid is proportional to the factor r4 for channels ( r = radius) and d3 for planar voids (d = width). Small differences in measured pore size values will therefore have a large effect on the calculated permeability. This is not a very good starting point for any method that tries to calculate K from morphometric pore data, let alone for methods based on “equivalent sizes.” 2. Occurrence of only a few pores may result in high permeabilities, even though their contribution to pore volume is very small. For example, a structure with blocks of 4 cm2 separated from adjoining blocks by 100 pm wide planes, has a high Kof 720 cm/day (Table I). However, they contribute only 1 volume % to porosity, which is well within the experimental error
16
J. BOUMA
when porosities of natural soils (let alone equivalent pore-size distributions) are determined by physical methods. These phenomena are even more pronounced for tubular voids (Table I). Physical methods are therefore inadequate as a tool to characterize size distributions of macropores, such as planes and channels, which are essential to flow in pedal soil.
B. PREDICTING
OF
SOILSWITH MACROPORES
The discussion in the previous section indicated that only a few pores can conduct much water if they are continuous. When studyingwater and solute movement in pedal soils or in soils with continuous macropores, it is necessary to express pore continuity. One techniqueis to use dyes. In our research, we have used Methylene Blue, but other dyes can be used as well. A solution of dyes in water is percolated through the soil. It is important to monitor the color of the solution as it changes during percolation. As dye is being adsorbed by the walls of the pores through which water moves, the color intensity of the water decreases and so does its staining potential. In studying water movement in pedal clay soils (Bouma et al., 1977, 1979), we used saturated blocks of soil and monitored outflow from the block to make sure that the color intensity of the outflow was identical to the intensity of inflow at the time that the staining experiment was stopped. Figure 15 shows a horizontal cross section from a heavy clay soil. The dark bands along the walls of some of the macropores indicate the presence of stained water during flow. The number and the lengths of stained pores were determinedby Quantimet and calculationswere made o f X , based on a pore interaction model, which expressed the effect of “necks” in the flow system (d,, = necks in the planar system and r,, = necks in the channel system in Fig. 16). Neck sizes are obtained by estimating the probability that pores of a given size class are continuous throughout the sample to pores equal to or larger than themselves. The critical probability is assumed to be 5%. Necks are considered in terms of number per unit area for channels (n/s)or length of planes per unit area (l/s). The reader is referred to the research of Bouma et al. ( 1977,1979)for more details. The main conclusions of the study can be summarizedas follows: ( 1) Necks in the planar system were very small and varied between 30 pm and 90 pm. The volume they occupied was low, ranging from 0.12 to 0.40%by volume; (2) Pore and neck sizes were measured in classes of 15 pm. A pore size of 30 pm has, in fact, to be interpreted in terms of 22.5 pm-37.5 pm. The corresponding Gtvalues, calculated with Eq. (3), ranged from 5.4 cm/day to 25 cm/day illustrating the major effect of small variationsin pore width on Kmt. When dealing with macroporous soils it is therefore not
INFLUENCE OF SOIL MACROPOROSITY
17
Fig. 15. Horizontal thin-section image of a freezedried wet heavy clay soil in which water-conducting planar voids were stained with Methylene Blue (dark bands on the walls of planar voids). Stains are discontinuous, illustrating vertical pore continuity.
a
Fig. 16. Schematic representation of tubular and planar “necks” in flow systems which determine the overall flow rates. &, values can be calculated when the number of length of voids per unit surface area S (nchannels or length I, of planar voids) is known [see Eqs. (2) and (3)] (Q, K - planar voids; &, K - channels).
18
J. BOUMA
realistic to talk about very specific K , values, but, rather, in terms of ranges of values; and (3) Gtvalues for six clay soils with macropores could be calculated using the micromorphology data. Of course, the method is cumbersome and direct measurement is preferred. The major contribution of this analysis is the increase in understandingof the basic processes involved during “saturated” flow in clay soils with macropores.
C. BREAKTHROUGH CURVES: DISPLACEMENT OF LIQUID The flux density (v) of water, as defined in soil physics, is a bulkflow velocity which is derived by observing the rate of movement of a free water surface. In other words, if a water surface on top of, say, a soil core moves down at a rate of 1 cm/day, we define v as 1 cm/day. The real flow velocity in the soil pores is higher because usually at least 40% of the soil is composed of solid particles. This is of practical importance for all soils but particularly for the finer textured soils where flow occurs along the ped faces and through other larger voids which may contribute very little to the total pore volume. If the soil were composed of simple capillary tubes of specific sizes, calculations ofthe real flow velocity in those pores would be easy. However, pores vary in shape, width, and direction, and the actual flow velocity in the soil pores is variable. At best, therefore, one can refer to some “average” velocity (v’) that can be calculated on the basis of the water-filled porosity at each tension, v/ = vie
(4)
where 8 is the water-filled porosity (cm3/cm3)as derived from the moisture retention curve. At unit hydraulic gradient, we find:
v/ = K/e
(5)
The average velocity v’ is often still a rather useless characteristic. Flow velocities vary considerably within irregular, natural porous systems. It is much more interesting to know the maximum and the minimum velocities for a given flow regime that occur simultaneously during flow through large and fine pores rather than the “average” velocity. Use of the tracing techniques can provide data that can be used to analyze such heterogeneousflow patterns. The percolatingwater can be replaced by a solution of, for example, CaCl,, and the chloride concentration in the column effluent can be monitored as a function of time to obtain a measure for the range of flow velocities within the sample. Chlorides are used because they are not adsorbed by the soil particles. If the maximum and minimum velocities are both close to the average velocity (for example, in a sand) we will find a relatively long period
INFLUENCE OF SOIL MACROPOROSITY
19
A
lllll
B
11111 C
Dlll CICO
0 8-
0 6-
Breakthrough curves
04
1
7 displacing liquid
=
displaced liquid
02
0
10
2 0 VIVO
Fig. 17. Schematic diagram showing breakthrough curves in three soil materials with different pore-size distributions. Type A corresponds with piston-type flow, Type B with a heterogeneous pore system, and Type C with occurrenceof macroporesin a fine-porous matrix. C/C,represents the relative concentration of the displacing liquid. (See text).
in which water (without chlorides) that was initially present in the column, is replaced (“pushed out”) by water with chlorides. There is a relatively well defined, sharp boundary that moves down the column, separating the two liquids (Fig. 17A). As soon as all the water is replaced, the column effluent will contain chlorides and the effluent concentration will be identical to the
20
J. BOUMA
concentration of the applied chloride solution. This type of displacement is called piston flow. Maximum and minimum velocities may differ considerablyin, for example, fine textured soil materials with channels and planar voids. Then, chlorides may be found very soon in the column effluent because fast movement occurs along the larger voids (such as macropores) during saturated flow, whereas very slow movement is found simultaneously through the fine porous peds (Fig. 17C). The concentrations of chlorides in the column effluent will therefore show quite a different pattern in time as compared with conditions where the maximum and minimum velocities are close together. The first chlorides will appear long before all the untraced initially present water is displaced from the column. The column effluent will reach the chloride concentration of the applied chloride solution after displacement of more than one pore volume (v/vo = 1)because the chloride solution (which flows rapidly through the larger voids) mixes in the column effluent with untraced water, which is slowly displaced from the fine porous aggregates at the same time. There is no well defined boundary moving down the column, separatingthe two liquids. These flow regimes, which are schematically represented in Fig. 17, can sometimes be physically characterized in terms of the apparent dispersion coefficient (D), if breakthrough curves follow an “ideal” pattern. This is not further explored here. It is emphasized that the shape of the breakthrough curve is a measure of macropore continuity. The rate of vertical movement of liquid waste or chemical fertilizer (particularly nitrates) through a soil may be highly underestimated if a hypothetical vertical flow rate is estimated by v’ = K/8, as discussed, where 8 refers to the total water-filled porosity. Few planar voids or channels occupy a much smaller porosity but may conduct large quantities of liquid in a short time. This can be illustrated by considering saturated flow through a hypothetical 30-cm-high soil core with a diameter of 11 cm and a porosity of 50%.Let us assume that the measured flux is 15 cm/day. If water is only shallowly ponded on top of the core, one finds that v = K. The “average” flow velocity in the pores (v’) is l00/50 X 15 = 30 cm/day. This implies that an estimate can be made of the time (“travel time”) needed for a particle of liquid to pass through the soil core. This would, obviously, be 30/30 = 1 day. However, if one assumes that the core consists of a very slowly permeable clay with one continuous channel with a diameter of 500 pm, the channel contributesonly 2 X volume %tototal porosity but accounts fully for the measured permeability of 15 cm/day (see Section 111,A). One can now calculate the real v in the channel as 100/(2 X lo-)) X 15 cm/day = 750,000 cm/day and a travel time of only 3.5 seconds rather than 1 day. This estimate is more realistic. Fortunately, such pores are not always continuous in larger natural soil samples, but still the example illustrates the problems involved, and the importance of macropore continuity.
INFLUENCE OF SOIL MACROPOROSITY
irrr {{rrri K(sa+)
0 water filled
21
irri Klunsati
6
air filled
l o g K (crn/day)
-2
0
-4
1
-8 -12 -16 hydraulic head (crn)
Fig. 18. Diagram illustratingthe effect ofthe Occurrence ofcontinuous macroporeson K,, and on K near saturation (A-D), and on measurement of K,, in attached and detached soil columns (E).
D. SATURATEDAND UNSATURATED FLOW The dominant effect of pore sizes on permeability is evident when comparing K values of a soil material that are measured at different degrees of saturation. Unsaturated soil below an infiltrating surface may have different causes, such as the occurrence of a physical bamer to flow at the surface of infiltration or an application rate which is lower than the saturated hydraulic conductivity. To illustrate the effect of macropores on K,, , a schematic diagram is shown in Fig. 18.A clay soil with relatively fine porous peds which also contains relatively large macropores is represented. The Kmtis relatively
22
J. BOUMA
high because the macropores are vertically continuous and free outflow is possible from the bottom of the macropore (Fig. 18A). When the application rate of water is decreased, there will be a rate at which not all pores can be filled with water. The small pores, which exercise the highest suction, will fill first and water will only flow in the macropore when the applied flux cannot be handled by the fine pores. An interesting condition is found when the small pores are filled while water is running down the walls of the macropore without Wing it entirely (Fig. 18B). Bouma ( 1982)described this condition which still defines Kmtbecause the hydraulic gradient is 1 m/m and the pressure is zero. However, the soil is not saturated. Bouma ( 1982)suggestedthe notation qmt1. When the application rate is reduced to a level where only the micropores are Wed and where a suction is measured, we are dealing with Kmt (Fig. 18C). Thus KMt.in soils with continuous macropores is not really defined (Fig. 18D).The picture is further complicated by macropore continuity. For example, even when Kmt is measured in a very large core, macropores such as worm channels may be vertically continuous. However, when the infiltration rate is measured into a column of soil that was carved out in situ, the worm channels may not be continuous. Even though a flux density at unit hydraulic gradient is measured at zero pressure (definingK,,), very different values are obtained (Fig. 18E) Lauren et al. (1988) reported a range from 0.36 to 43 m/day in a Glossaquic Hapludalf. Thus far, the effect of vertical macropores on Kmtand Kmt was considered. Horizontal shrinkage cracks may, however, also have a major effect on the upward movement of water and solutes. Studies in a Dutch clay soil (Bouma and de Laat, 1981;Bouma, 1984) showed the occurrence of horizontal air-filled cracks which increased in area as the suction increased. A field-stainingtechnique was devised to stain the area of air-iilled horizontal cracks as a function of the pressure head and a Kmt function was defined to calculate upward unsaturated flow, by reducing the Kmt values at corresponding suctions for the soil matrix, taking into account the stained areas at a number of suctions.
E. BYPASSFLOW Preferential movement along large pores during saturated flow was discussed in Section II1,D. However, such preferential movement may also occur in dry or moist soil, and this process is called bypass flow (Bouma, 1984). Bypass flow is defined as vertical movement of free water along continuous macropores through an unsaturated soil matrix. Thus, deep penetration of chemicals may occur in a short time. A schematic diagram is
INFLUENCE OF SOIL MACROPOROSITY
23
Fig. 19. Schematic diagram illustrating bypass flow as a function of rain intensity and duration.,.i Rain intensity;i, , infiltration rate into peds; cf, crack flow;iz, lateral infiltration into peds.
shown in Fig. 19 in which bypass flow is illustrated. Assume the presence of two peds, separated by a macropore which could be a crack or a large channel. When the application rate of water (i) be it rainfall or sprinkling irrigation, is lower than the infiltration rate at the soil surface, there will be no bypass flow. (i, 2 i&. However, as soon as i, becomes smaller than ia, there will be surface ponding ofwater, followed by water running down the walls of the macropore (flow cf in Fig. 19).In clay soilswith well developed peds, one sees that water runs down as small bands. This was demonstrated by sprinkling with water containing Methylene Blue as a dye (Bouma and Dekker, 1978; Bouma et al., 1981; see also Fig. 20). As water flows down the walls of macropores, it will be pulled into the dry or moist soil matrix (flux i2 in Fig. 19). Measurements have shown, however, that lateral adsorptionis relatively low (Hoogmoed and Bouma, 1980),because the area occupied by the vertical bands is very low and lateral infiltration is often reduced by the occurrence of coatings or pressure faces. At relatively high rainfall intensities (e.g., i, in Fig. 19),bypass flow may be almost instantaneous. At very low intensities (e.g., i, in Fig. 19)it will never occur. An intermediary condition is common where bypass flow occurs after a certain period of infiltration when the infiltration rate in the soil decreases to a level below the application rate. Bypass flow is a function of: (1) Soil microrelief- the higher the surface storage, the longer it takes for bypass flow to start; (2) Soil texture-infiltration rates in sandy soils are generally higher than those in clayey soils. Correspondingly, bypass flow is less likely to occur in the more sandy soils even though continuous macropores may be present; (3) Rain quantity and intensity as discussed using Fig. 19; and (4) Water content of the soil surface-dry soil may have water repellant properties which increase bypass flow. However, nonrepellant dry soil has relatively high infiltration rates and, therefore, relatively low bypass
Fig. 20. Experimental set-up to measure bypassflow in the field. Eight measurementscan be made at the same time (see picture) and data are directly available.
INFLUENCE O F SOIL MACROPOROSITY
25
flow. Moist or wet soil is associated with relatively high bypass flow due to relatively low infiltration rates in the matrix (see also Fig. 19, with the infiltration rate i , which decreases in time due to wetting of the soil). When studying water movement in soils in the field, we often find that macropores are not vertically continuous. Then, bypass water will accumulate in the bottom of the macropore constituting,in fact, subsurfaceinfiltration which has been called internal catchment. Van Stiphout et al. (1 987) have documented a field study with a clay soil where cracks were continuous to a depth of 60 cm and worm channels to a depth of 110 cm. Rainfall resulted in wetting patterns at the surface and at both 60 cm and 1 10 cm below the surface.
IV. METHODS TO STUDY MACROPORE FLOW A. MEASURING K AT
OR NEAR
SATURATION
Several methods for measuring k;u and Kmt were reviewed in Mute (1986). To obtain an accurate picture of the effect of desaturation on K in soils with macropores, steady-statemethods are attractive, because transient methods are associated with very rapid changes in a short time that cannot be characterized reliably. Steady-state methods may use the application of steady rainfall or a crust to induce a steady flow in underlying soil. We have obtained good results with the crust-test procedure, using one crust composed of coarse sand and quick-settinghydraulic cement. Different fluxes in the soil and corresponding negative pressure heads can be induced by varying the height of the outflow level of the Mariotte device (Booltink et af., 1991). This procedure is an improvement over the former crust-test procedure, where different crusts had to be applied. Measurement of Ltcan be achieved most easily by sampling large, undisturbed cores and by measuring steady outflow or inflow rates at shallow ponding and unit-gradient flow. This can also be done in situ using encased volumes of soil (Lauren et al., 1988).
B. MEASURING BYPASSFLOW Bypass flow is relatively easy to measure as it consistsof free water running along vertical macropores in an unsaturated soil matrix. The bypass water can be collected without the cumbersome use of suction cups or plates. Surprisingly little effort has so far been made in soil physics to measure
J. BOUMA
6 -
I
standard deviation
4 -
/ YATO’ --OUTFLOW
2 -
0
155
0
50
130
21020
180
21090
21070
21150
21175 21200 time (min)
Fig. 21. In-and outflow graphs for bypass flow as measured in a dry, cracked clay soil. Different slopes of the two lines indicate absorption of water into the soil groundmass.
bypass flow. Bouma et a/. (1981) measured what at the time was called “short-circuiting.” A large undisturbed soil core, including the soil surface, with a diameter and height of 20 cm was subjected to rainfall of different quantities and intensities (Fig. 20). The amount of water, leaving the core as a function of time, was registered and was a measure for the amount of bypass flow. A typical outflow graph is shown in Fig. 2 I. The slope of the line which represents the application rate is slightly steeper than the one which represents outflow. The difference is due to absorption of water by the soil which occurs: (1) at the upper surface where vertical infiltration takes place in soil between the macropores; (2) along the walls of macropores where free water moves downwards. In clay soils with peds, dye studies have indicated that this movement occurs in the form of small vertical bands with a width between 5 and 10 mm (Bouma and Dekker, 1978);and (3) from the bottom of dead-end pores where bypass water accumulates and infiltrates into the surrounding unsaturated soil matrix (Van Stiphout et al., 1987). For relatively short showers, the third process is usually the most important. Edwards et al. ( 1990)used an in situ method to collect bypass flow that followed wormchannels. They reported that infiltration could increase by more than 10 cm/year in watersheds farmed with no-tillage practices as compared to similar watersheds that were conventionally tilled.
INFLUENCE OF SOIL MACROPOROSITY
27
Fig. 22. Breakthroughcurves, using C1as a tracer, for saturated flow in a subangularblocky and a prismatic soil structurein a silt loam soil. Replicate curves are significantlydifferent.(-), Subangular blocky; (--), prismatic.
C. PEDOTRANSFERFUNCTIONS Pedotransferfunctions (Bouma, 1989b) relate different land characteristics with one another or to land qualities. Two types of pedotransferfunctions are distinguished: ( 1) Continuous pedotransferfunctions which relate continuous land characteristics, such as percent clay or organic matter or bulk density, to more complex land characteristics that are difficult to measure, such as moisture retention or hydraulic conductivity. Use is made of mathematical equations often derived by regression analysis; and (2) Class pedotransferfunctions which relate texture classes, or soil horizons to more complex land characteristics or land qualities. Class pedotransferfunctions are particularly useful to relate existing soil survey information, as available in soil profile descriptions,to physical and chemical land characteristicsthat are important for modern work in soil science. An example from Anderson and Bouma (1977a,b) is shown in Fig. 22. Five undisturbed 60-cm-long columns were sampled from the B2t and B3 horizons of different silt - loam soils in Wisconsin. Even though soil textures were identical, soil structures were quite different, as the B2t horizon consisted ofwell developed subangular blocky peds, while the B3 horizon had larger prismatic peds. Cross sections of both structures are shown in Fig. 23. Breakthrough curves for saturated flow, using chloride as a tracer, were significantlydifferent. More rapid
28
J. BOUMA
Fig. 23. Cross sections through the soil structures for which break-through curves were shown in Fig. 22. Subangular blocky, upper image.
initial breakthrough occurred in the B2 material with more, continuous planar voids between the peds. Occurrence of these continuous voids also resulted in tailing of the B2-breakthrough curves due to relatively slow displacement of untraced water from the peds (see Fig. 22). The concept of the class pedotransferfunction is now to associate measured types, in this case breakthrough curves, with the B2t or €33 horizon in a particular soil series that is identified in terms of its areal occurrence on soil maps. Soil series and horizons can be determined quite well and reproduci-
INFLUENCE OF SOIL MACROPOROSITY
29
bly by trained pedologists. Using class pedotransferfunctions allows efficient use of available measurements because little cost and effort is needed, certainly when data are present in databases. Wosten et al. (1990) have shown that use of soil horizons to predict hydraulic characteristics for simulation modeling can be quite successful, as compared with results obtained after directly measuring the necessary hydraulic characteristics. This illustrates the effective use of a class pedotransferfunction.
D. MODELING Different methods have been described so far to characterize water movement along macropores by direct measurement. Use of morphological techniques, including staining, can be helpful in explaining certain phenomena and in suggesting particular physical measurements, as has been demonstrated. However, there is also a need to independently predict flow phenomena as influenced by macropores. Much success has already been achieved by modeling water and solute fluxes in homogeneous soil. Much research is in progress to incorporate the effects of macropores in models for water and solute fluxes (Beven and Germann, 1982; White, 1985; Brusseau and Rao, 1990). Basically, four broad approaches can be distinguished. These are described here in general terms only as reference is made to the original papers.
1 . Stochastic Approach
The spatial variability of soil structure in the field is so large and complex that some researchers have concluded that mechanistic modeling is not feasible. Jury ( 1982)has suggested the use of a transfer function based on the probability density function of solute travel time through soil as proposed by Raats (1978). Dispersion of the applied solute is attributed to velocity variations in the soil. When the mean and standard deviation of the measured velocity distribution have been determined for a certain depth, the distribution of solute at greater depths and longer times can be accurately predicted, provided there is no marked change in the physical properties of the soil with depth (Jury et al., 1982). This approach, though attractive in principle, is site-specific, and extensive measurements must be made at any new site. A field example of this approach was recently published by Hornberger et al. ( 1990).
30
J. BOUMA 2. Deterministic: Schematized Porosity
Several models have been derived that consider a soil that consists of “mobile” water in macropores and “immobile” water in the soil matrix (Genuchten and Wierenga, 1976; Addiscott, 1977). The models describe transport through the macropores and diffusion between stagnant and mobile water. Clearly, all immobile water is not completely immobile. Steenhuis et al. ( 1990)describe a model which is essentially based on the equivalent pore-size model, as illustrated in Fig. 18. Water and solutes move with distinct velocities in certain pore groups. Complete mixing between the different pore groups can be specified, which reduces the model to traditional models which are based on the convective-dispersive equation. If no mixing is specified, the model behaves as a pore-bundle model as illustrated in Fig. 18. A crucial question centers on the problem of how to define the degree of mixing in different soils, which is clearly flux dependent. Another approach which fits in this broad category of models was described by Van Bronswijk ( 1988),who characterized water movement in swelling clay soils. Swelling characteristics are measured and water movement into cracks is simulated. Occurrence of cracks is, however, not based on direct observations but is derived from the swelling characteristics.
3. Deterministic: Morphometric Data
Descriptions of macropores in terms of type and number per unit crosssectional area can be used to define boundary conditions for macropore flow systems. Staining techniques are important for determining vertical and horizontal macropore continuity, which plays a major role in flow processes. Existing models are used to define the subprocesses of vertical infiltration at the soil surface, lateral infiltration from the macropores into the matrix, and internal catchment from discontinuous macropores. These processes are schematically shown in Fig. 24. Examples of this research approach have been summarized by Bouma (1989a, 1990).
4. Deterministic- Stochastic
Considering the large variability of macropore patterns in soil, it is difficult to see how a purely deterministic model for macropore flow can be derived which adequately expresses variability aspects. One new and intriguing ap-
INFLUENCE OF SOIL MACROPOROSITY
MODELING BYPASS FLOW
31
(c)
SUBMODELS :
(Six Factors)
1
SURFACE INFILTRATION (S)
2
LATERAL INFILTRATION ( L ) (Dye Tracing)
3
INTERNAL CATCHMENT (IC) (Dye Tracing)
4
OUTFLOW FROM SOLUM (0)
m:
a
*c 3L<=
. ... . . . .. ... ... .. .. . . .. . ......-. . . .....-. .. ... ~
Water table
c 0
No water table-
Fig. 24. Schematic diagram illustrating processes that occur during bypass flow.
proach is being developed at Cornell University. Using soil survey infonnation, simulation models for water and solute transport are being determined for observed soil profiles for which characteristic basic physical and chemical parameters have been estimated in major horizons. Thicknesses of soil honzons and the different parameters themselves are not only used in terms of average values, but their standard deviations, are also included. Monte Carlo techniques are used to draw at random samples from the very large number of possible calculations which are realized when ranges of parameter values are considered, rather than average values only (Petach and Wagenet, 1989). This approach could be extended to macropores, to be observed with morphometric techniques, when they are also defined stochastically in terms of number per unit area and depth of occurrence. At this moment it is not clear which method will turn out to be the best for practical field applications. Researchers would be well advised to continue their work at different levels of sophistication. Field applications in the context of agronomic studies require relatively simple procedures with which nonsoil scientists can iden-
32
J. BOUMA
tify. Here, the deterministic- morphometric approach could be particularly useful.
V. FIELD STUDIES ON MACROPORE FLOW AND ENVIRONMENTAL QUALITY
A. INTRODUCTION Many field studies have been reported in which macropore flow resulted in rapid movement of water and solutes from the soil surface to subsurface soil horizons or to groundwater aquifers.Less emphasishas so far been given to the development of methods that can be used to avoid macropore flow, which is generally unfavorable because water and nutrients are removed from surface soil, where most roots are generallypresent, to the subsoil where water and nutrients are beyond reach of the roots. Moreover, formation of horizontal cracks in clay soil will strongly inhibit upward unsaturated flow, which is a major sourceofwater and nutrient supply to crops in sandy or silty soils. Macropore flow often results in groundwater pollution in soils with shallow aquifers because a major part of the adsorptive capacity of the macroporous soil is not used. In Section IV,D,2, “mobile” and “immobile” water were discussed. Correspondingly, one should also discuss the “active” and “inactive” cation exchange of a soil, which is governed by flow patterns. The exchange capacity, as defined in soil survey reports, is a maximum value for a given soil material obtained by making analyses of finely ground soil samples. Whatever value applies for any given time in a macroporous soil is a function of the flow system at that moment, as discussed in this chapter and elsewhere. Occurrence of macropores can also have a major impact on rooting patterns and the associated extraction of water and solutes. Many simulation models for water and nutrient movement play little attention to solute extraction by plants. At best, an “effective”rootingdepth is assumed. Within this depth water and nutrients are implicitly assumed to be within diffusion distance of the roots. Only the concept of “water availability” has been defined in terms of pressure heads and corresponding water contents between 0.1 or 0.3 bars and 15 bars. However, if roots follow only macropores in the soil (which they often do), there may be substantialdistances between roots in the soil, with the effect that water in soil between roots cannot diffuse into the roots. Thus we should also consider the “accessibility” aspect, in juxtaposition to the “availability” aspect. Also, soil morphological analyses can help to specify the accessible soil volume (Bouma, 1990).
INFLUENCE OF SOIL MACROPOROSITY
33
B. SOMEENVIRONMENTAL EFFECTSOF MACROPORE FLOW CASESTUDIES Several studies have used tracers to demonstrate deeper penetration of solutes in soils with macropores than would have occurred under ideal conditions of miscible displacement (Blake et al., 1973; Thomas et al., 1978). Shaffer et al. (1979) recorded increases of Cd2+ in drainage water sampled from macropores at a depth of 1.2 m, while Cd2+in the matrix of the soil was very low. Kanchanasut et al. ( 1978) reported a 1 0-fold increase in P concentration in tile drain flow from a soil where rainfall occurred immediately after the application of superphosphate. Similar results for nitrate were obtained by Dekker and Bouma (1 984). They reported on a practice used by farmers whereby nitrogen fertilizers were applied to dry, clay soils and then sprinkler irrigation was initiated. About 50% of the N and P were lost by bypass flow. An alternative approach which started with briefwetting of the soil, followed by fertilization and sprinkling with lower intensities and quantities reduced losses strongly. They pointed out that “gaps” in the N balance are often attributed to denitrification, while bypass flow may sometimes be a more important process. Thomas and Phillips (1979) and others have suggested that nitrate within aggregatesis protected from leachingwhen bypass flow occurs. Barraclough et al. ( 1983) also observed rapid movement of nitrate along macropores. Edwards et al. (1 990) reported on the pronounced effects of wormchannels in no-till corn fields as compared with fields under traditional tillage which did not contain channels. Luxmoore et al. ( 1990) described pronounced macropore flow in a forested hillslope. Other studies also indicate pronounced effects of macropores on flow patterns under forested conditions which generally have a relatively loose soil structure, due to active biologtcal processes.
VI. CONCLUSIONS A number of conclusions can be formulated about the influence of macroporosity on environmental quality. These are:
1 . Numerous field studies have demonstrated that the occurrence of macropores in soils results in rapid, downward movement of solutes which may lead to groundwater pollution. Some of these studies were made more than one hundred years ago and several reviews are available in which these studies are summarized. 2. Efforts to describe preferential movement of solutes along macropores
34
J. BOUMA
in quantitative terms have not yet resulted in a comprehensive theoretical framework that allows for the independent prediction of such phenomena in different soils. In fact, the process of bypass flow (movement of free water through macropores in an unsaturated soil groundmass) is still not being considered in soil physics textbooks and manuals. 3. In this article a suggestion is made for developing prediction procedures at different levels of sophistication, including procedures that allow field assessment by trained agronomists. Use of relatively simple soil morphological techniques that describe macropores in terms of type, size, number of occurrence, and vertical continuity, can be useful in predicting infiltration patterns of solutes in quantitative terms. In addition, it is necessary to know the hydraulic characteristics of the soil groundmass and other determining factors for bypass flow such as the microrelief, rain duration and intensity, soil wettability, and the actual water content. Use of a relatively simple simulation model which requires both morphological and physical basic data was demonstrated. 4. Prediction of the dynamic character of swelling soils with macropores or of soils with a high biological activity which results in continuous formation and destruction of macropores is extremely difficult. Rather than focus on this type of prediction, it may be advisable first to focus on the quantitative prediction of water and solute behavior in certain static well defined soil structural conditions, as discussed under point 3. Each condition should be well defined in terms of the associated soil processes which are bound to be a function of soil management. A dynamic characterization may result from combining a series of static conditions by introducing the time factor. 5. The implications of preferential flow of solutes along macropores for environmental quality are extremely important. Some standard soil physical and chemical concepts need modification. Concepts of “immobile” or “dead” water, that does not significantly participate in flow processes, have already been used. However, the same principle should apply to the capacity of soils to adsorb cations. In addition to the well established concept of “available” water for plants, we should also add the aspect of “accessibility.”
REFERENCES Addiscott, T. M. 1977. A simplecomputer model for leachingin structuredsoils. J. SoilSci. 28, 554-563. Anderson, J. L., and Bouma, J. 1977a. Water movement through Pedal soils: I. Saturated flow. Soil Sci. SOC.Am. J. 41,413-418. Anderson, J. L.,and Bouma, J. 1977b. Water movement through Pedal soils: I. Unsaturated flow.Soil Sci. SOC.Am. J. 41.419-423.
INFLUENCE OF SOIL MACROPOROSITY
35
Barraclough, D., Hyden, M. J., and Davies, P. 1983. Fate of fertilizer nitrogen applied to grassland. I. Field leaching results. J. Soils Sci. 34,483-497. Beven, K., and Germann, P. 1982. Macroporesand water flow in soils. Water Resour. Res. 18, I 3 1 I - 1325. Blake, G., Schlichting, E., and Zimmermann, U. 1973. Water recharge in a soil with shrinkage cracks. Soil Sci. SOC.Am. 37,669-672. Booltink, H . W. G., Bouma, J., and Gimenez, D. 1991. A suction crust infiltrometer for measuring hydraulic conductivity of unsaturated soil near saturation. Soil Sci. SOC.Am. J. 55,566-568. Bouma, J. 1969. “Microstructure and Stability of Two Sandy Loam Soils with Different Soil Management,” Agric. Res. Rep. No. 724. PUDOC, Wageningen, Netherlands. Bouma, J. 198 1 . Comment on “Micro, meso and macroporosity of soil.” Soil Sci. Soc. Am. J. 45, 1244- 1245. Bouma, J. 1982. Measuringthe hydraulic conductivity of soil horizonswith continuous macropores. SoilSci. SOC.Am. J. 46,438-441. Bouma, J. 1984. Using soil morphology to develop measurement methods and simulation techniques for water movement in heavy clay soils. ISSS Symp. Water Solute Movement Heavy Clay Soils, ILRI, Wageningen,Neth. (J. Bouma and P.A.C. Raats, eds.), pp. 298-
315. Bouma, J. 1989a. Using soil survey data for quantitative land evaluation. Adv. Soil Sci. 9, 117- 123. Bouma, J. 1989b. Land qualities in space and time. In “Land Qualitiesin Space and Time.” (J. Bouma and A. K. Bregt, eds.), pp. 3- 13. PUDOC, Wageningen, Netherlands. Bouma, J. 1990. Using morphometric expressions for macropores to improve soil physical analyses of field soils. Geoderma 46,3 - 13. Bouma, J., and Dekker, L. W. 1978.A case study on infiltration into dry clay soil 1. Morphological observations. Geoderma 20,27 -40. Bouma, J., and de Laat, P. J. M. 1981. Estimation of the moisture supply capacity of some swelling clay soils in The Netherlands. J. Hydrol. 49,247-259. Bouma, J., Jongerius, A., Boersma, O., Jager, A., and Schoonderbeek,D. 1977.The function of different types of macropores during saturated flow through four swelling horizons. Soil Sci. SOC.Am. J. 41,945-950. Bouma, J., Jongerius, A., and Schoonderbeek, D. 1979. Calculation of saturated hydraulic conductivity of some pedal clay soils using micromorphometric data. Soil Sci. SOC.Am. J . 43,26 1 -264. Bouma, J., Dekker, W., and Muilwijk, C. J. 198 1 . A field method for measuringshort-circuiting in clay soils. J. Hydrol. 52, 347-354. Bouma, J., Belmans, C. F. M., and Dekker, L. W. 1982. Water infiltration and redistribution in a silt loam subsoil with vertical worm channels. Soil Sci. SOC.Am. J. 46, 9 17 - 92 1. Brewer, R. 1964. “Fabric and Mineral Analysis of Soils.” Wiley, New York. Brusseau, M. L., and Rao, P. S. C. 1990. Modeling solute transport in structured soils: a review. Geoderma 46, 169- 193. Childs, E. C. 1969. “An Introduction to the Physical Basis of Soil Water Phenomena.” Wiley, New York. Dekker, L. W., and Bouma, J. 1984. Nitrogen leaching during sprinkler irrigation of a Dutch clay soil. Agric. Water Manage. 8, 37 -41. Edwards, W. M., Shipitdo, M. J., Owens, L. B., and Norton, L. D. 1990. Effect of lumbrucus terristris L. Burrows on hydrology of continuous no-till corn fields. Geoderma 46,73 -85. Fitzpatrick,E. A., Mackie, L. A., and Mullins, C. E. 1985.The use ofplaster ofParis in the study of soil structure. Soil UseManage. 1,70-72.
36
J. BOUMA
Hoogmoed, W. B., and Bouma, J. 1980. A simulation model for predicting infiltration into cracked clay soil. Soil Sci. SOC.Am. J. 44,458 -46 1. Hornberger, G. M., Beven, K. J., and Germann, P. F. 1990. Inferencesabout solute transport in macroporous forest soils from time series models. Geoderma 46,249 - 263. Jury, W. A. 1982. Simulation of solute transport using a transfer function model WaterResour. Res. 18, 363-368. Jury, W. A., Stolzy, L. H., and Shouse, P. 1982. A field test of transfer function model for predicting solute transport. Water Resour. Res. 18,369- 375. Kanchanasut, P., Scotter, D. R., and Tillman, R. W. 1978. Preferential solute movement through larger soil voids. 11. Experiments with saturated soil.Aust. J. Soil Res. 16,269276.
Klute, A,, ed. 1986. “Methods of Soil Analysis. Part 1: Physical and Mineralogical Methods,” 2nd ed. Agric. Ser. Monogr., Soil Sci. SOC.Am., Madison, Wisconsin. Lauren, J. G., Wagenet, R. J., Bouma, J., and Wdsten, J. H. M. 1988. Variability of saturated hydraulic conductivity in a GlossaquicHapludalf with macropores. Soil ScI. 145,20-28. Luxmoore, R. J., Jardine, P. M., Wilson, G. V., Jones, J. R., andZelazny, L. W. 1990. Physical and chemical controls of preferred path flow through a forested hillslope. Geoderma 46, 139-155.
Mackie, L. A. 1987. Production of three dimensional representations of soil macropores with a microcomputer. Geoderma 40,275-280. Murphy, C. P., Bullock, P., and Turner, R. H. 1977a. The measurement andcharacterization of voids in soil thin sections by imagine analyzer: I. Principlesand techniques. J. SoilSci. 28, 498 - 508.
Murphy, C. P., Bullock, P., and Biswell, K. J. 1977b. The measurement and characterization of voids in soil thin sections by image analyzer: 11. Applications. J. Soil.Sci. 28, 509 - 5 18. Petach, M., and Wagenet, R. J. 1989. Integration and analysing spatially variable soil properties for land evaluation. In “Land Qualities in Space and Time” (J. Bouma and A. K. Bregt, eds.), pp. 145- 155. PUDOC, Wageningen, Netherlands. Raats, P. A. C. 1978. Convective transport of solutes by steady flows. I. General theory. Agric. Water Manage. 1,201 -218. Ringrose-Voase, A. J., and Bullock, P. 1984. The automatic recognition and measurement of soil pore types by image analysis and computer programs. J. Soil Sci. 35,673 - 684. Shaffer, K. A., Fritton, D. D., and Baker, D. E. 1979. Drainage water sampling in a wet, dual-pore soil system. J. Environ. Qual. 8, 241 -246. Spaans, E. J. A., Baltissen, G. A. M., Bouma, J., Miedema, R., Lans, A. E. L. and Schoonderbeek, D. 1989. Changes is physical properties of young and old volcanic surface soils in Costa Rica after clearing of tropical rain forest. Hydrol. Processes 3,383- 392. Steenhuis,T. S., Parlange, J. Y., and Andreini, M. S. 1990. A numerical model for preferential solute movement in structured soils. Geoderma 46, 193 - 209. Thomas, E. W., and Phillips, R. E. 1979. Consequencesofwater movement in macropores. J. Environ. Qual. 8, 149 - 152. Thomas, E. W., Phillips, R. E., and Quisenbeny, V. L. 1978. Characterization of water displacement in soils using simple chromatographic theory. J. Soil Sci. 29,32-37. Van Bronswijk,J. J. B. 1988. Modelling ofwater balance, cracking and subsidenceof clay soils. J. Hydrol. 91, 199 - 2 12. Van der Plas, L., and Tobi, A. C. 1965. A chart for judging the reliability of point counting results. Am. J. Sci. 263.87-90. Van Genuchten, M. T., and Wierenga, P. J. 1976. Mass tranfer in sorbing porous media. I. Analytical solutions. SoilSci. SOC.Am. Proc. 40,473-480.
INFLUENCE OF SOIL MACROPOROSITY
37
Van Genuchten, M. T., Rolston, D. E., and German, P. F. 1990.Transport ofwater and solutes in macropores. Geoderma, Spec. Issue 46, 1 -291. Van Stiphout, T. P. J., van Lanen, H. A. J., Boerma, 0.H., and Bouma, J. 1987. The effect of bypass flow and internal catchment of rain on the water regime in a clay loam grassland soil. J. Hydrol. 95, 1 - 1 I . Warner, G. S., Nieber, J. L., Moore, I. D., and Geise, R. A. 1989. Characterizingmacropores in soils by computed tomography. SoilSci. SOC.Am. J. 53,653-660. White, R. E. 1985. The influence of macropores on the transport of dissolved and suspended matter through soil. Adv. Soil Sci. 3, 95 - I2 I . Wosten, J. H. M., Schuren, C. H. J. E., Bouma, J., and Stein, A. 1990. Comparing four methods to generate soil hydraulic functions in terms of their effecton simulated soil water budgets. Soil Sci. SOC.Am. J. 54. 821-832.
This Page Intentionally Left Blank
DNA MARKERS IN PLANTIMPROVEMENT Andrew H. Paterson,’ Steven D. Tanksley,2 and Mark E. SorrellsZ ‘Texas A&M University Department of Soil & Crop Sciences College Station, Texas 77843 and Department of Plant and Soil Sciences University of Delaware Newark, Delaware 1971 I ’Department of Plant Breeding and Biometry Cornell University Ithaca, New York 14853
I. Introduction: Agricultural Genetics and DNA Markers 11. How and Why are Genetic Maps Made? A. Classes of Genetic Markers B. Identifying Linkages between Genetic Markers 111. Using Genetic Markers to Study and Improve Agricultural Productivity A. Linkage between Genetic Markers and Genes Influencing Simply Inherited Traits B. Linkage between Genetic Markers and Genes Influencing Quantitative Traits IV. Describing Individual Quantitative Trait Loci A. Chromosomal Location B. Effects of Gene Dosage (or Gene “Action”) C. Multiple Effects of Single Genes (Pleiotropy) D. Environmental Effects V. Cumulative Effects of Many Quantitative Trait Loci o n Phenotype of an Individual A. Hybrid Vigor (Heterosis) B. Progeny Superior to Both Parents (Transgression) C. Interactions between Genes (Epistasis) VI. Improving the Efficiency of Breeding Programs A. Identifying Genetic Variation B. Recombination and Selection VII. Marker-Facilitated Study of Complex Populations VIII. Marker-Facilitated Study of Polyploids IX. Utilization of Exotic Germplasm in Crop Improvement X. Revealing Evolutionary Relationships among Crop Species and Their Wild Relatives -Utility in Comparative Genetic Mapping XI. Cloning Genes from Map Position? XII. Challenges for the Future XIII. Conclusion References 39 Adwnrrr in Agmnomy, Val. 46 Copyrighr 0 1991 by Academic Press, lnc. All rights of reprnducuon in any form reserved.
40
ANDREW H. PATERSON e t
a/.
1. INTRODUCTION: AGRICULTURAL GENETICS AND DNA MARKERS The productivity of domestic crop plants has evolved through the collective efforts of agronomists since the dawn of agriculture, and represents one of mankind’s greatest achievements. Genetics, the study of resemblance between relatives, has made a major contribution to improved agricultural productivity. On a day-to-day basis, different crop varieties or animal strains are chosen for particular markets or particular environments. From a historical perspective, improved crop yield has been influenced perhaps more by genetic improvement than by any other single factor (Fehr, 1984). Despite the breeding progress already achieved, additional gains in agricultural productivity are demanded at an ever-faster pace, by population growth, and by changes in agricultural practices, biotic and abiotic environments, and consumer preferences. Technological advances have provided contemporary plant breeders with efficient field machinery, high-precision laboratory instruments, and rapid data collection and computation systems. However, the tools available for directed genetic manipulation of crop plants have changed much less rapidly, until recently. In recent years, the burgeoning field of molecular biology has provided tools suitable for rapid and detailed genetic analysis of higher organisms, including agricultural species. Perhaps the most fundamental of these tools are DNA markers, simply detected differences in the genetic information carried by two or more individuals. Information from DNA markers serves many diverse purposes, such as forensic science, paternity testing, identifying genes responsible for genetic diseases, and inferring evolutionary relationships among organisms. Perhaps the most widespread application of DNA markers is in the construction of genetic maps, which can be used to determine the chromosomal location of genes affecting either simple or complex traits (Paterson et al., 1988; Lander and Botstein, 1989). In regard to making and using genetic maps, DNA markers are not fundamentally different from other types of genetic markers which have been used since the early 1900s-but they can be found in far greater numbers. With such a large number of genetic markers, one can build a complete genetic map, which is informative about all regions of all the chromosomes in an organism. By knowing the “map position” of a gene, one can use nearby DNA markers to diagnose the presence of the gene, without having to wait for the gene’s effects to be seen. In human genetics, for example, adult-onset genetic diseases can be diagnosed (or at least predicted) in children -permitting
DNA MARKERS IN PLANT IMPROVEMENT
41
early treatment. Camers of a genetic disease, who will never suffer the disease but might transmit it to progeny, can be identified, permitting informed decisions to be made regarding childbearing. In agriculture, DNA markers might be used to predict future attributes of a tomato seedling, a pine sapling, or a newborn calf. Informed decisions might then be made promptly, regarding which individuals should be retained as elite breeding stock, which are satisfactory for agricultural production, and which might best be culled. In this chapter, we will address applications of DNA markers to plant breeding. Much of what we will discuss is also relevant to animal breeding. A review of related research is included, but our primary purpose will be to discuss the new information which can be obtained using DNA markers, and how that information might expedite improvement of agricultural productivity. To date, breeding efforts using DNA markers have emphasized diploid species and inbreeding methods; in addition to discussing these experiments, we will offer some ideas regarding situations which are genetically more complex. The use of DNA markers in extracting agriculturally valuable traits from exotic germplasm will be discussed at length, as we feel this important area of research is especially amenable to marker-facilitated breeding. Finally, we will briefly consider emerging technologies which may further expand the repertoire of molecular tools available to the plant breeder. In the interest of realism, we remind the reader here, and again in the conclusion, that present-day technology associated with DNA markers is tedious, laborious, and expensive, and that DNA markers are not an omnipotent solution to the limitations of classical breeding. Classical plant and animal breeding make an indispensible contribution to society, and will continue to do so for the forseeable future. However, we feel that DNA markers can significantly accelerate many breeding endeavors. Further, DNA markers may provide new approaches to some objectives which have proven difficult to achieve with classical techniques, such as introgression of valuable traits from exotic germplasm into domestic cultivars. In order that the nongeneticist might see how these new tools and techniques have evolved, and also see the value of the new information they provide, we have attempted to provide some background information which will no doubt be elementary to the geneticist or breeder. We hope that our colleagues will forgive us the occasional venture into background material, in the interest of communicating the significance of contemporary developments in plant genetics to a diverse audience. In particular, the breederjgeneticist may well find Section I1 to be elementary, and may wish to proceed to Section 111.
42
ANDREW H. PATERSON et al.
11. HOW AND W H Y ARE GENETIC MAPS MADE? To make a genetic map, one takes advantage of the fact that the genetic information of an organism (DNA) is passed from parent to progeny as long strings, which are packaged together with proteins to form units called chromosomes. Most higher organisms have several to many different chromosomes, with two copies of each,' one from each parent. Different chromosomes carry different repertoires of genes, while the two copies of a chromosome carry the same repertoire of genes, in the same linear order. However, differences within genes carried on each copy, called alleles, can cause individuals to appear different from one another. Such differences are transmitted from an individual to its progeny, and serve as landmarks, or genetic markers, useful in determining which parent was the source of a particular chromosome segment in a progeny individual.
A. CLASSES OF GENETIC MARKERS Differencesbetween the chromosomes of two individuals can be detected in several ways-some by a casual glance at the individuals (visible markers), others by an assay of enzymes (isozymes/allozymes) from body tissues, and still others by analysis of the DNA itself (DNA markers). To be useful as a genetic marker, a trait must meet two criteria: (1) the trait must differentiate between the parents: and (2) the trait must be accurately reproduced in the progeny. For example, blood type is often different between two parents, and the parental blood types are accurately reproduced in the progeny, so blood type might be used as a genetic marker. In contrast, height is often different between two parents, but the parental heights are usually not accurately reproduced in the progeny, so height is generally not useful as a genetic marker (except in extreme cases, such as genetic dwarfism). Genetic markers were being used in biology well before it was known that DNA was the hereditary material. Visible markers, mutations in genes with visible consequences such as dwarfism or eye color, have been used in genetic studies since early in the Twentieth Century (Morgan, 191 1). Markert and Moller (1959) showed that genetic differences in easily assayed enzymes (isozymes)changed their rate of migration through a gel matrix, in response to an electrical field. Using visible markers together with isozymes, substantial genetic maps of some organisms had already been assembled by the late 1970s.
' Except the sex-determiningchromosomes(X and Y) in animals. Most plants do not have sex chromosomes,however many plants are exceptionsto the rule of two copies per chromosome in that they are polyploid,with more than two copiesofeach chromosomewhich leads to very complex genetics (see Section VIII).
DNA MARKERS IN PLANT IMPROVEMENT
43
In 1980, it was suggested that large numbers of genetic markers might be found by studying differences in the hereditary DNA molecule itself, revealed as restriction fragment length polymorphisms (Fig. 1) (Botstein et al., 1980). Restriction enzymes2are highly specific “molecular shears,” which cleave the DNA at particular sequences (restriction sites). If two individuals differ by as little as a single nucleotide (a “letter” in the genetic code) in the restriction site, the restriction enzyme will cut the DNA of one but not the other, ’generatingrestriction fragments of different lengths, which can then be separated (in an electrical field) and visualized (by specific binding of a radioactive probe). In principle, visible markers, and isozymes are as useful as DNA markers; however, in practice much greater numbers of DNA markers can be readily found. Crop plants have about 1O8 - 1O ’ O nucleotides of DNA in total. If even a small fraction of these are different between two individuals, an enormous . ~contrast, relatively few visible number of potential DNA markers r e ~ u l tIn markers or isozymes tend to be different between two randomly chosen individuals. In order to have a sufficient number of visible markers for a particular experiment, they must first be collected, through breeding, into a multiply marked stock. Isozymes suffer the same limitation, and are further limited by the number of biochemical assays known to detect enzyme activities (perhaps 100; Tanksley, 1983b).
B. IDENTIFYINGLINKAGES BETWEEN GENETIC MARKERS A genetic map represents the relative order of genetic markers, and their relative distances from one another, along each chromosome of an organism. During sexual reproduction in higher organisms, the two copies of each chromosome pair, aligning themselves closely with one another. Occasionally, the two strings comprising the DNA of each chromosome can break, and rejoin with opposite partners, producing a reciprocal exchange or recombination. The new recombinant chromosome thus is a mosaic of the two parental chromosomes. Genetic markers which lie close to one another on the chromosome are seldom recombined, and thus are usually found together in the same progeny individuals. In contrast, markers which are
* Restriction enzymesare synthesized by bacteria and other simple organisms, and appearto restrict the ability of viruses to parasitize the bacteria, by degrading the DNA of said viruses. Note that the technique of restriction fragment length polymorphism cannot detect all differencesbetween nucleotides-just those differences which happen to occur in a restriction site, or differences which change the distance between restriction sites. New techniques for identifymg DNA markers, using special separation methods (Fischer and Lerman, 1983; Kornher and Livak, 1989), more sensitive DNA probes (Jeffreys ef a/., 1985; Ali ef al., 1986; Vergnaud, 1989; Tautz, 1989), or simply larger numbers of candidates (Williams et al.. 1991) may partly alleviate this limitation in the future.
44
ANDREW H. PATERSON et a/.
0 2) Cleave DNA with restriction enzyme
Fig. 1. Detection of DNA markers by the method of restriction fragment length polymorphism. DNA is extracted from two (or more) different individuals, then is digested by a restriction enzyme, which cleaves the DNA in or near a specific recognition sequence (or restriction site). In the example, the recognition sequenceofthe commonly used restriction enzyme EcoRI is presented (5'-GAATTC). The two individuals differ in DNA sequence at one potential recognition sequence. The restriction enzyme cuts in one but not the other (indicated by X), generating restriction fragments of different length. The fragments are separated by gel electrophoresis, and visualizedby binding of a specific radioactive DNA probe. In practice, restriction fragment length polymorphisms are found empirically, by randomly testing different DNA probes with different restriction enzymes, until a combination is found which distinguishes between the genotypes of interest.
distant from one another, or on different chromosomes, are frequently recombined, often occurring in different combinations in the progeny than they did in the parents (Fig. 2). One estimates the relative distance between two markers by determining the percent recombination, that is, the percent of progeny which received the two markers from different parents. Markers which lie close together show a small percent recombination, and are said to be linked. By studying many markers, one eventually finds enough linked pairs to build up a genetic map, including many markers lying short distances apart from one another along each chromosome (Fig. 3). Genetic
Segregating F2 ( s e l f e d or intercrossed) progeny 1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20
Visible
markei (seed color)
00 c
Isozyme (dirner)
0
c 0-
RFLP
- I= B
m 0 0 5
1
0 0
0 0
- 3 6 1
0
Markers are randomly a s s o c i a t e d i n progeny; hence not l i n k e d
Fig. 2. Fundamentals of genetic mapping. (A) A visible marker (seed color), an isozyme (allozyme) marker, and a restriction fragment length polymorphism (RFLP) marker have been scored on two inbred parents, their F, hybrid, and 20 F, self-pollinatedprogeny ofthe F, hybrid. The visible marker, isozyme, and RFLP are each useful as genetic markers, because (1) the traits differ between the parents; and (2) the progeny fall into discrete classes regarding the traits. (B) Of the three possible pairwise comparisons between the markers (3 X 3 tables: seed color X isozyme, seed color X RFLP, isozyme X RFLP), only seed color and the RFLP show an association. Eighteen of the 20 F, progeny show the same combination of seed color and RFLP genotype exhibited in the parents or F,, with only two being different (recombinant). The two observed recombinants (Nos. 12 and 17 in Fig. 20) can each be explained by recombination along just one of the two informative chromosomes in each of these individuals, therefore recombination has occurred between the seed color gene and RFLP in 2/40 (5%) of the cases examined. This indicates that the seed color gene and the RFLP lie on the same chromosome, and are separated by a distance of only 5% recombination, or about five map units.
92C9l IS621
96
4 2'11 Z'C ('9
9'21
(zc I -l.l
L'C 1'2 2'2 0'2
1.1 0'1
9'9 6'2
9'b
1'5
I'C
9'0
1'1
9'1
C'tl b'Z b'2 b'l 1'1
P'C
9'1
5'0 9'21
11-
1'91
IIC'LII
1'2
9'2 f'b
9'21
C'b 5' I S'5
C'C 9'1 9' I
6
0
L
c
I
5
4
6
16211
16116
112.711
11.0
1 1.81) 1 1.21) 1 7.31)
1.8 3.0 1,2 7,l
(b.111
b.1
1 9.211
9.1
1 2.91)
2.9
I 0.81) lb.11) I 9.711
0.8 6.5 9.8
I 1.011
1P12 16170
TGl82 lt325
115,111
15.1
-11
COll
CDb 7 lG97 16232
16287
76208 1628+1 1662 16127
111.41)
11.1
I665
16352 16151
-11
1625
Ttlb1 I6217
TGISS
10366 1GW1
11S.lII 15.8
16233
It371 11.b11
I6361
1.7
16115 16292 IGlb2
16113
116.31)
1190 16218
16.9
16198
1118) 76185
1G273
16221
76220 16215 I6122
lwb)
10120
1 2.811 12.IIl I 2.211 1 9.611
2,8
123 I ) 16285
2.1 2.2 9.7
1
111.11)
11.B
( 9.91)
10.0
c.n rrec.
01.1
nori.r
m
Id
111.11)
11.5
IE13W 121, I l l
-
22.5
1
1101) IGI18
111Sl I6230
12521 I629b
16191
19) i G I O 3 571 1652
12s.s11
28.1
I 1.81) 1 2.811
4.8 2.8 2.8 2.8
121.21)
11.51)
I 2.81) I 2.811 111.91)
1 I91 CO2 1217) CHS-5
12351 16280
1259) I6301
130.01) 31.b
12
11
10
I 1.311 I 0.81) 1 5.81) 1 I.3lI 1 2.01)
1q.2
25.1 I .5 1.1 0.8
5,9 1,3 2.0
1 1.11) L 1 1IO.II) 10.2 119.811
115.11) 15.8
20.9
I 70) C O 2 1 1 I 3bl T G l l l 1112) 16361 1373) 16391 c 201 cot8
I 73) IGSW I 39) CObR 1135) 16160 12181 lGZI1
1 121 1Gb8 I 8.91)
9.0
IIS.~Il 15.9
1 2.81) 11.311
2.1
1 1.511
1.5
I 6.61)
6.b
1I 0 , 2 1I
lo,*
III.OI) 11.2 1113) 16180
1.1
47
48
ANDREW H. PATERSON et a/.
maps are said to be complete when newly added markers invariably fall near previously mapped markers. Having a complete genetic map is important in locating genes of unknown map position, because one can be confident that all regions of all chromosomes are searched-no stones have been left unturned. To complete a genetic map, one must map a large number of markers, a number which can be estimated from the number and approximate length of chromosomes, and the desired average distance between markers (Lange and Boehnke, 1982). The principles of genetic mapping have been known for nearly a century; however, recently developed molecular techniques for finding large numbers of genetic markers quickly (Botstein et al., 1980) have stimulated the construction and use of genetic maps in a variety of plants (Table I), at least two plant pests (pathogenic fungus of lettuce Bremia lacrucae; Michelmore and Hulbert, 1987; Hulbert et al., 1988; pathogenic fungus of rice Magnaporthegrisea; B. Valent, F. Chumley, and J. Sweigart, personal communication), and several animals (Womack et al., 1990), including human (DonisKeller et al., 1987).
111. USING GENETIC MARKERS TO STUDY AND
IMPROVE AGRICULTURAL PRODUCTIVITY Agricultural productivity is the result of growing superior genotypes in an environment which allows them to express their superiority (Boyer, 1982). To create such a superior genotype, the breeder must assemble many genes which work well together, and which adapt the genotype to the target environment. Soil and weather conditions, as well as incidence of diseases, insects, and other pests, can vary substantially over short distances. The importance of environment on agricultural productivity is reflected in the ongoing successes of international agricultural centers, and state “Agricul-
Fig. 3. Genetic Linkage map oftomato, based upon isozymeand restrictionfragment length polymorphism (RFLP)markers. This framework map shows the Linear order and relative distances between markers which could be ordered with statistical confidence (to L.O.D. > 3) using the MapMaker computer program (Lander et a/.. 1987). This map does not show multiple markers which mapped to the same locus (ie., showed no recombination with each other).In total, the tomato RFLP map nresently includes about 800 markers,and spans a length of 1480 centiMorgans. (Canal and Tanksley, unpublished data.)
DNA MARKERS IN PLANT IMPROVEMENT
49
tural Experiment Stations,” in tailoring crop varieties to suit local conditions. While variation in environment creates the need for local breeding stations, these stations are united in practicing genetic manipulation to improve agricultural productivity. Genetic manipulation of agricultural species is not nearly so elegant as that possible for experimental organisms such as Escherichia coli (gut bacterium), Saccharomyces cerevisiae (yeast), or Drosophifa melanogaster (fruitfly), for several reasons. First, higher plants and animals tend to have relatively large quantities of DNA in their chromosomes, making it difficult to fish out particular genes of interest. Further, higher plants and animals have relatively long generation times, months to years (versus a few minutes for E. coli). Finally, measures of agricultural productivity usually reflect the effects of many genes, acting at different times during the lengthy period ofgrowth and development ofthe organism. Such multigenic or polygenic traits have proven especially difficult to improve through classical breeding methods. Deoxyribonucleic acid markers can be used to obtain much information about the genes which influence an agriculturally important trait, thus facilitating breeding efforts. In particular, one might estimate: 1. Number of genes: How many genes influence a trait? 2. Chromosomal location: Where are the genes located on the chromosomes ( e g , near which genetic markers)? 3. Phenotypic effect: How much does each gene affect the trait? 4. Gene dosage (or “gene action”): Does an individual carrying two copies of a gene look different from an individual carrying only one? 5. Pleiotropy: Do individual genes affect more than one trait? 6. Environmental sensitivity: Do genes function similarly in different environments? 7. Epistasis: Does the effect of one gene influence the effect(s) of others?
Answers to these questions might help the breeder to “design” crop vaneties that consistently perform well in a defined environment. By incorporating genes which protect against environmental hazards, such as drought, heat, cold, salinity, or particular races of pests, a variety might retain some productivity even under suboptimal conditions (which are the rule, rather than the exception; Boyer, 1982). A number of scientific developments, spanning the past century, have only recently coalesced into a means by which the breeder might obtain detailed answers to these questions. First was the discovery by Mendel ( 1865) that genetic factors behave as discrete particles, when passed from parent to offspring. Next was the discovery by Morgan (19 1 1) of linkage between genetic factors (described previously). Finally, several workers showed that
ANDREW H. PATERSON et al.
50
Table I Lmkage Maps of DNA Markers in Crop Plantsa6 Taxon Soloname Lycopersicon esculentum Solanum tuberosum
Haploid chromosomes
Markers
-
12
800 -400
12
- 150
7
19
10 12
468 -400
Triticum aestivum
21
79
Sorghum bicolor Hordeum vulgare
10
Saccharum oficinale
64 (32-80)
Capsicum annum Petunia hybrida Poaceae Zea mays Oryza sativa
Pemisetum glaucum Fabaceae Glycine max Phaseolus vulgaris
12
7
52
- 90 20
7
in progress
20 II
- 550
I
-100
Reference
S. D. Tanksley (personal communication) Bonierbale et al. (1988);Gebhardt et al. (1989b) Tanksley et al. (1988);J. P. Prince and S. D.Tanksley (unpublished observations) Cornu et al. ( 1990) Coe et ul. ( 1990) McCouch er al. (1988);S. D.Tanksley (personal communication) Sharp et al. (1989);Chao ef al. (1989); Kam-Morgan and Gill ( 1990) Hulbert et al. ( 1990) M. Huen and M. E. Sorrells (personal communication); Blake (1990) W. Burnquist and M. E. Sorrells (personal communication) Ozias-Akins (I 989)
J. A. Rafalski (personalcommunication) C. E. Vallejos (personal communication); Gepts ef al. (1990) Weeden and Wolko ( 1990) G.Bauchan (personal communication); C. Echt (personal communication)
Pisum sativum Medicago sativa
6
in progress
Brassicaceae Brassica oleracea Brassica campestris
9 10
-450 - 150
19
117
Slocum et al. ( 1990);Landry ( 1990) Slocum ( 1989); R. Bernatzky (personal communication) Landry ( 1990)
9
160
Kesseli et al. (1990)
Brassica napus Asteraceae Lactuca sativa F’inaceae Pinus taeda Rosaceae Prunus persica Rosa spp. Linaceae Linum usitatissimum
14
12
in progress
Neale ef al. (1989)
7
in progress in progress
Morgens et al. (1 989); Rogers et al. (1 989) Hubbard et al. (1989)
15
in progress
Cullis et al. (1990)
‘Based on the most current information we were able to obtain by the time of submission. Taxa, and species within taxa, are ordered by the number of markers mapped. Assignment to familiesaccordingto Radford, A. E., Ahles, H. E., and Bell, C. R. 1979.“Manual ofthe Vascular Flora of the Carolinas.” Univ. of North Carolina Press, Chapel Hill.
’
DNA MARKERS IN PLANT IMPROVEMENT
51
quantitative variation in phenotype might be explained by Mendel’s discrete particles, setting the stage for studies of association between genetic markers and virtually any measurable trait.
A. LINKAGEBETWEEN GENETIC MARKERS AND GENES INFLUENCINGSIMPLYINHERITED TRAITS Two nearly simultaneous discoveries early in the Twentieth Century set the stage for applying linkage analysis to crop improvement. One was the discovery by Morgan (191 1) of genetic linkage, i.e., that Mendelian genetic factors which lie close together on a chromosome are usually cotransmitted from parent to progeny. This meant that genetic factors with striking effects could serve as a “proxy” for other nearby genes whose effects are difficult to discern. Many such associations between easily identified genetic markers and simply inherited traits have been found, and put to good use. A particularly good example is the close linkage between nematode resistance and the allozymic variant Apsl (Rick and Fobes, 1974). Scoring plants for an allozyme (or DNA) marker can be done at seedlingstage,and is much easier than measuring susceptibility to a subterranean pest. One need not actually expose plants to the pest, but instead can select for resistance based on the presence or absence of the marker. This reduces dependence on unreliable natural infestations of the pest, and circumvents the need to artificially introduce noxious pests into a greenhouse or experimentalfield. The use of near-isogenic lines for rapidly finding markers associated with particular traits (Young et ul., 1988) will no doubt add greatly to the list of known markers for simply inherited traits in the near future.
B. LINKAGEBETWEEN GENETIC MARKERS AND GENES INFLUENCINGQUANTITATIVE TRAITS After the discovery of genetic linkage, another important step toward genetic analysis of agricultural productivity was the multiplefactor hypothesis (East, 1915; see also Yule, 1906)of quantitative inheritance. For simply inherited traits, a difference between parents in one or two genes explains virtually 100%of the differences among the progeny-such a trait is itself a genetic marker. However, most measures of agricultural productivity, such as size, shape, yield, or quality, are influencedby many genes. Further, these traits tend to be quantitative, in that the progeny do not fall into discrete classes but show a continuous range of phenotypes between the parental
52
ANDREW H. PATERSON et d
phenotypes (and occasionally beyond the parental phenotypes; see Transgression, in Section V,B). For a long time, it was not clear how Mendel’s discrete genetic factors could give rise to individuals whose phenotype (e.g., height, or weight, or yield, or other traits)was intermediate between those of their parents. Experiments with beans (Johanssen, 1909), wheat (Nilsson-Ehle, 1909), and tobacco (East, 191 9 , showed that such continuous variation in phenotype could be explained by the independent actions of many discrete genetic factors, each factor having a rather small effect on the overall phenotype. These “polygenes” or quantitative trait loci (QTLs) (Geldermann, 1975) were hypothesized to differ little from genes affectingsimply inherited traits; i.e., they were expected to exhibit the fundamental Mendelian properties of segregation and recombination. Consequently, it seemed likely that one might identify linkages between genetic markers and QTLs, and thus determine the locations of individual QTLs. Linkage between a genetic marker and a QTL was first demonstrated by Sax ( 1923), who found that Phaseolus genotypes with different seedcoat colors also differed in average seed size. Associationsof many different traits with visible markers were promptly found in severalother species (see Smith, 1937, for citations of early studies on tobacco, tomato, maize, pea, bean, barley, and Drosophila, using visible markers). More recently, associations of single isozyme and restriction fragment length polymorphism (RFLP) markers with many quantitative traits have been found, particularly in tomato (Tanksley et al., 1982; Vallejos and Tanksley, 1983; Nienhuis et al., 1987; Osborn et al., 1987; Weller, 1987; Weller et al., 1988; Tanksley and Hewitt, 1988;Martin etal., 1989),andinmaize(RussellandEberhart,1970; Stuber and Moll, 1972; Stuber et al., 1980, 1982, 1987; Frei et al., 1986; Edwards et al., 1987),but also in several other species (Everson and Schaller, 1955; Breese and Mather, 1957; Law, 1967). These pioneering efforts contributed significantly to understanding the genetic basis of quantitative variation, and stand as models for design of subsequent experiments. Many of these early studies had rather few genetic markers to work with. This was a serious limitation, as one could show that QTLs were somewhere in the vicinity of a particular marker, but seldom had enough additional markers nearby to help pinpoint the location of the QTL (Fig. 4a). As genetic maps came to include more markers, it became possible to more precisely estimate the location of a QTL, by studyingseveral markers along a chromosome (Fig. 4b) (Thoday, 1961; Tanksley et al., 1982)(see Section IV). Thoday ( 1961) emphasized that “the main practical limitation . . . (to localizing QTLs) . . . seems to be the availability of suitable markers.” This limitation was remedied by the construction of complete RFLP linkage maps, permitting systematic searches of an entire genome for QTLs in-
53
DNA MARKERS IN PLANT IMPROVEMENT I
a
aa: mean=l.8 g
Aa: mean=2.8 g
AA: mean=3.8 g
L 3 L
L
a a'-
cn
r+ I
cram
k
L
cn
0
I I
.rrr _.rrr
I
Ib
Fig. 4. Using genetic markers to map genes atfectng quantitative traits. (a) Single-marker association. A number of individuals are measured for the trait of interest (in this case, tomato fruit size), and scored at a genetic marker locus. The average trait phenotype is determined for each marker genotype. In some cases, there is an association between trait phenotype and marker genotype;for example, in the case shown, each copy of the A allele is associated with an increase in mass per fruit of about 1 gram. Hence, it is inferred that a genetic factor affecting mass per fruit lies somewhere in the region of marker A. (b) Multiple-marker association (intervalmapping). By studyingother markers along the chromosome (e.g., markers X,Z, G, N, and T), in addition to marker A, it can be determined more precisely where the gene lies, and how much it affects the trait (Paterson ef al., 1988, 1990b; Lander and Botstein, 1989). In this case, five additional markers were chosen which had previously been mapped, and were known to lie at intervals of about 20%recombination, along the length ofthe chromosome (asindicated along horizontal axis ofgraph). The peak effect on the trait was quitedistant from marker A, and was considerably greater than had been estimated based on marker A. The peak effect of about 1.5 grams, suggested that the most likely location of the gene@) responsible (indicated by arrow) was about halfway between markers G and N. A likelihood interval for the gene(s), in which one can be 90% confident that the gene is located (drawn as a horizontal bar), approximately coincides with the G-N interval.
+
54
ANDREW H. PATERSON et al.
fluencing a trait (Paterson et al., 1988). Further, new algorithms for QTL mapping minimized the number of individuals and number of genetic markers needed to map QTLs (Lander and Botstein, 1989; see also Weller, 1987, for a related technique).
IV. DESCRIBING INDIVIDUAL QUANTITATIVE TRAIT LOCI How does one describe a gene? In a simple organism such as E. coli, one can readily isolate individualgenes associated with a particular function. In a few Petri dishes, one might select among literally millions of individuals to identify mutants in the function of interest, then artificially introduce small bits of wild-type (e.g., nonmutant) DNA into the mutants and identify those bits (genes)which restore normal function. Thus having a gene in hand, one might determine the sequence of genetic information comprising the gene, the protein product encoded by this information, the function of the protein, the regulation ofactivityof the gene or protein by environmentalfactors,and so on. Such elegant schemes for gene cloning are rarely, if ever, applicable to crop plants. First, one is seldom able to study millions of individuals! Further, one seldom knows enough about the function of a gene to isolate it from the many thousands of other genes in the organism. Finally, for traits influenced by several genes, the effects of any one gene are often partly masked by other genes and/or by environment. Thus, it is difficult to discern the effect of any one gene by merely looking at the appearance (phenotype) of one individual. Using DNA markers, QTLs can be described by their chromosomallocation, dosage effect, phenotypic effect@),and sensitivity to environment. Such a description has long been possible for genes affecting simply inherited traits; however, only with the availabilityof high-density genetic maps has it become possible to obtain this information for individual QTLs. In the next section (IV,A), we will discuss how this information is obtained. In a later
‘
However, for some traits, one can study millions of genotypes, if the trait is expressed by cultured cells (Chaleff, 1983), or by pollen (Freeling and Bennett, 1985). Further, in the tiny plant Arabidopsis thaliana. one can screen enormous numbers of individuals for mutationsin genes of unknown function by creating the mutations with a DNA element containing an antibioticresistance gene (Feldmann et al., 1989). By collectinga large number of such mutations, one might ultimately assemble a “library,”including mutationsin many of the expressed genes in the organism.
DNA MARKERS IN PLANT IMPROVEMENT
55
section (XI), we will discuss how this information might eventually facilitate cloning of QTLs, permittinga molecular description of the underlyinggenes. Before proceeding, we note that the sort of description of QTLs which we will present has been obtained in a rather small number of cases, to date. Some of the reasons for this are that DNA markers have only recently been available, and are still not available in some species, polymorphism between parents is not always optimal, and the necessary experiments are laborious and expensive. These and related issues are addressed in more detail later (Section XII). Despite the fairly limited number of test cases, however, the principles discussed in the following sections should prove broadly applicable.
A. CHROMOSOMAL LOCATION One determinesthe chromosomal location of agene by identifjmg nearby genetic markers which are usually cotransmitted with the gene from parent to progeny. This principle applies both to genes with large effects on phenotype (simply inherited traits) and genes with small effects on phenotype (QTLs). Using even a single genetic marker, one can determine that a QTL (or other gene) lies somewhere in the vicinity of the marker (Fig. 4a). However, to determine the location of a QTL (or other gene), one needs to study several markers along the chromosome (Fig. 4b). Thoday ( 1961) and Tanksley et al. ( 1982) demonstrated the fundamentals of QTL mapping. With a complete map ofgenetic markers, one can employ powerful statistical methods (Lander and Botstein, 1989), to determine likelihood intervals for the locations of QTLs (Paterson et al., 1988). In tomato, numerous QTLs have been mapped to individual likelihood intervals measuring about one-fifth of a chromosome; e.g., about 2% of the genome (Paterson et al., 1988, 1990b). Having mapped a QTL to a likelihood interval, one might narrow the location of the QTL down to perhaps 1/50- 1/100 of a chromosome (about 0.1% of the genome in tomato), by comparing individuals which carry different portions of the likelihood interval (Fig. 5 ) (Paterson et al., 1990a). Such fine mapping is necessary for several purposes (Section IV,C; Pleiotropy, and Section IX, Utilization of Exotic Germplasm). However, even with this high-resolution mapping, it is difficult to prove that a QTL represents only one gene. It is generally acknowledged that a typical higher plant genome includes perhaps 10,000 to 100,000 genes. Consequently, 0.1% of the genome would include an average of 10- 100 genes. Several genes lying close together, each with a small effect on a trait, could appear to be a single QTL of large effect (Michelmoreand Shaw, 1988; Paterson et al., 1988).To
A
B
1 x
Z
A
G
N
T
C E
D
Fig. 5. High-resolution mapping ofquantitative trait loci (QTLs) by substitution mapping (Paterson et al., 1990a). Starting from a cross between two genotypes which differ in a trait of interest (e.g., fruit size), and whose chromosomes can be distinguished by genetic markers (e.g., filled versus open) (A), one performs a first-generation analysis to determine the approximate location of genes of interest [(B), from Fig. 41. Having established this likelihoodinterval for the location of a gene of interest, one can use additional mapped DNA markers (C) to identify individual progenieswhich carry differentchromosome segments in the region of the likelihood interval (D). One then determines the phenotypes associated with each of these different chromosome segments [(E); note that this usually involves measuring many individuals of the same genotype, especiallywhen the effect of the gene is small]. Usually, not all of the chromosome segments will be associated with the trait. By identifyingthe chromosomal region unique to those segmentswhich do show the trait, one can map the location of the gene@)involved to very fine resolution (F). Using this technique (Paterson et al.. 1990a),the primary limitation to map resolution is having enough genetic markers to distinguish between chromosome segments with small differences. RFLP, Restriction fragment length polymorphism.
DNA MARKERS IN PLANT IMPROVEMENT
57
reveal just what lies at a quantitative trait locus, different techniques with much higher resolution will probably be necessary (Section XI).
B. EFFECTSOF GENEDOSAGE (OR GENE “ACTION”) Many plants and most animals are diploid, having either zero, one, or two copies (alleles) of a gene.5 The phenotypic effect of an allele often depends upon the number of copies present in an individual, and the nature of the alternate allele (when two allelesare present together). For example, consider a hybrid between two parents which each carry two copies of a different allele, and each have a different phenotype. Thus, the hybrid is heterozygous, carrying one copy of each parental allele. If the phenotype of the hybrid is intermediate between those of the parents, the two alleles are said to have an “additive” relationship- one copy elicits an effect, and two copies produce twice the effect of one (Fig. 6A). If the phenotype of the hybrid is identical to that of one parent, the two alleles are said to have a dominant/recessive relationship. The dominant allele is that which determinesthe phenotype of the hybrid; thus only one copy of the allele is necessary to elicit a maximal effect (Fig. 6B). The alternate allele is said to be recessive, because its effect is only seen in the absence of the dominant allele (e.g., when the recessive allele is present in two copies, such as in the parent) (Fig. 6B). Finally, an allele is overdominant if one copy has a greater effect than two (Fig. 6C).6 Strategies for using genetic markers to study effects of gene dosage on quantitative traits have been worked out in detail (Mather and Jinks, 1971; Tanksley et al., 1982; Soller and Beckmann, 1983; Edwards et al., 1987; Lander and Botstein, 1989;Paterson et al., 1990b).Fundamentally, one first studies genetic markers near an allele of interest, to determine whether an individual is carrying 0, 1, or 2 copies of the allele. Next, one determines the average phenotype of individuals carrying 0, 1, or 2 copies of the allele. Finally, statistical techniques are used to determine whether the change in phenotype per additional copy of the allele most closely fits additivity,dominance/recessiveness, or overdominance. Individual QTLs have been found which show additive (d = 0),dominant Polyploids, which can have more than two copies, will be discussed in Section VIII. We note that gene dosage is determined in reference to particular alternate alleles; allele B may be dominantto allele A, while allele C could be dominant to B. Thus, the dosage effect of a particular allele may be different in crosses to individualscarrying different alternate alleles. In addition, it is noted that the reciprocal of overdominanceis underdominance, where the hybrid shows lower expression of the trait than either parent.
ANDREW H. PATERSON et ul.
58
A 0
-
Midparent value
/'
--bb
Bb
BB
GENE DOSAGE
-
88 . B
/d
/
- -
bb
BB
Bb
GENE DOSAGE
C w
E P B
-
\
0 '
\
0/-
-
\
-- m
I
m
I
m
m
Midparent value
DNA MARKERS IN PLANT IMPROVEMENT
59
(d = a), recessive (d = -a), and overdominant (a = 0, d > 0) gene action (Paterson et al., 1990a,b).Further, some QTLs cannot be clearly assigned to any of these classes, as they appear intermediate between two classes (e.g., partial dominance; a > 0, d > 0). Thus, these classes of gene action are perhaps more realistically interpreted as extremes in a continuum, rather than discrete entities. The ideal population for studyingeffects of gene dosage in diploidsis an F, self- or intercross, segregatingin the classical 1 :2 : 1 Mendelian ratio for the three possible genotypes. In several other types of populations widely used in breeding, dosage effects of alleles are not readily determined, because one of the three possible genotypes does not occur. For example, in a backcross (Brigs, 1938), homozygotes for the donor genotype do not occur; thus one sees no effect of recessive factors, and consequently cannot distinguish between additive, dominant, or overdominant factors. In single-seed descent (Brim, 1966)or recombinant inbred populations (Bailey, 198l), the heterozygous class is rare or absent; thus, one sees no effect of overdominant factors, and consequently cannot distinguish between additivity and dominance/recessiveness.
C. MULTIPLE EFFECTSOF SINGLEGENES (PLEIOTROPY) Plant breeding decisions frequently involve tradeoffs- maximizinggains in one trait while minimizing losses in another. The need for such tradeoffs can be a consequence of different genes which are closely linked, or a single
Fig. 6. Effects of gene dosage on individual quantitative trait loci (QTLs). (A) Additivity. The phenotype of the heterozygote falls exactly halfway between the parents (e.g., at the midparent mean). Additivity is expressed as the additive effect (a)of an allele on phenotype, which is thus half the differencebetween phenotype of the parents. (B) Dominance/Recessiveness: The phenotype of the heterozygote is identical to that of one parent, and the allele from that parent is designated dominant. The allele of the alternate parent is designated recessive, as its effect is only seen when the dominant allele is not present. Dominance is expressed as the dominance deviation (4 of an allele, or the difference between the midparent value (e.g., the average of the parent phenotypes) and the observed phenotype of the heterozygote. (C) Overdominance. The phenotype of the heterozygote is superior to that of either parent. In this case, the additive effect is zero (since the parents are the same), but the dominance deviation is significant. (Note: the reciprocal is underdominance, where the phenotype of the heterozygote is inferior to that of either parent, and the dominance deviation is negative.)
60
ANDREW H. PATERSON et al.
gene which affects several traits (pleiotropy; Gruneberg, 1938).' One can sometimes use genetic mapping to determine whether an association between different traits is due to several linked QTLs which might be separated by recombination, or to a single QTL with several inseparableeffects. Techniques have been described for narrowing the location of a QTL down to about 0.1%of the genome (Paterson et al., 1990a).However, as suggested in Section IV, A, this might include 10- 100genes. Consequently,QTLs which map to different locations can be inferred to represent different genes, but QTLs which map to similar locations cannot be assumed to represent the same gene. Different techniques are needed to demonstrate that two traits are caused by the same gene (see Section XI).
D. ENVIRONMENTAL EFFECTS Plant breeders routinely find that genotypes which perform well in one environment are not well suited to different environments. Sometimesthese differences among genotypes can be attributed to relatively simply inherited attributes such as susceptibility to particular strains of pest, photoperiod response, or vernalization requirement. However, differencesin adaptation of plant or animal genotypes to particular environmentsmay also be due to environment-sensitiveQTLs. In a study by Paterson et al. (1990b), it was found that among a total of 29 QTLs mapped in three different environments, 4 (14%) were expressed in all three environments, 10 (34%) were expressed in two environments, and 15 (52%) were expressed only in a single environment. The observation that individual QTLs show a range of sensitivities to environment represents both a difficultyand an opportunity. It would be simpler for the breeder to work with QTLs which function consistently over a range of environments. However, by combining several QTLs with different environment specificities into a single genotype, one might elicit an improvement in phenotype which is somewhat buffered against the vagaries of environment. It also must be emphasized that the parameters which define an environment are seldom, if ever known (although methods for evaluating the relative importance of measured environmentalvariables have been suggested) (Brown et al., 1983). However, with sufficient testing, it may be possible to
'
Of course, such trade-offs can also be due to independent genes which are completely unlinked, but this situation is easily remedied.
DNA MARKERS IN PLANT IMPROVEMENT
61
assert that particularQTLs, like particular cultivars,perform satisfactorilyin some production regions but not others.
V. CUMULATIVE EFFECTS OF MANY QUANTITATIVE TRAIT LOCI ON PHENOTYPE OF AN INDIVIDUAL A. HYBRID VIGOR(HETEROSIS) One of the most significant events in the history of plant breeding was the discovery of “hybrid vigor” (Shull, 1908, 1911; East, 1908), also called heterosis (Shull, 1914).By mating two unrelated maize inbreds of rather low productivity, heterozygous F, progeny resulted which were dramatically larger and more fruitful than either parent. This finding was ultimately to transform U.S. maize breeding and production, from the use of open-pollinated landraces to nearly 100%hybrid varieties. This shift to hybrid varieties partly accounts for sharply improved yields of maize through the mid Twentieth Century (Fehr, 1984).The principle of heterosis applies to many other crops as well, especially those which are naturally cross-pollinated. Heterosis is clearly related to heterozygosity. If an F, hybrid which shows heterosis is selfed or intermated to its siblings, the average performance of progeny declines with each subsequent generation, concomitant with declining heterozygosity (East, 1908).The average grain yield of a maize plant increaseswith increasingnumber of heterozygous marker loci in the genome (Edwards et al., 1987).Crossesbetween closely related maize inbreds tend to show less hybrid vigor (Lee et al., 1989)than more distant matings. While the association of heterosis with heterozygosity is clear, it has long been debated why heterozygosity results in heterosis. An F, hybrid is heterozygous for all genetic factors which differ between its inbred parents. Thus, dominant factors from either parent could mask deleterious recessive mutations from the other parent, generating progeny with the strengths of both parents but none of the weaknesses- the dominance theory of heterosis (Bruce, 1910; Keeble and Pellew, 1910). Alternately, individuals carrying two different parental factors at a locus could be inherentlysuperior to either parental type -the overdominance theory of heterosis (Shull, 1908, 191 1; East, 1908). A complete discussion of each of these theories has been the topic of many papers, and at least one book (Gowan, 1954),and is beyond the scope of this chapter. Briefly, it can be difficult to distinguish between dominance and overdominance as a basis for heterosis; for example, two dominant genes
62
ANDREW H. PATERSON et ul.
closely linked in repulsion could behave essentially as a single heterotic locus (Jones, 1917). Almost certainly, both dominance and overdominance account for some cases of hybrid vigor. Our point is this: Genetic markers improve one’s ability to study the effects of individual genes (with the limitations discussed previously; see SectionsIVA and C). Consequently, it might more readily be determined whether a trait which shows heterosis is influenced largely by dominant genes, overdominant genes, or a mixture ofthe two. Through the use of genetic markers, the near future may hold a better understanding of the genetic basis of heterosis, and perhaps additional gains in agricultural productivity by further exploitation of heterosis.
B. PROGENY SUPERIORTO BOTH PARENTS (TRANSGRESSION) New cultivars can also be superior to either of their parents for reasons other than hybrid vigor. For example, many of the crops that feed the world are grown largely as pure lines, with virtually no heterozygosity, yet genetic improvement of such species (e.g., wheat and rice) has been credited with producing a “Green Revolution” (Borlaug, 1983). In crosses between individuals which are similar, one occasionally finds progeny which are superior to either parent, a phenomenon referred to as transgression (Simmonds, 1979). This is especially true of crosses between elite cultivars, each of which shows superior performance, but by virtue of carrying different genes. Thus, by recombination of genes from different elite cultivars, a few progeny may receive a more complete repertoire of “favorable” genes than either parent. The genetic basis of transgression may be described at the level of individual genes, using genetic markers. For example, two wild relatives of tomato (Lycopersicon esculentum) which have been studied in detail, namely L. chmielewskii and L. cheesmanii, are known to have fruit pH only slightly higher than that of tomato. Yet, DNA markers have shown that each wild species carries some alleles which increase pH (+), and other alleles which decrease pH (-), relative to the correspondingtomato alleles(Paterson et al., 1988,1990b).As a consequence, occasional progeny carry most of the (+) or most of the (-) alleles;thus, the progeny show a greater range in pH than the parents. Identification of rare transgressant individuals is fundamental to plant breeding progress, and might be facilitated by genetic markers. Transgressants for fruit pH in tomato will undoubtedly not solve the world hunger problem; however, the genetic principle explaining their occurrence is equally applicableto other traits, such as yield, disease resistance, or productivity under adverse environments.
63
DNA MARKERS IN PLANT IMPROVEMENT
bb
Bb
BB
GENE DOSAGE
Fig. 7. Epistasis. The effects of one gene can depend upon the presence of another. Allele B has no effect in the absence of allele C, and has much more effect in CC than in Cc individuals.
C. INTERACTIONS
BETWEEN
GENES (EPISTASIS)
An organism results from the collective actions of many different genes, with no single gene functioning completely independently of all others. The whole is not simply the sum of the parts. Considerable evidence from classical quantitative genetics indicates that interaction between different genes, or epistasis,can influence the phenotype of the individual (Wright, 1968; Allard, 1988). As was the case for gene dosage, the use of genetic markers to study epistasis has been addressed by numerous workers (Spickett and Thoday, 1966; Mather and Jinks, 197 1 ;Tanksley et al., 1982; Soller and Beckmann, 1983; Weller et al., 1988; Edwards et al., 1987; Allard, 1988; Paterson et al., 1988, 1990a,b),and is simplistically illustrated in Fig. 7. One first determines whether an individual is carrying 0, I , or 2 copies of gene A. Then one subdivides these classes, based on which individuals are carrying 0, 1, or 2 copies of gene B (generating nine possible combinations, from 0 0 to 2 2). When the effect of allele dosage at one gene is altered by allele dosage at another gene, epistasis is indicated. Genetic mapping experiments have documented only a few cases of epistasis, among many candidates which have been studied (Spickett and Thoday, 1966; Weller et al., 1988; Edwards et al., 1987; Allard, 1988; Tanksley and Hewitt, 1988; Paterson et al., 1988, 1990a,b). This may indicate that epistasis is not especially prominent, at least in the species and traits which have been studied. Alternately, genetic mapping experiments may simply not have the statistical resolution to detect epistasis, due to studying rather
+
+
64
ANDREW H. PATERSON e t a/.
small populations (Paterson et al., 1988, 1990a,b), or using sparsely distributed geneticmarkers (Weller et al., 1988;Edwards et al., 1987,Allard, 1988). Understanding epistasismay be of much practical importance-classical plant and animal breeders know from long experience that genetic background can have a pronounced effect on both simply inherited and quantitative traits; thus, some degree of epistasis is likely to affect many traits. It would be good news if it turned out that epistasis has a relatively small influence on the overall phenotype, as it is far easier to identifyand transfer genes individually than in pairs, trios, et cetera. Further study of epistasis is clearly warranted.
VI. IMPROVING THE EFFICIENCY OF BREEDING PROGRAMS Plant breeders have developed many sophisticated strategies for accomplishing particular breeding objectives in different species, and under different situations(Allard, 1960;Simmonds, 1979).However, these strategiesare similar to one another in that breeding progress is largely determined by three factors: 1. Identifying genetic variation among different individuals, cultivars, races, or species which show different attributes, or which show a common attribute but due to different genes. 2. Making crosses and obtaining recombination between genotypes with different attributes, creating genotypes with new sets of attributes (or superior levels of a particular attribute). 3. Accurate selection of these occasional genotypes with new sets of attributes, from large populations grown in variable environments and including many inferior genotypes.
Each of these areas can benefit from use of genetic markers, especially DNA markers. Genetic markers represent genetic variation, permitting one to estimate relatedness between different genotypes, and consequently to predict which matings might produce new and superior gene combinations. Further, by having markers for genes of interest, one can readily detect recombination between these genes, and perform accurate selection for genetically superior individuals, from among the masses of candidatesinclud-
DNA MARKERS IN PLANT IMPROVEMENT
65
ing many “pretenders” who were favored by environment, rather than genetics.
A. IDENTIFYINGGENETIC VARIATION Within cultivated crop varieties, one can use geneticmarkersto determine relatedness between individuals. Much of this work has employed isozymes, largely because they are simply detected, at low cost. With as few as 20 markers (Soller and Backmann, 1983),one might readily distinguish among unrelated maize inbreds, and many such efforts have been conducted, especially in maize (Stuber and Goodman, 1983;Smith, 1984, 1988;Smith et al., 1985).In the extreme, one can “fingerprint” varieties, providing a reproducible genetic description which is useful for variety protection (Burr et al., 1983). Genetic markers also can be used to portray diversity (or lack thereof) within cultivated germplasm (Smith, 1988), and to identify groupings of cultivars which are adapted to particular regions (Souza and Sorrells, 1989) or perform similarly in crosses to other cultivars (Lee et al., 1989). Again, maize has been prominent in this area of research, partly because it shows a high level of genetic diversity among cultivated types (although only a small portion of this is represented in the most widely grown maize hybrids; Smith, 1988). In addition to maize, other outcrossing species such as Brassica (Figdore et al., 1988), and potato (Gebhardt et al., 1989a)also show considerable variation within cultivated types. This abundance of variation is in contrast to inbreeding species such as soybean (Glycine m u ; Apuya et al., 1988; Keim et al., 1989;Tingey et al., 1989),tomato (L.esculenturn;Miller and Tanksley, 1991), wheat (Triticum aestivum; Chao et al., 1989), and cotton (Gossypium hirsutum, G. Barbadense; A. H. Paterson, unpublished observations), which show relatively little variation in genetic markers among cultivars. This lack of DNA variation, despite clear morphological variation among cultivated varieties, illustrates two points: ( 1) Genetic markers identified (and examined) to date represent a tiny fraction of the genomes of most crop plants. For example, the 800 markers mapped in the tomato genome represent about 8 X lo5 base pairs (bp) (S. D. Tanksley, personal communication), of a total of about 6 X lo8 bp (Galbraith et al., 1983), or about 0.13% of the genome. A change of as little as a single nucleotide among these 6 X lo8, or about 0.0000000016% of the genome (for tomato), can have a pronounced visible effect. Consequently, it is not surprising that cultivars which look quite dissimilar in the field can be difficult to distinguish in the laboratory. Finding such a change as a random
66
ANDREW H. PATERSON et al.
DNA marker is analogous to seeking a needle in a haystack. (2) Although genetic markers comprise only a tiny portion of the genome,lack of variation in genetic markers may reflect less than desirable levels of variation in agriculturallyimportant genes. This points to an urgent need for collecting, cataloging, and utilizing exotic germplasm in crop improvement, and DNA markers may aid in this objective (see Section IX).
B. RECOMBINATION AND SELECTION After establishingthat a gene of interestlies near a particular DNA marker, one can assay the marker (genotype), and predict with high likelihood whether the gene is present or absent, even before the trait (phenotype) can actuallybe seen. This can be quite valuable. A DNA marker can be assayedat the seedling stage, permitting one to make selectionsbefore many traits can be seen, thus reducing the number of individuals which must be grown to maturity. Further, many traits may be more accurately selected for by using genotype at DNA markers than by relying solely on appearance, which may be due either to genotype or to environment (Burr et al., 1983; Stuber and Edwards, 1986; Soller and Beckmann, 1988; Lande and Thompson, 1990; Paterson et al., 1990b). Finally, unlike many traits, genetic markers can be reliably assayed in nontarget environments such as the growth chamber, greenhouse, or winter nursery, permitting more rapid progress in breeding many species. By assaying DNA markers, one also reveals which individuals in a population are genetically similar to one another. Classical breeding generally employs rather weak selection,choosingperhaps 25% of the individuals in a population for further study. By using genetic markers to determine which individuals in a population are similar, one might reduce the selected fraction to fewer individualswhich collectivelycarry all the genes ofinterest, and advance only these few to the next generation. For example, from about 1000 tomato plants segregating for unlinked segments of seven L. chrnielewskii chromosomes, an experienced tomato breeder chose about 25 individuals as sufficiently desirable for further experiments. Restriction fragment length polymorphism analysis revealed that these individuals included only eight different genotypes (A. H. Paterson, J. Deverna, M. Kuehn, and S. D. Tanksley, unpublished observations); thus, the size of subsequent experiments was reduced by 3-fold in a single step. The efficiencies of scale and time accorded by DNA markers are valuable in breeding most species, but are of special value in breeding species with large stature or long generation time, such as orchard or forest trees, where
DNA MARKERS IN PLANT IMPROVEMENT
67
fewer individuals might save hectares, and fewer generations might save decades.
VII. MARKER-FACILITATED STUDY OF COMPLEX POPULATIONS Most natural populations of plants, as well as many experimentalpopulations, are genetically quite different from the classical linkage mapping populations discussed in Section 111. While linkage mapping populations are commonly derived from two-generationcrossesbetween two parents, many natural and experimental populations are derived from multi-generation matings between an assortment of different parents, resulting in a massive reshuffling of genes. Individuals in such populations carry a complex mosaic of genes, derived from a number of different founders of the population. Gene frequenciesin the population as a whole may be modified by natural or artificial selection, or by genetic drift (e.g., chance) in small populations. Study of such complex populationshas yielded much information on both adaptation by natural selection (Allard, 1988), and on genetic change as a result ofartificial selection (Stuber et al., 1980).These two modes ofselection are quite different, since natural selection tends to favor an intermediate phenotype, which is adequately suited to a number of diverse environments (Lande, 1976), while artificial selection tends to favor an extreme phenotype, which gives superior performance in a rigidly defined environment. Classicalquantitative genetics was a startingpoint for investigatinggenetic change under long-term selection regimes, however, simply measuring the phenotype of an individual provided insufficient information about the underlying genetic factors. In the words of Allard ( 1988): Emphasis was therefore shifted . . . to a particulate approach . . . determining the individual effects of single marker loci on adaptive change, then determining the joint effects of pairsofloci . . .
In experimental barley populations subjected to natural selection, extensive study by Allard and colleagues (summarized in Allard, 1988) documented continually increasing reproductive capacity over 50 generations, associated with highly significant changes in allele frequency at numerous marker loci. Individual marker loci were associated with additive effects on traits such as number of kernels per plant, number of tillers per plant, kernel size, kernel weight, harvest index, vegetative biomass, plant height, leaf area, spike length, and awn length. Superior reproductive capacity (such as num-
68
ANDREW H. PATERSON et ul.
ber of kernels) was usually associated with the most common allele at each marker locus. Significantinteractionsbetween loci were found in a number of cases. Finally, different genetic marker alleles were associated with adaptation to different environments, even “environments” separated by as little as 1 m (Allard et al., 1972). These experiments have demonstrated, in extraordinary detail, that genetic adaptation involves selection at individual loci, as well as combinations of loci, and occurs differently in different environments. Regarding artificial selection, selection for high grain yield in four maize populations by two different strategies(full-sibfamily selection, and reciprocal recurrent selection)has been associated with changes in allele frequency at eight different enzyme (isozyme) loci (Stuber et al., 1980). Four of the eight isozymes showed similar effects of selection in all of the populations studied. Further, the greatest shifts in allele frequency were found in the populations which showed the greatest yield improvement. While no direct test was performed demonstrating that genes near these loci influence yield, the changes observed were statistically unlikely to have arisen by chance. The experiment indicated that selection for a complex trait such as yield was associated with many genetic changes at different places throughout the genome. These studies are two of many such cases which have provided much basic information on microevolution, and on the genetic structure of plant populations under selection. Such studies emphasizethat responseto selection for reproductive capacity or for agricultural productivity is complex, involving many genes and interactions between genes. Further, these studies reveal the genetic complexity which might be expected in various wild or synthetic populations carrying particular traits of interest to a breeder. How might the breeder most effectively extract these traits? Can this process be facilitatedby genetic markers? Given a complex population with superior average expression of a trait, the breeder might wish to ( 1) maintain or improve the expression of the trait of interest, while maintaining desirable levels of other traits; and (2) maintain sufficient genetic diversity that rare desirable alleles influencing the trait(s) of interest are not lost before their frequency can be increased by selection. Genetic markers might be especially valuable in accomplishing this second objective;for example, one might select a fraction of the population based on favorable phenotype (perhaps for several traits- one might readily employ index selection), then apply genetic markers to this fraction and keep a subset which represent much of the allelic diversity within the population. Most natural populations and many synthetic populations are genetically complex. Genetic markers have provided a glimpse at the nature of changes
DNA MARKERS IN PLANT IMPROVEMENT
69
which underlie the continuous process of phenotypic selection,both natural and artificial. Strategies for extracting a maximum of desirable phenotypic variation from such populations remain an important area of breeding research. An integrated approach, merging classical phenotypic selection with a genetic marker-based analysis, may aid in extracting valuable genes from heterogeneous populations. Again, strategies for studying heterogeneous populations would be of special value in species of large stature or long generation time, where one might gain access to heterogeneous natural or cultivated populations, rather than waiting for two generations of crossing to produce ideal mapping populations.
VIII. MARKER-FACILITATED STUDY OF POLYPLOIDS Unlike the diploid species we have emphasized thus far, several agriculturally important species are polyploid, containing more than two copies of each chromosome. Some polyploids, called allopolyploids, are similar to diploids in segregation and recombination. Allopolyploids contain several sets of chromosomes which are derived from a common ancestor, but have diverged enough to be readily distinguishable and thus, usually pair with only one partner during sexual reproduction. Examples of allopolyploids include wheat (6x), oat (6x), soybean (4x), cotton (4x), tobacco (44,and rapeseed (4-4. Other polyploids, called autopolyploids, are very different from diploids in segregation and recombination. Autopolyploids contain several sets of chromosomes which are not readily distinguishable.Thus, a particular chromosome may pair with different partners in different regions. Some examples are potato ( 4 4 , alfalfa (4x3,and banana (3x or greater). Genetics and gene mapping in autopolyploids is complicated by the presence of several copies of each chromosome. Rather than being limited to having 0, 1, or 2 copies of an ancestral gene like diploids, polyploidscan have from 0 to x copies (x being the number of copies of each chromosome). Hence, study of gene dosage effects, and also of epistasis, is much more complicated than in diploids. Further, study of recombination is quite complicated in autopolyploids, because of multiple opportunities for chromosome pairing. Finally, one seldom has enough genetic markers to distinguish each of (for example)four homologous chromosomes;more commonly, one chromosome can be distinguished from the other three. Consequently, only a fraction of the total number of recombination events can be studied using genetic markers; thus, one must study more individuals to compensate for this loss of information.
70
ANDREW H. PATERSON et a'. DNA Band Parent Probe no. P1 P2
Progeny 1 2 3
4
5
6
7
8
9
10 1 1 1 2
Fig. 8. Identificationof singledose restriction fragment(SDRF) polymorphism (Wu eta/., 1991). Hypothetical patterns of two DNA probes, a and b, on both parents (P, and Pz)and their progeny. Fragments 2, 3, 5, and 6 are all polymorphic between the parents. However, only fragments 2 and 5 are SDRFs, as confumed by a 1 : 1 ratio of presence versus absence in the progeny, indicating that the polymorphism is present on only a single chromosome in parent 1. Fragments 3 and 6 do not segregate 1 :1, indicating that the polymorphism is present on more than one chromosome in parent 1.
One method of linkage analysis in polyploids (Wu et al., 1991) involves the identification of bands that represent a single-doserestriction fragment (SDRF) (Fig. 8). In this approach, each restriction fragment is analyzed for its presence or absence in the progeny. If the fragment is represented by a single dose, it will segregate 1 : 1 (presence :absence)in the gametes or 3 : 1 in the selfed progeny, regardlessof whether the speciesis allo- or autopolyploid. Single-dose restriction fragment markers in autopolyploids represent the genotypes of one parent or the other, and thus do not discriminate between individuals which are heterozygous from those which are homozygous for the parental fragment (unlike codominant markers used in the study of diploids; see Section 11). Estimation of recombination distance between two SDRF markers is similarto that for diploids, except that the case of repulsion phase linkage (presence of a marker at one locus Linked to absence of a marker at a nearby locus) is very difficult to detect using SDRFs. The principle of SDRF markers was used to construct a linkage map of potato (Bonierbale et al., 1988) from a cross between heterozygous diploids.* Population or family sizes required for detecting SDRFs and their Linkage have been estimated by Wu et al. ( 1991). If one considersa crossbetween two individuals, one of which has a particular band that is lacking in the other, the first step is to test the hypothesis that the band represents a SDRF (not a multiple-dose band). This is done by studying segregation of the band in the
* However, we note that in diploids, identificationof SDRFs is not essential, since nearly all fragments are single dose, and virtually all genotypes can be identified.
DNA MARKERS IN PLANT IMPROVEMENT
71
progeny: single-dose bands will segregate at 1 : 1 (presence:absence), while double-dose markers will segregate at 3 : 1 or greater (Fig. 8). Therefore, the population size must be large enough to distinguish between 1 : 1 and 3 : 1 segregation ratios, at a particular level of probability. If a population of 75 individuals is used, the inadvertent selection of multiple dose bands will occur less than 1% of the time (Wu et al., 1991). Population size is also important for determining linkage distance. If a population size of 75 is chosen for identifyingSDRFs, it will be adequate for detecting approximately 1 centimorgan (cM) (about 1% recombination) linkages with a 99% level of confidence (with the exception of repulsion phase linkage in autopolyploids) (Wu et al., 1991). Maximum detectable recombination fraction for repulsion in autopolyploids decreases as the number of chromosomes increases. For a population of 75, only linkages of 15 cM or less can be detected in autotetraploidsand none can be identified at higher ploidy. Because of this large difference between allo- and autopolyploids in detection of repulsion phase linkage, a comparison of the number of coupling and repulsion phase linkages in a species with unknown pairing behavior will provide an indication of the degree of preferential pairing among chromosomes in each linkage group; e.g., an indication of whether the species is allo- or autopolyploid. In view of the complicated genetics associated with autopolyploids, it comes as little surprise that the genetic maps of these species are as yet less detailed than those for many diploids and allopolyploids.At the same time, the inherent complexity of polyploid genetics may add even greater value to the simplicity achieved by identifying markers for a trait. Markers might permit the breeder to identifyindividuals carrying different desirable genes, and help him/her to select transgressive progeny from crosses between such individuals. Breeding objectives which have classically been cumbersome, relying heavily upon phenotypic selection among random-mated individuals of unknown genotype, might be accomplished more rapidly and efficiently with the information added by DNA markers.
IX. UTILIZATION OF EXOTIC GERMPLASM IN CROP IMPROVEMENT Wild relatives of crop speciesare a rich source of valuable traits (Stephens, 1961; Harlan, 1976; Hawkes, 1977; Stalker, 1980; Rick, 1982; Brown et al., 1985;Goodman et al., 1987; Vaughan, 1989),from which we have realized only a small fraction of the potential gains to crop improvement. The range of genetic variation for a trait is often much greater in exotic germplasm than
72
ANDREW H. PATERSON e t al.
among cultivated types, as cultivated types are usually derived from a small number of ancestors, and have been selected intensely over many centuries (Hawkes, 1977; Vaughan, 1989). Consequently, the current breeding pool for many crops, especially self-pollinated species (such as wheat, rice, soybean, cotton, and tomato), contains only a small fraction of the existing genetic variation. Many simply inherited traits such as disease and insect resistances (Harlan, 1976), and even quantitative traits such as elevated soluble solids in tomato (Rick, 1974), have been extracted from exotic germplasm. In addition to agriculturally valuable traits, wild species exhibit many properties which are important to survival in their natural environment, but are illsuited to agricultural productivity (Stephens, 1961; Harlan, 1976). Some examples of such properties are seed dormancy, shattering, tall stature, excessive vegetative growth, unpleasant flavor or odor, toxins, thorns, small seedy fruit, and nonuniform maturity. Breeders are often reluctant to make crosses with exotic germplasm, for fear that such undesirable traits will be difficult to separate from valuable attributes. An otherwise outstanding crop variety can be condemned by even small departures from the rigorous standards of the producer, processor, or consumer; hence, conservativebreeding strategies have traditionally been the key to success. A consequence of this demand of the agricultural marketplace for conservatism is the ever-shrinking genetic diversity in many crop species, and ever-greater need for making wide crosses to access new genetic variation. Such gene introgression experiments are well suited to use of genetic markers, and may ultimately yield substantial contributions to agricultural productivity. Stephens ( 1961) eloquently summarized the hazards of interspecific breeding, by classical techniques: When the chromosomes of different species are sufficiently alike to recombine more or less freely their recombinant products are likely to be greatly inferior to the parental combinations. At present there is no way of controlling this total disruption of the parental genotypes which results from segregation.
Genetic markers provide a means of controllingthe degree of disruption of the parental genotype. One can use genetic markers to determine which chromosomal regions carry agriculturally desirable genes, and also to reveal which individuals carry those chromosomal regions. One can then restore most of the domestic parent genotype (and consequently, phenotype) by backcrossing, while retaining the desired genes (and traits) from the wild parent by selectingfor nearby genetic markers. Genetic markers may even be used to accelerate the restoration of the domestic parent’s attributes, by
DNA MARKERS IN PLANT IMPROVEMENT
73
selecting against markers from the wild parent outside the regions carrying target genes (Fig. 9). (Tanksley, 1983a; Burr et al., 1983; Soller and Beckmann, 1988; Paterson et al., 1988, 199a,b; Young and Tanksley, 1989). This strategy seems simple. However, one frequently encounters complications. The chromosomal segments or linkage blocks which are readily transferred by backcrossing tend to be quite large, including not only the target genes but also many additional genes (Hanson, 1959; Zeven et al., 1983). Many such genes from wild species have effects which are agriculturally undesirable. Normally, such undesirable linkages might be overcome by recombination, which acts to “homogenize” chromosomes. However, recombination appears to occur less frequently between chromosomal segments which are so different from one another as those from wild and domestic species(Rick, 1969;Paterson et al., 1990a);thus, the homogenization appears to be least effective in wide crosses,just where it is needed the most. As a consequence, rather large chromosomal segments from the wild donor may remain intact for many generations (Young and Tanksley, 1989),coupling valuable traits to undesirable traits. Numerous examples of such associationshave been found. In our own studies, some cases of association between improved soluble solids concentration and reduced fruit size in tomato have been shown to represent such linkages (Paterson et al., 1990a). As discussed in Section IV (Pleiotropy),genetic markers aid in determining when such trait associationsare due to effects of different genes (Paterson et al., 1990a).One can then use the same markers to identifl rare individuals in which the associations are broken by recombination. Rather than performing many backcrosses and slowly trimming an introgressed chromosomal segment, one might use genetic markers to screen a large population, and obtain the desired recombinants in fewer generations (Tanksley, 1983b). Thus, the disruption of the parental genotype (Stephens, 1961) in crosses with exotic germplasm might be minimized, maintaining much of the desirability of the adapted cultivar while adding valuable traits from exotic sources. The potential of these new tools for utilizing wild germplasm can only be realized if the germplasm is readily available. Although extensive collections of certain species of wild plants have been made, many species are less extensively collected, and in most species there are surely novel variants not yet found. Further, facilities for preserving and propagating these precious heirlooms operate under limited budgets and with minimal staff and thus are barely able to maintain viable stocks of the existing collections, not to speak of adding to the collections or beginning to characterize individuals in the collections. The potential value of wild germplasm, and the substantial risk of extinction of as yet uncollected types in some areas of the world, create an
Lycopersicon esculenturn
Lycopersicon chmielewskii
v BC1: Plant #135
BC2Fl: Plant # 135-5-11
#135-5-11-768-13
~
U I
Fig. 9. Use of restriction fragment length polymorphisms (RFLPs) to eliminate undesirable regions of donor genome. Quantitative trait loci (QTL) mapping showedthat a segment of chromosome 1 from Lycopersicon chrnielewskii (indicated by 6lled chromosomes) was valuable in elevating the soluble solids concentration of Lycopersicon esculentum (tomato; indicated by open chromosomes) (Patemon et al., 1990a). Using RFLPs to identify chromosome segments from L. chrnielewskii, one plant in the first backcross generation (#I 35) was found which carried only five undesirable segments of L. chmielewskii genome, in addition to the desired region of chromosome 1. M e r two additional backcrosses to L. esculenturn, using RFLPs to select against these undesirable segments, a single plant (#135-5-11-768-13) was found which was carrying only the segment of interest.
DNA MARKERS IN PLANT IMPROVEMENT
75
urgent need for accelerating the collection and characterization of wild plants.
X. REVEALING EVOLUTIONARY RELATIONSHIPS AMONG CROP SPECIES AND THEIR WILD RELATIVES -UTILITY IN COMPARATIVE GENETIC MAPPING An important use of genetic markers has been to attempt to discern evolutionary relationships, within and between species, genera, or larger taxonomic groupings. Such studies involve studying similarities and differences among taxa, using numerous genetic markers. Although phylogenetic trees have previously been established for many species on the basis of visible and isozyme markers and chromosome homology (e.g., tomato; Rick, 1979;wheat; Riley, 1965;cotton; Beasley, 1942; Kimber, 1961; Phillips, 1962), DNA markers have recently added to the breadth of phylogenetic information available for a number of species (Song et al., 1988, 1990; Galau et al., 1988; Wendel, 1989; Debener et al., 1990; Miller and Tanksley, 1991; Wang and Tanksley, in preparation; Goffreda and Sorrells, in preparation). Such studies are not only of academic interest, but are also important in classifying newly discovered germplasm, and in establishing possible sources from which valuable traits might readily be transferred to crop species. This latter objective also relies heavily upon studies of crossability among related species (see, e.g., Beasley, 1942; Rick, 1979; Dewey, 1984). In a few cases, it has been possible to reveal the consequences of evolutionary divergence on chromosome organization in crop species. For example, Brassica campestris and B. oleracea show a high degree of conservation of linkage arrangement (Slocum, 1989). The chromosomes of tomato are remarkably similar to those of potato. Based on mapping of 134 common markers in the two species, seven chromosomes showed no detectable rearrangement, and the remaining five showed a total of seven paracentric inversions (Bonierbale et al., 1988). Pepper and tomato, which are more distantly related, retain homology to many common cDNA probes, but differ by a larger number of rearrangements (Tanksley et al., 1988; Prince and Tanksley, 1991). These consequences of evolutionary divergence are of academic interest, and also have much practical value. By defining the sites of chromosomal rearrangement, one also defines intervening regions in which genes are
76
ANDREW H. PATERSON et al.
arranged similarly in different organisms. This might permit extrapolating results from tomato to pepper (Tanksley et al., 1988), or from human to bovine (Womack et al., 1990), or vice versa. Hence, exhaustive study of tomato (or human) may help to fill in the gaps left by less detailed study of pepper (or cow), using this technique of comparative mapping. To a lesser degree, comparative mapping may also be possible in regard to QTLs. A recent study (Paterson et al., 1990b) compared the locations of QTLs in two different wild species of tomato, L. chmielewskii and L. cheesmanii, which are only distantly related (Rick, 1979; Miller and Tanksley, 1991) but are similar in having very small fruit with high soluble solids. About half of the QTLs mapped in the two species fell at similar chromosomal locations (8 of 16 possible matches), suggestingthat some of the same genetic factors influence quantitative traits in the two distantly related species. More experiments are needed to see how broadly this result applies, to other speciesand other traits. However, it suggeststhat QTL mapping information from one pedigree might be somewhat predictive of QTL locations in other pedigrees, or in this particular case, other species.
XI. CLONING GENES FROM MAP POSITION? The sophisticated molecular tools and techniques available for studying individual genes are of much potential use to the crop breeder. Some of this potential can be realized by marker-facilitatedselection, that is, using linked DNA markers to select for nearby genes which are difficult to manipulate by other means. However, new strategies for crop improvement might result, if one could isolate genes relevant to particular components of agricultural productivity. Because the biochemical function of such genes is usually unknown, the elegant strategies which exist for cloning genes of known function are of little help to the breeder. Further, because measures of agricultural productivity are influenced by many genes, together with environment, one cannot readily identifyindividuals in which only one of the many genes has been turned off, by insertion of an easily identified DNA element (Federoff et al., 1984; Cone el al., 1986; Schmidt et al., 1987; Feldmann et al., 1989). One piece of information the breeder can readily obtain is the chromosomal location, or map position of genes affecting agricultural productivity (Section IV,A). However, as has been previously discussed, crop plants have perhaps 10,000to 100,000genes, scattered through a total of lo8- loLobp of DNA. How does one fish out a single gene (- lo3bp) responsible for an incremental improvement in an end-point measurement, while ignorant of
DNA MARKERS IN PLANT IMPROVEMENT
77
the biophysical basis of the change? In other words, how does one achieve sufficient resolution of map position to clone a gene? In most systems, map-based cloning is presently a long-term undertaking that involves enormous expenditures of time and money, and has only reached fruition in a few cases of single genes with striking effects (RoyerPokora et al., 1986;Rommens et al., 1989;Fearon et al., 1990;Gessler et al., 1990; Wallace et al., 1990). Map-based cloning of some single genes in crop plants may well be accomplished in the next few years, but new genetic stocks, new molecular tools, and new laboratory technology will be necessary before map-based cloning can become a routine exercise. However, the rewards may be substantial; the capability of cloning genes by map position would help to bring the tools of molecular biology to bear on important practical questions which are presently limited to classical approaches. Obtaining the tools and information necessary to clone agriculturally important genes will be expedited by cooperation among several disciplines, each of which can provide important information at different steps in the process (Fig. 10). The agronomist, physiologist, pathologist, and/or entomologist can contribute by identifying properties of the whole plant (or perhaps plant cell) which affect agricultural productivity, as well as biotic and/or abiotic environmental conditions necessary to optimize detection of those properties. The breeder makes crosses between genotypes which differ in these properties, and shows that somewhere among the lo8- lolo bp which comprise the genome of a typical crop plant, there are genes responsible for variation in the relevant properties. By backcrossing, with selection for the trait to make near-isogenic stocks, the location of a gene might be delimited to a fraction of a chromosome, perhaps 1 -2% of the genome (Hanson, 1959). The cytogeneticist,employing genetic stocks with abnormal numbers of chromosomes or chromosome arms, might obtain similar resolution somewhat more quickly than could be achieved by backcrossing (Khush, 1973; Riley and Law, 1983). At this point, the geneticist enters the picture. Given near-isogenic stocks, DNA markers can be used to identify the 1 -2% of the genome carrying a gene of interest (Young et al., 1988; Tanksley and Hewitt, 1988). Given a segregatingpopulation, DNA markers can be used to map either single genes or QTLs to a portion of a chromosome, or perhaps 1-2% of the genome (Paterson et al., 1988).With additional generations of crossing, the location of a gene can be narrowed down to perhaps 0.1% of the genome (Paterson et al., 1990a). Hence, in trying to identify a particular lo3 bp of DNA from lo9 bp, DNA markers can narrow the field ofcandidates by 99.9%. This still leaves a stretch of at least lo6 bp to be dissected, a daunting task. The molecular biologist may ultimately be able to provide the next steps. In yeast, it is possible to construct artificial chromosomes (Burke et al.,
78
ANDREW H. PATERSON et a[.
} Maize genome }
Tomato genome
}
Tomato chromosomes
}
QTL likelihood mapping
}
QTL substitution mapping
}
Yeast artificial chromosomes
}
P 1 bacteriophage
100,000,000
10,000,000 1,000,ooc 100,000
1 Cosmids
10,000 1000
} Lambda bacteriophage )
Bacterial plasmids
}
M13 bacteriophage
.AGCTCGTGATGAC
. . .
Fig. 10. Resolution of different methods of gene manipulation, as measured by number of nucleotide bases. Classical plant breeding works at the level of the whole genome, 10"- 10'O nucleotides. Cytogenetics works at the level of individual chromosomes or chromosome arms, each about 5 - 10% of the genome in most crops. Quantitative trait loci (QTL) likelihood mapping (by interval analysis; see Fig. 4b) or isogenic lines, offer slightly finer resolution than cytogenetic stocks, perhaps one-fifth of a chromosome, or about I -2% of the genome. Quantitative trait loci (QTL) substitution mapping (Paterson et aL, 1990a) narrows the location of a QTL down to perhaps 1/50- 1/100 of a chromosome, or about 0.1% of the genome. Yeast artificial chromosomes(YACs)can carry an intact DNA segment of lo5- 106nucleotides,about 0.001 -0.1% of the genome. Such a DNA segment can then be subdivided into smaller pieces, in Pl phage (Sternberg, 1990), and/or smaller cloning vectors such as cosmids, lambda phage, plasmids, and M 13 phage (Sambrook et al., 1989). Finally, pieces measuring 300 nucleotides can be sequenced, to determine the identity of each nucleotide in the region.
-
DNA MARKERS IN PLANT IMPROVEMENT
79
A
B t
I C
Fig. 11. A strategy for cloning a gene, known only by its effect on phenotype, using “chromosome walking.” Starting with the knowledge that a gene lies between two mapped DNA clones [(A), from Fig. 51, one restriction fragment length polymorphism (RFLP) clone can be used to identify a YAC which includes the clone, by screening a large number of candidate YACs (usually a library of lo5or more). A technique called inverse PCR (Ochman et al., 1988) can be employedto obtain a small bit of DNA lying at the end of aYAC.This terminal DNA is then used as a probe to identlfy a second YAC,and the cycle is repeated until a YAC is identified which includes the second RFLP. In practice, one seldom has the luxury of knowing which direction to walk in, and OCcasionaUy the required YAC is not present in the library for any number of reasons;thus, this walkingexperiment is invariably more difficultthan it sounds, In principle, however, all the genomic DNA which Lies between the two starting DNA clones might be identified. In another laborious experiment, all of this genomic DNA can then be introduced, a small bit at a time (C), into plants lacking the trait ofinterest (Schell, 1987). Those are inferred to harbor the target gene, bits which cause the plant to exhibit the trait of interest (D) and may be dissected further, to study at the level of detail described at the begmning of Section IV.
1987), which can carry > lo5 bp of foreign DNA. By randomly making a large number of yeast artificial chromosomes (YACs) from a genome of interest, one might identify a YAC near the gene of interest, by using the closest FWLP as a probe. Next, one might use the first YAC as a probe, to identify a second YAC, which overlaps with, but also extends beyond, the first. By repeating this “chromosome walking” (Steinmetz et al., 1982)(Fig. 11) several times, a series of YACs might be identified which, lying end to end, span the chromosomal region of interest. Having this, one might use
80
ANDREW H. PATERSON et
a].
DNA vectors which carry smaller bits of recombinant DNA to dissect the region (cosmids, phages, plasmids, see Fig. lo), and perhaps introduce bits of it into some control genotypes which do not express the trait (Fig. 11). Individuals which subsequentlyexhibit the trait have presumably received a DNA segment carrying the gene of interest. With this tiny bit of DNA in hand, an extensive repertoire of molecular techniques might be applied to describe the genetic basis of the trait in exacting detail, and perhaps even alter the trait by either exaggerating it or eliminating it. This scheme is somewhat idealized, and will no doubt be improved upon by new technologies or new genetic strategies. The principle of map-based cloning is currently being applied in several favorable systems. However, the ability to clone the anonymous genes selected for by even a single breeder working on one crop in one season, lies far into the future, if ever. Nonetheless, the association of agriculturally important genes with specific DNA markers provides an important medium for communicating information between production agriculture and molecular biology. By integrating the strengths of a number of diverse biological disciplines, new developments in molecular biology may be most expedientlybrought to bear on questions of agricultural productivity.
XII. CHALLENGES FOR THE FUTURE Plant breeding is widely recognized as making a substantial contribution to long-term improvement of agricultural productivity. We consider genetic markers to be a potentially valuable tool in accomplishing many plant breeding objectives. Marker-facilitatedbreeding is already showing its value in specializedbreeding experiments, such as introgressionof traits from wild species, and may stimulate interest in this important (but often difficult) undertaking. The broader impact of genetic markers on the overall process of crop improvement will likely be influenced by many factors, such as: 1. The genetic basis of variation among elite varieties. A superior variety is often distinguished from its lesser brethren by tiny differences in phenotype. Such small differences might prove difficult to map, in that an enormous number of individuals would have to be studied (Lander and Botstein, 1989). Further, these differences could be partly due to epistasis, which has proven elusive in mapping experiments (see above). Carefully designed mapping experiments using a small number of elite crosses might offer valuable information on the genetic basis of variation among elite breeding lines.
DNA MARKERS IN PLANT IMPROVEMENT
81
2. The availability of diverse germplasm, with known properties. Wild speciesand feral accessions of crop species hold much of the existing genetic variation in plant species (Hawkes, 1977; Rick, 1982; Brown et al., 1985; Vaughan, 1989). In some species, extensive germplasm collections have been accumulated, while other species have been less thoroughly collected. Further, much of the collected germplasm has been characterized only minimally, if at all. We believe that DNA markers are an especially valuable tool for introgression of desirable traits from exotic germplasm into elite cultivars. However, the germplasm must be available, and its attributes must be known. Thus, collection and characterization of exotic germplasm continues to be a priority, and is made ever more urgent as human activitiesrapidly alter the global environment. 3. Identifyingbiophysical measures of agricultural productivity. Classical plant breeding methods rely heavily upon end-point measurements of agricultural productivity, which are influenced by different parametersand consequently different genes, in different environments. If more specific measures of agricultural productivity can be identified, for example, physical or chemical properties of the plant which relate directly to productivity under a particular environmental stress, it will be much more feasible to identify the underlying genes by mapping. 4. The ability to identify polymorphic genetic markers in cultivated germplasm. While cross-pollinated species such as maize generally show an abundance of genetic variation, it has proven difficult to find polymorphism among cultivars of many self-pollinated species such as wheat, soybean, and tomato, impeding genetic mapping experiments. New classes of highly polymorphic genetic markers (Jeffreys et al., 1985; Vassart et al., 1987; Rogstad et al., 1988; Dallas, 1988), techniques for detecting tiny genetic differences (Fischer and Lerman, 1983; Myers et al., 1985), and techniques for rapidly screening many candidate markers for polymorphism (Williams et al., 1991), may help in this regard. 5 . Development of tools and techniques for high-resolution analysis of large genomes. Much progress in plant breeding might be made by having markers to more easily select for traits of interest. However, this marker-facilitated selection represents only one facet of how molecular biology might contribute to plant breeding. Cloning of individual genes responsible for variation in agriculturallyimportant traits would undoubtedly contribute to basic biology, and might also create new opportunities for improving agricultural productivity. Tools for physical analysis of large genomes, such as YACs, along with improved strategies for interfacing genetic and physical maps, will be a high priority. 6. Simplification of DNA marker technology. Current procedures for using DNA markers in genetic mapping require considerable laboratory
82
ANDREW H. PATERSON et
a/.
infrastructure, beyond the present capabilities of many breeding stations. Techniques are needed which require a minimum of sophisticated equipment, and are easily scaled up to handle large numbers of individuals. New developments such as nonradioactive detection of DNA probes (Langer et al., 198l), allele-specific DNA probes not dependent on restriction enzymes or electrophoretic separations (Conner et al., 1983), and finally, amplification and direct visualization of particular DNA sequences (Mullis and Faloona, 1987; Williams et d.,1991) are all promising steps toward simpler DNA marker technology.
XIII. CONCLUSION For much of the Twentieth Century, biologists have been aware of the potential that genetic markers held for detailed investigation of complex questions in quantitative biology. The advent of DNA markers, greatly expanding the number of genetic markers available, is allowing researchers to begin to tap the potential of this technology, to the benefit of both basic biology and agricultural productivity. We believe that genetic markers will make a major contribution to the biological sciences, especially agriculture, for the foreseeable future.
REFERENCES Ali, S., Muller, C. R., and Epplen, J. T. 1986. DNA fingerprintingby oligonucleotideprobes specific for simple repeats.Hum. Genet. 74,239-243. Allard, R. W. 1960. “Principles of Plant Breeding.” Wiley, New York. Allard, R. W. 1988. Genetic changes associated with the evolution of adaptedness in cultivated plants and their wild progenitors. J. Hered. 79,225-238. Allard, R. W., Babbel, G. R., Clegg, M. T., and Kahler, A. L. 1972.Evidence for coadaptation in Avena barbata. Proc. Natl. Acad. Sci. U.S.A.69,3043-3048. Apuya, N. R., Frazier, B. L.,Keim, P., Roth, E. J., and Lark,K. G. 1988. Restriction fragment length polymorphisms as genetic markers in soybean, Glycine mar (L.) merrill. Theor. Appl. Genet. 75,889-901. Bailey, D. W. 1981. Recombinant inbred strains and bilineal congenic strains. In “The Mouse in Biomedical Research” (H. L. Foster, J. D. Small, and J. G. Fox, eds.). Academic Press, New York. Baker, R. J. 1986. “Selection Indices in Plant Breeding.” CRC Press, Boca Raton, Florida. Beasley, J. 0. 1942. Meiotic chromosome behavior in species, species hybrids, haploids, and induced polyploids of Gossypium species. Genetics 27,25 -54. Blake, T. 1990. Barley (Hordeum vulgare).In “Genetic Maps” (S. J. OBrien, ed.), 5th Ed. Cold Spring Harbor Press, Cold Spring Harbor,New York. Bonierbale, M. W., Plaisted, R. L.,and Tanksley, S. D. 1988. RFLP maps based on a common
DNA MARKERS IN PLANT IMPROVEMENT
83
set of clones reveal modes of chromosomal evolution in potato and tomato. Genetics 120, 1095- 1103. Borlaug, N. E. 1983. Contributions of conventional plant breedingto food production. Science 219,689-693. Botstein, D., White, R. L., Skolnick, M., and Davis, R. W. 1980. Construction of a genetic linkage map in man using restriction fragment length polymorphisms.Am. J. Hum. Genet. 32,314-331. Boyer, J. S . 1982. Plant productivity and environment. Science 218,443-448. Breese, E. L., and Mather, K. 1957.The organisationof polygenicactivitywithin a chromosome in Drosophila. I. Hair characters. Heredity 11,373-395. Bridges, C. B. 1935. Salivary chromosome maps. J. Hered. 26,60-64. Brim, F. N. 1938. The use of the backcross in crop improvement. Am. Nut. 72,285-292. Brim, C. A. 1966. A modified pedigree method of selection in soybeans. Crop Sci. 6,220. Brown, A. H. D., Grant, J. E., Burdon, J. J., Grace, J. P., and Pullen, R. 1985. Collection and utilization of wild perennial Glycine. World Soybean Rex ConJ III: Proc. (R. Shibles, ed.). Westview Press, Boulder, Colorado. Brown, K. D., Sorrells, M. E., and C o h a n , W. R. 1983. A method for classification and evaluation of testing environments. Crop Sci. 23,889-893. Bruce, A. B. 1910. The Mendelian theory of heredity and the augmentation of vigor. Science 32,627-628. Burke, D. T., Carle, G. F., and Olson, M. V. 1987.Cloning of largesegmentsof exogenousDNA into yeast by means of artificial chromosome vectors. Science 236,806 - 8 12. Burr, B., Evola, S. V., Burr, F. A., and Beckmann, J. S. 1983. The application of restriction fragment length polymorphism to plant breeding. In “Genetic Engineering” (J. K. Setlow and A. Hollaender, eds.), Vol. 5. Plenum, New York. Chalef, R. S. 1983. Isolation of agronomically useful mutants from plant cell cultures. Science 219,676-682. Chao, S.,Sharp, P. J., Worland, A. J., Warham, E. J., Koebner, R. M. D., and Gale, M. D. 1989. RFLP-based genetic maps of wheat homoeologous group 7 chromosomes. Theor. Appl. Genet. 78,495 - 504. Coe, E. H., Jr., Hoisington, D. A., and Neuffer,M. G. 1990.Linkage map ofcorn (MAIZE)(Zea mays L.) (2n = 20). In “Genetic Maps” (S. J. OBrien, ed.), 5th Ed. Cold Spring Harbor Press, Cold Spring Harbor, New York. Cone, K. C., Burr, F. A,, and Burr, B. 1986. Molecular analysis of the maize anthocyanin regulatory locus CI. Proc. Natl. Acad. Sci. U.S.A. 83,9631 -9635. Comer, B. J., Reyes, A. A., Morin, C., Itakura, K., Teplitz, R. L., and Wallace, R. B. 1983. Detection of sickle-cell /F-globin allele by hybridization with synthetic oligonucleotides. Proc. Natl. Acad. Sci. U S A . 80,278-282. Cornu, A., Farcy, E.,Maizonnier, D., Haring, M.,Veerman, W., and Gerats, A. G. M. 1990. Petunia hybrida (2n = 14). In “Genetic Maps” ( S . J. OBrien, ed.), 5th Ed. Cold Spring Harbor Press,Cold Spring Harbor, New York. Cullis, C. A., Gorman, M. B., Bader, J., Alldridge,N., and Alldridge,K. J. 1990. RFLP mapping in flax,with particular reference to isolation of flax rust resistance genes. J. Cell Biol. 14E, 315. Dallas, J. F. 1988. Detection of DNA “fingerprints” of cultivated rice by hybridization with a human minisatellite DNA probe. Proc. Natl. Acad. Sci. U.S.A.85,6831-6835. Debener, T., Salamini, F., and Gebhardt, C. 1990. Phylogeny of wild and cultivated Sofanum speciesbased on nuclear restriction fragment length polymorphisms(RFLPs). Theor.Appl. Genet. 79,360-368. Dewey, D. R. 1984. The genomic system of classification as a guide to intergeneric hybridization with the perennial Triticeae. Stadler Genet. Symp. 16, 209 - 28 1.
84
ANDREW H. PATERSON et al.
Dobzhansky, T. 1954. Nature and origin of heterosis. In “Heterosis” (J. W. Gowan, ed.). Iowa State Coll. Press, Ames. DonisKeller, H., Green, P., Helms, C., Cartinhour, S., Weiffenbach, B., Stephens, K., Keith, T. P., Bowden, D. W., Smith, D. R., Lander, E.S., Botstein, D., Akots,G., Rediker, K. S., Gravius, T., Brown, V. A., Rising, M. B., Parker, C., Powers, J. A., Watt, D. E., Kauffman, E. R., Bricker, A., Phipps, P., Muller-Kahla, H., Fulton, T. R., Ng, S., Schumm, J. W., Braman, J. C., Knowlton, R. G., Barker, D. F., Crooks, S. M., Lincoln, S. E., M y , M.J.,andAbrahamson, J. 1987.Ageneticlinkagemapofthehumangenome.Ce1151,319337. East, E. M. 1908. Inbreeding in corn. Rep. Conn. Agric. Exp. Stn., 1907 pp. 419-428. East, E. M. 1915. Studies on size inheritance in Nicotiana. Genetics 1, 164- 176. Edwards, M. D., Stuber, C. W., and Wendel, J. F. 1987. Molecular-marker-facilitatedinvestigations ofquantitative trait loci in maize. I. Numbers, genomicdistribution, and types ofgene action. Genetics 116, 113- 125. Everson, E.H., and Schaller, C. W. 1955. The genetics of yield differencesassoCiatedwith awn barbing in the barley hybrid (Lion X Atlaslo)X Atlas. Agron. J. 47,276-280. Fearon, E. R., Cho, K. R., Nigro, J. N., Kern, S. E., Simons, J. W., Ruppert, J. M., Hamilton, S. R., Preisinger,A. C., Thomas, G., Kinzler, K. W., andVogelstein, B. 1990. Identification of a chromosome 18q gene that is altered in colorectal cancers. Science 247,49-56. Federoff,N. V., Furtek, D. B., and Nelson, 0.E. 1984. Cloningof the bronze locus in maize by a simple and generalizable procedure using the transposable controlling element Activator (Ac). Roc. Natl. Acad. Sci. U.S.A. 81, 3825-3829. Fehr, W. 1984. “Genetic Contributions to Yield Gains ofFive Major Crop Plants,” Spec. Publ. No. 7. Crop Sci. Soc.Am., Madison, Wisconsin. Feldmann, K. A., Marks, M. D., Christianson, M. L., and Quatrano, R. S. 1989. A dwarf mutant of Arabidopsis generated by T-DNA insertion mutagenesis. Science 243, 1351 1354. Figdore, S.S., Kennard, W. C., Song, K. M., Slocum, M. K, and Osborn, T. C. 1988. Assessment of the degree of restriction fragment length polymorphism in Brassica. Theor.Appl. Genet. 75,833-840. Fischer, S . G., and Lerman, L. S. 1983. DNA fragments differing by single base-pair substitutions are separated in denaturing gradient gels: Correspondencewith melting theory. Proc. Natl. Acad. Sci. U.S.A. 80, 1579-1583. Freeling, M., and Bennett, D. C. 1985. Maize Adhl. Annu. Rev. Genet. 19,297-323. Frei, 0.M., Stuber, C. W.,and Goodman, M. M. 1986. Use of allozymes as genetic markers for predicting performance in maize singlecross hybrids. Crop Sci. 26,37-42. Galau, G. A., Bass, H. W., and Hughes, D. W. 1988. Restriction fragment length polymorphisms in diploid and allotetraploid Gossypium: Assigning the late embryogenesisabundant (Lea) alloalleles in G. hirsutum. Mol. Gen. Genet. 211,305 - 3 14. Galbraith, D. W., Harkins, K., Maddox, I. M., Ayres, N. M., Sharma, D. P.,and Firoozabady, E. F. 1983. Rapid flow cytometric analysis of the cell cycle in intact plant tissues. Science 220, 1049-1051. Gebhardt, C., Blomendahl, C., Schachtschabel, U., Debener, T., Salamini, F., and Ritter, E. 1989a. Identificationof 2n breeding lines and 4n varieties of potato (Solanum tuberosum. ssp. tuberosum) with RFLP fingerprints. Theor.Appl. Genet. 78, 16-22. Gebhardt, C., Ritter, E.,Debener, T., Schachtschabel, U., Walkemeier, B., Uhrig, H., and Salamini, F. 1989b. RFLP analysis and linkage mapping in Solanum tuberosum. Theor, Appl. Genet. 78,65 - 75. Geldermann, H. 1975. Investigations on inheritance of quantitative characters in animals by gene markers. I. Methods. Theor. Appl. Genet. 46,319-330.
DNA MARKERS IN PLANT IMPROVEMENT
85
Gepts, P., Singh, S., Nodari, N., Garrido, B., and Koinange, E. 1990. Toward an integrated linkage map of common bean (Phaseolus vulgaris L.). J. Cell Biol. 14E, 28 I . Gessler, M., Poustka, A., Cavenee, W., Neve, R. L., Orkin, S. H., and Bruns, G. A. P. 1990. Homozygous deletion in Wilms tumours of a zinc-finger gene identified by chromosome jumping. Nature (London) 343,774. Goodman, R. M., Hauptli, H., Crossway, A., and Knaut, V. C. 1987. Gene transfer in crop improvement. Science 236,48-54. Gono, D. M., Jermstad, K. D., Tauer, C. G., and Neale, D. B. 1989. RFLP mapping of the loblolly pine (Pinus taeda L.) genome. Proc. Hortic. Biotechnol. Symp., Davis, Calif:p. 53. Gowan, J. W. 1954. “Heterosis.” Iowa State Coll. Press, Ames. Gruneberg, H. 1938. An analysis of the “pleiotropic” effects of a new lethal mutation in the rat (Musnowegicus). Proc. R. Soc. London, Ser. B 125, 123- 144. Hanson, W. D. 1959. Early generation analysis of lengths of heterozygous chromosome segments around a locus held heterozygous with backcrossing or selfing. Genetics 44,833837. Harlan, J. R. 1976. Genetic resources in wild relatives of crops. Crop Sci. 16,329-333. Hawkes, J. G. 1977. The importance of wild germplasm in plant breeding. Euphytica 26, 6 15 -621. Hubbard, M., Kelly, J., Abbott, A., and Ballard, R. 1989. Construction ofan RFLP linkage map in roses. Proc. Hortic. Biotechnol. Symp., Davis, Calif:p. 53. Hulbert, S . H., Ilott, T. W., Legg, E. J., Lincoln, S. E., Lander, E. S., and Michelmore, R. W. 1988. Genetic analysis of the fungus, Bremia lactucae, using restriction fragment length polymorphisms. Genetics 120,947-958. Hulbert, S. H., Richter, T. E., Axtell, J. D., and Bennetzen, J. L. 1990. Genetic mapping and characterization of sorghum and related crops by means of maize DNA probes. Proc. Nail. Acad. Sci. U.S.A.87,4251 -4255. JefFteys, A. J., Wilson, V., and Thein, S. L. 1985. Hypervariableminisatelliteregionsin human DNA. Nature (London) 314,67-73. Johanssen, W. 1909. “Elemente der exakten Erblichkeitsllehre.” Fischer, Jena. Jones, D. F. 19 17. Dominance of linked factorsas a means of accounting for heterosis. Genetics 2,466-479. Kam-Morgan, L. N. W., and Gill, B. S. 1990. DNA restriction fragment length polymorphism: A strategy for genetic mapping of the D genome of wheat. Genome 32,724732. Keeble, F., and Pellew, C. I9 10. The mode of inheritance of stature and floweringtime in peas (Pisum sativum). J. Genet. 1,47-56. Keim, P., Shoemaker, R. C., and Palmer, R. G. 1989. Restriction fragment length polymorphism diversity in soybean. Theor. Appl. Genet. 77,786-792. Kesseli, R. V., Paran, I., and Michelmore,R. W. 1990. Genetic linkage map oflettuce (Lactuca sativa, 2n = 18).In “Genetic Maps” (S. J. OBrien, ed.), 5th Ed. Cold Spring Harbor Press, Cold Spring Harbor, New York. Khush, G. 1973. “Cytogenetin of Aneuploids.” Academic Press,New York. Kimber, G. 196 1. Basis of the diploid-like meiotic behavior of polyploid cotton. Nature (London) 191,98- 100. Kornher, S., and Livak, K. J. 1989. Mutation detection using nucleotide analogs that alter electrophoretic mobility. Nucleic Acids Res. 17,7779 - 7784. Lande, R. 1976. The maintenance of genetic variability by mutation in a polygenic character with linked loci. Genet. Res. 26,22 1-235. Lande, R., and Thompson, R. 1990. E5ciency of marker-assisted selection in the improvement of quantitative traits. Genetics 124, 743-756.
ANDREW H. PATERSON et a/.
86
Lander, E. S., and Botstein, D. 1989.Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121, 185- 199. Lander, E. S.,Green, P., Abrahamson, J., Barlow, A., Daly, M. J., Lincoln, S. E., and Newburg, L. 1987.MAPMAKER An interactivecomputer package for constructingprimary genetic linkage maps of experimental and natural populations. Genomics 1, 174- 18I . Landry, B. 1990.RFLP mapping in Brassica spp. CruciferGenet. Coop. Annu. Meet., Ithaca, New York Abstr. Lange, K., and Boehnke, M.1982.How many polymorphic marker geneswill it take to span the human genome? Am. J. Hum. Genet. 34,842-845. Langer, P.R., Waldrop, A. A., and Ward, D. C. 1981.Enzymatic synthesis of biotin-labelled polynucleotides: Novel nucleic acid affinity probes. Proc. Natl. Acad. Sci. U.S.A. 78,
6633-6637.
Law, C.N. 1967.The location of factors controlling a number of quantitative characters in wheat. Genetics 56,445-46 1. Lee, M., Godshalk, E. B., Lamkey, K. R., and Woodman, W. W. 1989.Association of restriction fragment length polymorphisms among maize inbreds with agronomicperformance of their crosses. Crop Sci. 29, 1067- 107 1. Markert, C. L., and Moller, F. 1959. Multiple forms of enzymes: tissue, ontogenetic, and species-specific patterns. Proc. Natl. Acad. Sci. U.S.A. 45,753-763. Martin, B., Nienhuis, J., King, G., and Schaefer,A. 1989.Restriction fragmentlength polymorphisms associated with water use efficiency in tomato. Science 243, 1725- 1728. Mather, K.,and Jinks, J. L. 1971. “Biometrical Genetics.” Cornell Univ. Press,Ithaca, New
York.
McCouch, S. R., Kochert, G., Yu,Z. H., Wang, Z. Y., Khush, G. S., C o h a n , W. R., and Tanksley, S. D. 1988.Molecular mapping of rice chromosomes. Theor.Appl. Genet. 76,
815-829.
Mendel, G. 1865.Versuche iiber Pflanzen-Hybriden. Verh.Naturforsch. Ver.Brunn IV,3-47. (Eng. transl., Harvard Univ. Press, Cambridge, Massachusetts, 1925.) Michelmore, R. W., and Hulbert, S.H. 1987.Molecular markers for genetic analysis of phytopathogenic fungi. Annu. Rev.Phytopathol. 25,383-404. Michelmore, R. W., and Shaw, D. V. 1988.Character dissection.Nature (London)335,672-
673.
Miller, J. D., and Tanksley, S.D. 199I. RFLP analysisof phylogenetic relationshipsand genetic variation in the genus Lycopersicon. Theor. Appl. Genet. (in press). Morgan, T. H. I91 1. Random segregation versus coupling in Mendelian inheritance. Science
34,384.
Morgens, P. H., Cohen, R., Wright, P., and Callahan, A. 1989. Fruit cDNAs detect RFLPs among peach cultivars (Prunus persica). Proc. Hortic. Biotechnol. Symp., Davis, Calif:p. 52. Mullis, K. B., and Faloona, F. A. 1987.Specific synthesis of DNA in vitro Via a polymerasecatalyzed chain reaction. Methods Enzymol. 155,335-351. Myers, R.M., Lumelsky, N., Lerman, L. S., and Maniatis, T. 1985.Detection of single base substitutions in total genomic DNA. Nature (London) 313,495-497. Neale, D. B. 1989. RFLP mapping of loblolly pine (Pinus taeda). Prm. Hortic. Biotechnol. Symp. (Davis, California). Nienhuis, J., Helentjaris, T., Slocum, M., Ruggero, B., and Schaefer, A. 1987.Restriction fragment length polymorphism analysisof loci associated with insect resistancein tomato. Crop Sci. 21,797-803. Nilsson-Ehle, H. 1909.“Kreuzunguntersuchungen an Hafer und Weizen.” Lund. Ochman, H., Gerber, A. S., and Hartl, D. L. 1988.Genetic applications of an inverse polymerase chain reaction. Genetics 120,621 -623.
DNA MARKERS IN PLANT IMPROVEMENT
87
Osborn,T. C., Alexander, D. C., and Fobes, J. F. 1987. Identification of restriction fragment length polymorphisms linked to genes controlling soluble solids content in tomato fruit. Theor. Appl. Genet. 73,350-356. Ozias-Akins, P. 1989. Restriction fragment length polymorphism mapping in apomictic interspecific hybrids of Penniseturn.Proc. Hortic. Biotechnol. Syrnp., Davis, Calif:p. 52. Paterson, A. H., Lander, E. S., Hewitt, J. D., Peterson, S., Lincoln, S. E., and Tanksley, S. D. 1988. Resolution of quantitative traits into Mendelian factors, using a complete linkage map of restriction fragment length polymorphisms. Nature (London) 335,721 -726. Paterson, A. H., Deverna, J. W., Lanini, B., and Tanksley, S. D. 1990a. Fine mapping of quantitative trait loci using selected overlapping recombinant chromosomes, in an interspecies cross of tomato. Genetics 124,735 -742. Paterson, A. H., Damon, S.,Hewitt, J. D., Zamir, D., Rabinowitch, H. D., Lincoln, S. E., Lander, E. S., and Tanksley, S. D. 1990b. Mendelian factors underlying quantitative traits in tomato: Comparison across species, generations, and environments. Genetics (in press). Phillips, L. L. 1962. Segregation in new allopolyploidsof Gossypium. N.segregation in New World X Asiatic and New World X wild American hexaploids. Am. J. Bot. 49, 5 1-57. Rick, C. M. 1969. Controlled introgression ofchromosomes ofSolanurnpennelliiinto Lycopersicon esculentum: segregation and recombination. Genetics 62,753 -768. Rick, C. M. 1974. High soluble-solidscontent in large-fruited tomato lines derived from a wild green-fruited species. Hilgardia 42,493-510. Rick, C. M. 1979. Biosystematicstudies in Lycopersicon and closelyrelated speciesofSolanurn. In “The Biology of the Solonaceae”(J. C. Hawkes, R. N. Lester, and A. D. Skelding, 4s.). Academic Press, New York. Rick, C. M. 1982. The potential of exotic germplasm for tomato improvement. In “Plant Improvement and Somatic Cell Genetics” (I. K. Vasil, W. R. Scowcroft, and K. J. Frey, eds.). Academic Press, New York. Rick, C. M., and Fobes, J. F. 1974.Association ofan allozyme with nematode resistance. Rep. Tomato Genet. Coop. No. 24, p. 25. Riley, R. 1965. Cytogeneticsand evolution of wheat. In “Essays on Crop Plant Evolution” (J. B. Hutchinson, ed.), Cambridge Univ. Press, Cambridge, England. Riley, R., and Law, C. N. 1983. Chromosome manipulation in plant breeding: progress and prospects. Stadler Genet. Symp. 16,301 -322. Rogers, H., Dotson, R., Eldredge, L., Ballard, R., Baird, V., and Abbott, A. 1989. Construction of a saturated RFLP linkage map in peaches. Proc Hortic. Biotechnol.Syrnp.,Davis, Calif: p. 52. Rogstad, S. H., Patton, J. C., 111, and Schaal, B. A. 1988. M13 repeat probe detectes DNA minisatellite-like sequences in gymnosperms and angiosperms. Proc. Natl. Acad. Sci. U.S.A.85,9176-9178. Rommens, J. M., Iannuzzi, M. C., Kerem, B., Drumm, M. L., Melmer, G., Dean, M., Rozmahel, R., Cole, J., Kennedy, D., Hidaka,N., Zsiga, M., Buchwald, M., Riordan, J. R., Tsui, L.-C., and Collins, F. S. 1989. Identification of the cystic fibrosis gene: Chromosome walking and jumping. Science 245, 1059- 1065. Royer-Pokora,B., Kunkel, L. M., MOMCO,A. P., Goff, S. C., Newburger, P. E., Baehner, R. L., Cole, F. S.,Curnutte, J. T., and Orkin, S.H. 1986.Cloningthe gene for an inherited human disorder-chronic granulomatous disease-on the basis o f its chromosomal location. Nature (London) 322,32-38. Russell, W.A., and Eberhart, S. A. 1970. Effects of three gene loci in the inheritance of quantitative characters in maize. Crop Sci. 10, 165- 169. Sambrook, J., Fritsch, E. F., and Maniatis, T. 1989. “Molecular Cloning: A Laboratory Manual,” 2nd Ed. Cold Spring Harbor Press, Cold Spring Harbor, New York.
88
ANDREW H. PATERSON et al.
Sax, K. 1923. The association of size differences with seed-coat pattern and pigmentation in Phaseolus vulgaris. Genetics 8, 552- 560. Schell, J. 1987. Transgenic plants as tools to study the molecular organization of plant genes. Science 237, 1176- 1182. Schmidt, R. J., Burr, F. A., and Burr, B. 1987. Transposon taggingaud molecular analysisofthe maize regulatory locus opaque-2. Science 238,960-963. Sharp, P., Chao, S., Desai, S., and Gale, M. D. 1989. The isolation, characterization, and application of the Triticeae of a set of wheat RFLP probes identifyingeach homoeologous chromosome arm. Theor. Appl. Genet. 78,342 - 348. Shull, G. H. 1908. The composition of a field of maize. Rep. Am. Breed. Assoc. 4,296-301. Shull, G. H. 191 1. The genotypes of maize. Am. Nut. 45,234-252. Shull, G. H. 1914. Duplicate genes for capsule form in Bursa bursa-pastoris. Z.I.A.V. 12, 97-149. Simmonds, N. W. 1979. “Principles of Crop Improvement.” Longman, New York. Slocum, M.K. 1989. Analyzing the genomic structure of Brassica speciesand subspeciesusing RFLP analysis. In “Development and Applications of Molecular Markers to Problems in Plant Genetics” (T. Helentjaris and B. Burr, eds.). Cold Spring Harbor Press, Cold Spring Harbor, New York. Slocum, M.K., Figdore, S. S., Kennard, W., Suzuki, J. Y., and Osborn,T. C. 1990. RFLP linkage map of Brussicu oleracea(2n = 18). I n “Genetic Maps” (S. J. OBrien, ed.), 5th Ed. Cold Spring Harbor hess, Cold Spring Harbor, New York. Smith, H. H. 1937. The relation between genes affecting size and color in certain species of Nicotiana. Genetics 22, 36 1. Smith, J. S. C. 1984. Genetic variability within U.S. hybrid maize Multivariate analysis of allozyme data. Crop Sci. 24, 1041- 1046. Smith, J. S. C. 1988. Diversity of United States hybrid maize germplasm; Isozymic and chromatographic evidence. Crop Sci. 28,63-69. Smith, J. S. C., Goodman, M. M., and Stuber, C. W. 1985. Genetic variability within US. maize germplasm. 11. Widely-used inbred lines 1970 to 1979. Crop Sci. 25, 68 1 685. Soller, M., and Beckmann, J. 1983. Genetic polymorphism in varietal identification and genetic improvement. Theor. Appl. Genet. 67,25 -33. Soller, M., and Beckmann, J. 1988. Genomic geneticsand the utilization for breeding purposes of genetic variation between populations. Proc. Znt. Con$ Quant. Genet. 2nd (B. S. Weir, M. M. Goodman, E. J. Eisen, and G. Namkoong, eds.). Sinauer, Sunderland, Massachusetts. Song, K. M.,Osborn, T. C., and Williams, P. H. 1988. Brassica taxonomy based on nuclear restriction fragment length polymorphisms (RFLPs). 1. Genome evolution of diploid and amphidiploid species. Theor.Appl. Genet. 75,784-794. Song, K. M., Osborn, T. C., and Williams, P. H. 1990. Brassica taxonomy based on nuclear restriction fragment length polymorphisms (RFLPs). 3. Genome relationshipsin Brassica and related genera, and the origin of B. oleracea and B. rapa. Theor. Appl. Genet. 79, 497 - 506. Souza, E., and Sorrells, M. E. 1989. Pedigree analysis of North American oat cultivars released from 1951 to 1985. CropSci. 29,595-601. Spickett, S. G., and Thoday, J. M. 1966. Regular responses to selection 3: Interaction between located polygenes. Genet. Res. 7,96 - 12 1. Stalker, M. T. 1980. Utilization of wild species for crop improvement. A h . Agron. 33, 1 11 147. Steinmetz, M., Minard, K., Horvath, S., McNicholas, J., Srelinger, J., Wake, C., Long, E.,
DNA MARKERS IN PLANT IMPROVEMENT
89
Mach, B., and Hood, L. 1982. A molecular map of the immune response region from the major histocompatibility complex of the mouse. Nature (London) 300,35 -42. Stephens, S. G. 1961. Speciesdifferentiationin relation to crop improvement. Crop Sci. 1 , l - 5 . Sternberg, N. 1990. Bacteriophage P1 cloning system for the isolation, amplification, and recovery of DNA fragments as large as 100 kilobases. Proc. Natl. Acad. Sci. U.S.A.87, 103-107. Stuber, C. W., and Edwards, M. 1986. Genotypic selection for improvement of quantitative traits in corn using molecular marker loci. Proc. Annu. Seed Trade Assoc. 41, 70-83. Stuber, C. W., and Goodman, M. M. 1983. “Allozyme Genotypes for Popular and Historically Important Inbred Lines of Corn, Zea mays L.,” USDA-ARS, ARR-S-16. U.S. Gov. Print. Off., Washington, D.C. Stuber, C. W., and Moll, R. H. 1972. Frequency changes of isozyme alleles in a selection experiment for grain yield in maize (Zea mays L.). Crop Sci. 12, 337-340. Stuber, C. W., Moll, R. H., Goodman, M. M., Schafer, H. E., and Weir, B. S. 1980. Allozyme frequency changes assoCiated with selection for increased grain yield in maize (Zea mays L.). Genetics 95,225 -236. Stuber, C. W., Goodman, M. M., and Moll, R. H. 1982. Improvement of yield and ear number resulting from selection at allozyme loci in a maize population. Crop Sci. 22, 737-740. Stuber, C. W., Edwards, M., and Wendel, J. F. 1987. Molecular-marker facilitated investigations in maize. 11. Factors influencing yield and its component traits. Crop Sci. 27,639648. Tanksley, S. D. 1983a.Introgression of genes from wild species. In “Isozymes in Plant Genetics and Breeding’’ (S. D. Tanksley and T. J. Orton, eds.). Elsevier, Amsterdam. Tanksley, S. D. 1983b. Molecular markers in plant breeding. Plant Mol. Biol. Rep. 1, 3-8. Tanksley, S. D., and Hewitt, J. D. 1988. Use of molecular markers in breeding for soluble solids in tomato-a reexamination. Theor. Appl. Genet. 75,8 1 1-823. Tanksley, S. D., Medina-Filho, H., and Rick, C. M. 1982. Use of naturally-occurringenzyme variation to detect and map genes controlling quantitative traits in an interspecific backcross of tomato. Heredity 49, I 1-25. Tanksley, S. D., Bernatzky, R. B., Lapitan, N. L., and Prince, J. P. 1988. Conservation ofgene repertoire but not gene order in pepper and tomato. Proc. Natl. Acad. Sci. U.S.A. 85, 6419-6423. Tautz, D. 1989. Hypervariability of simple sequences as a generalsource for polymorphic DNA markers. Nucleic Acids Res. 17,6463-6472. Thoday, J. M. 1961. Location of polygenes. Nature (London) 191,368-370. Tingey, S. V., Sebastian, S., and Rafalski, A. R. 1989. An RFLP map of the soybean genome. Abstr., Crop Sci. Soc. Am. Annu. Meet. Vallejos,C . E.,and Tanksley, S. D. 1983. Segregationof isozyme markers and cold tolerance in an interspecific backcross of tomato. Theor. Appl. Genet. 66,241 -247. Vassart, G., Georges, M., Monsieur, R., Brocas, H., Lequarre, A. S., and Christophe, D. 1987. A sequence in M 13 phage detects hypervariable minisatellites in human and animal DNA. Science 235,683-684. Vaughan, D. A. 1989. Collection, conservation, and potential use of the wild relatives of rice in Asia and Australia. Rev.Adv. Plant Biotechnol.,I985 - 1988, Int. Symp. Genet. Manipulation Crops, 2nd (A. MujeebKazi and L. A. Stich, eds), pp. 179- 190. Vergnaud, G. 1989. Polymers of random short oligonucleotides detect polymorphic loci in the human genome. Nucleic Acids Res. 17,7623-7630. Wallace, M. R., Marchuk, D. A., Andersen, L. B., Letcher, R., Odeh, H. M., Saulino, A. M., Fountain, J. W., Brereton, A., Nicholson, J., Mitchell, A. L., Brownstein, B. H., and
ANDREW H. PATERSON et ul.
90
Collins, F. S. 1990. Type 1 Neurofibromatosis gene: Identification of a large transcript disrupted in three NF 1 patients. Science 249, 18 1- 186. Weeden, N. F., and Wolko, B. 1990. Linkage map for the garden pea)Pisum sativum). In “Genetic Maps” (S. J. O’Brien, ed.), 5th Ed. Cold Spring Harbor Press, Cold Spring Harbor, New York. Weller, J. I. 1987. Mapping and analysisofquantitative trait loci in Lycopersicon(tomato) with the aid of genetic markers using approximate maximum likelihood methods. Heredity39, 413-421.
Weller, J. I., Soller, M., and Brody, T. 1988. Linkage analysis of quantitative traits in an interspecific cross of tomato (L. esculentum X L. pimpinellifohm) by means of genetic markers. Genetics 118,329-339. Wendel, J. F. 1989. New World tetraploid cottons contain Old World cytoplasm. Proc. Natl. Acad. Sci. U.S.A. 86,4132-4136. Williams, J. G. K., Kubelik, A. R., Livak, K. J., Rafalski, J. A., and Tingey, S. V. 1991. Oligonucleotide primers of arbitrary sequence amplify DNA polymorphisms which are useful as genetic markers. Submitted. Womack, J. 1990. “Mapping the Genomes of Domestic Animals.” Cold Spring Harbor Press, New York. Wright, S. 1968. “Evolution and the Geneticsof Populations.” Univ. ofChicago Press,Chicago, Illinois. Wu, K. K., Burnquist, W., Sorrells, M. E., Tanksley, S. D., Tew, T. L., Moore, P. H., and Heinz, D. J. 199 1. The detection and estimation of linkage in polyploids using single dose restriction fragments. Submitted. Young, N. D., and Tanksley, S. D. 1988. Restriction fragment length polymorphism maps and the concept of graphical genotypes. Theor. Appl. Genet. 77,95- 101. Young, N. D., and Tanksley, S. D. 1989. RFLP analysis of the size of chromosomal segments retained around the Tm-2 locus of tomato during backcross breeding. Theor.Appl. Genet. 77,95-101.
Young, N. D., Zamir, D., Ganal, M.W., and Tanksley, S. D. 1988. Use of isogenic lines and simultaneousprobing to identifyDNA markers tightly linked to the Tm-2agene in tomato. Genetics 120, 579-585. Yule, G. V. 1906. On the theory of inheritance of quantitative compound characters on the basis of Mendel’s laws-a preliminary note. Rep. Int. Conf:Genet., 3rd 140- 142. Zeven, A. C., Knott, D. R., and Johnson, R. 1983. Investigations of linkage drag in near-isogenic lines of wheat by testing for seedling reaction to races of stem rust, leaf rust, and yellow rust. Euphytica 32,319-327.
SOILSCIENCEAPPLICATIONS OF NUCLEAR MAGNETIC RESONANCESPECTROSCOPY William F. Bleam Department of Soil Science University of Wisconsin-Madison Madison, Wisconsin 5 3 706
I. Introduction and Overview 11. A Layman’s View of Modern NMR Spectroscopy
111.
IV.
V. VI.
A. Vector Model of Magnetic Resonance B. Statistical Model of Magnetic Resonance C. Relaxation Mechanisms NMR Spectroscopy as an Experiment The Design of NMR Studies A. Radio-frequency Pulses B. Free Induction Decay and Detection C. Elementary Pulse NMR Experiments D. Experimental Considerations E. Spectral-Editing and Selective-Excitation F. Experimental Considerations: Summary Current Developments and Their Future Implications A. Quadrupolar Nuclei B. Nuclear Magnetic Resonance of Interfaces and Adsorbates Conclusions Appendices A. Convention for Representing the Magnetic Field B. The Force on a Charge Moving in a Magnetic Field C. The Potential Energy of a Magnetic Moment in a Magnetic Field Symbols References
I. INTRODUCTION AND OVERVIEW In the past ten years technological breakthroughs have completely changed the face of nuclear magnetic resonance (NMR)spectroscopy. The most far reaching of these are superconducting electromagnets, Fourier transform techniques, and magic-angle sample spinning. These develop 91 ,#mmc#~ in Apnamy, VoI. 46 Copyright 0 1991 by Academic P m s , Inc. AU rights of qroduction in any form -4.
92
WILLIAM F. BLEAM
ments have increased the sensitivity of NMR instruments permitting the study of a greater range of less sensitive nuclei, expanded and redefined the experimental options available to NMR spectroscopy, and extended highresolution spectroscopy to polymers, colloidal systems, and solids. The magnetic field B (Appendix A) of superconducting electromagnetsin modern NMR instruments ranges from 7.0to 14.1 T, three to five times that of conventional electromagnets. The magnitude of the magnetic field is important because resolution is directly proportional to B. Fourier Transform NMR was made practical mainly by computer technology which allows acquisition and accumulation of many scans, thereby increasing the signal-to-noise ratio (S/N). Many isotopes with low natural abundance or low sensitivity have become accessibleto NMR study through the combination of high magnetic fields and Fourier transform techniques. Computer control of radiofrequency (rf ) generation and data acquisition has much broader implications than simply increased sensitivity. As we will see, modern NMR is founded on the capacity to manipulate nuclear-spin populations through “pulses” of rf energy and to observe the dynamics of these spin populations as they interact with their chemical environment. Modem pulse NMR is more than a spectroscopy; it is a physical means of performing experiments. The importance of this concept cannot be overemphasized and it underlies the explosive growth in NMR spectroscopy in the last decade. High-resolutionsolid-stateNMR dates from the mid- 1970swhen the first magic-angle sample spinning (MAS) devices were built. Rapid spinnin of the sample in a rotor inclined at the magic angle, 6- = cos-l( I/ 3) = 54.74”,to the static magnetic field B dramatically reduces the effect of several line-broadeninginteractions. Among these are chemical shift anisotropy, which becomes important for materials that are static or in very slow motion on the time scale of the NMR experiment (= lo-’sec), nuclear dipole- dipole interactions, and quadrupole-coupling interactions (Andrew, 1981).If the frequencyof MAS is on the order of the line broadening (expressed in frequency units) caused by the interactions listed above, then the MAS linewidth will be substantially narrowed. This article is written for researchers with little or no background in modern “pulse” NMR spectroscopy who wish to evaluate the applicability of this technique to their areas of study. It is the author’s judgment that successful use of this or any other physical technique requires the capacity to communicate with specialists and read the relevant scientific literature. As a user, you must be able to understand enough about NMR in its current form to recognize whether the information you are seeking would be provided by NMR experiments and to identify those experiments suited to your needs.
P
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
93
A full grasp of NMR fundamentals is not essential to evaluate potential applications of NMR spectroscopy, but informed discussions with NMR specialists and evaluation of suitable NMR experiments does require some basic knowledge of NMR physics. Only the most essential concepts will appear in the text, supplemented by citations for further reading. The following is a list ofexcellent NMR books and reviewsthe interested reader may wish to consult: Fukushima (1981); Fyfe (1 983); Engelhardt and Michel (1987); Farrar, (1987); Turner ef al. (1987); Blumich and Spiess (1988); Clayden (1988); Hams et al. (1988); Kessler ef al. (1988); Kirkpatrick (1988); King and Williams (1989a,b); Williams and King (1990a,b). The first section of this article will summarize concepts and terminology commonly used by NMR specialists and widely encountered in the NMR literature. The concept of modem NMR spectroscopy as a body of “experiments” will be examined next. The reader will find that the scientific literature contains references to literally scores of NMR “experiments,” each designed to reveal some aspect of the chemical environment. Effective application of modem NMR spectroscopy requires the informed selection of the proper experiments in the design of an NMR study. Developments in NMR spectroscopy of special interest to soil scientistswill be the final topic.
11. A LAYMAN’S VIEW OF MODERN NMR SPECTROSCOPY Nuclear spins (e.g., ‘H, I3C, 27Al,29Si,31P),when placed in a magnetic field B, align to form a net nuclear magnetic momentp whose dynamics can be represented using classical physics. Though the classical model is not the only way of representing NMR, nor is it completely adequate, it has several advantages. Almost invariably, pulse NMR experiments are illustrated in the scientific literature as “pulse sequences” which are a formalized description of the effect of a series of rf pulses on the nuclear magnetic moments in the sample. Most pulse sequence diagrams are easily visualized using a classical model for the dynamics of a magnetic moment.
A. VECTORMODELOF MAGNETIC RESONANCE 1. Classical Physics: Magnetic Moment in a Static Magnetic Field
Frames of reference figure prominently in all discussions of the dynamics of magnetic moments in magnetic fields. The dynamics, and hence the mechanics, of complicated systems can be simplified by the judicious choice
94
WILLIAM F. BLEAM
Fig. 1. Magnetic moment m arising from a particle with charge q and position r(l) moving at an angular velocity w in a closed orbit.
of a frame of reference wherein the dynamics are represented. Once the characteristics of that frame of reference are defined, the phenomena of magnetic resonance reduces to a very simple dynamic process described by equally simple equations. The classical representation of a magnetic moment m is a particle with charge q moving in a closed orbit (Fig. 1):
m=
I
(r x 4 d(4/2)
(1)
The magnetic moment of such a charged particle has a form similar to the angular momentum L:
L = m(r X v) m =L
d(q/2m)
m = (q/2m)L
(4)
m = yL (5) The constant of proportionality relating m and L is called the gyromagnetic ratio y. When m is placed in a static B it will experience a moment offorce, or torque, N (Appendix B). The mathematical expressionfor N can be rewritten to yield the following differential equation for the dynamics of the magnetic moment m in B: dm/dt = ym X B
(6)
The solution to Eq. (6) is a time-dependent magnetic moment m(t)that is
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
9.5
Fig. 2. A frame of reference (solid lines labeled:x’, y’, z) rotatingat an angular velocity u , about the z-axis of a static frame of reference (dashed lines labeled: x, y, z).
contant in magnitude but precessing at a constant frequency about B. By convention the z axis is defined to be in the direction of the static magnetic field, henceforth symbolized: Bo. The angular velocity at which m(t) precesses about Bo is known as the Larmor angular velocity 0,: 00
(7)
= -7Bo
The precession of m presents the opportunity to define two different frames of reference: a “laboratory” or static frame of reference, and a rotating frame of reference (Fig. 2). If we imagine a frame of reference that rotates , denoting the time rate of about the z axis with an angular velocity a change of m in the rotating frame as (dmldt),, then (using Eqs. (6) and (7)):
dmldt = (dm/dt), - m X a ,
(8)
Notice that Eq. (8)expresses precession rate in the static frame ofreference as the sum of two terms. The first term on the right-hand side is the apparent precession rate of m in a frame of reference rotating at an angular velocity a,,, while the second term is the precession rate of the rotating frame itself
(dmldt), (dmldt),
= ym X = ym
Bo
+ m X w,,
x (Bo + w o t / Y )
(9) (10)
+
The quantity B’ = (B, w,&) can be thought of as the apparent magnetic field in the rotating frame (Fig. 3) in which m precesses with the apparent angular velocity: wo - 0, = - Y(B0 + qoJr)
(1 1)
That is, m appears stationaryin a frame that rotates at the Larmor frequency
96
WILLIAM F. BLEAM
Fig. 3. The magnetic fields in the static frame of reference (B,)and a frame of reference rotating at an angular velocity w, (B’).
0,. Defining a frame of reference that rotates at 0, does not, of course, cause
Bo to become zero in the laboratory frame. 2. Classical Physics: Magnetic Resonance
Supposethat m is subjected to a time varying magnetic field B(t)consisting of two components. The component of the magnetic field along the z axis (B,) is time independent while the component of the magnetic field confined to the xy plane [B,(t)] rotates with an angular frequency o1(Fig. 4): B(t) = Bo
+ B, cos wit
Bo*B,cos ~ 0 1 = t 0
(12)
(13)
Fig. 4. Alignment of a time-varying magnetic field Bl(t) = BI cos(w,t) in both the static frame of reference(dashed lines labeled:x, y, z)and in a frame of referencerotatingat an angular velocity w,, = wI(solid lines labeled: x’, y’, z) about a static magnetic field Bo.
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
97
When B ,(t) is generated by an rf-transmittercoil in an NMR instrument it consists of two circularly polarized components rotating in opposite senses. For our purposes we need consider only one of the two circularly polarized components, equivalent to a B I(t)created by linearly polarized radiation, because only one of the two circularlypolarized componentscouples with m. Consider the dynamicsof m in a rotating frame of reference whose angular = 0 , . Since B, is stationary in this rotating frame it is velocity is: om, convenient to align B with the rotating-frame x axis, denoted x’ (Fig. 4). B, is created by a superconducting magnet and B,(t) by a rf-transmitter coil with the result: IBl(t)l Q lBol. If w , Q w,, the apparent field in the rotating frame of reference will be aligned close to the z axis and the dynamics of m will hardly show the effect of B,(t), i.e., B’ = B,. If the resonance condition is nearly met (i.e., w , = a,) then, even though IB,(t)I Q IBol, the apparent magnetic field in the rotating frame will be dominated by B I , i.e., B’ = B . Under these circumstances m will begin to precess about B with an angular frequency: yB .At resonance (i.e., w , = 0,)m will appear to precess about x’ under the sole influence of B, , i.e., B’ = B, . The precession of m about the rotating-frame Y-axis at resonance causes the potential energy of m to vary (cf., Appendix C) sincethe angle 4 between m and the magnetic field Bo is changing.
,
,
3. Classical Physics: Summary
A magnetic moment m when placed in a static magneticfield B, will experience a moment of force N and begin to precess at a characteristic angular velocity w, = - yB,. The gyromagnetic ratio y is the constant of proportionality relating the frequency of precession o, (the Larmor frequency) to the magnitude of the magnetic field B,. The magnetic field B’ and, hence the rate at which m precesses about B,, in a frame of reference rotating at a frequencyo, about an axis parallel to B, [Eq. ( 1 l)], depends on a,. If o, = o,then the rate of m precession and the apparent magnetic field B’ in that rotating frame is zero. Classical magnetic resonance of m in a time-varying magnetic field (B, B cos w ,t) is most easily visualized in a frame of reference rotating at a frequency o, = w , . When o1= y lBol in such a rotating frame (i.e., at or near resonance), B’ = B and m precesses about an axis normal to B, ,even though lBll 4 JBo(.In other words, the magnetic field along the z axis (the axis of B,) is zero in this rotating frame with the result that B , is the only apparent magnetic field influencingthe dynamics of m.The precession of m about B, when o1= - y IB,( in a frame of reference rotating at a frequency w, = w , is the spin flipping called magnetic resonance.
,
+
WILLIAM F. BLEAM
98
4. Quantum Physics: Nuclear Spin and Nuclear Magnetic Moments
Atomic nuclei are composed of fundamental particles known as protons and neutrons. Angular momentum or spin is one of the intrinsic properties of these particles. There are two quantized properties associated with the angular momentum of nuclei, the square ofthe nuclear angular momentum J2 = (h/2a)Z12
(14) and the component of the nuclear angular momentum in the z direction J, = (h/2a)Z, (15) The symbols I and Z, denote the nuclear spin vector and the z component of the nuclear spin. The energy of a nucleus depends on the component of the nuclear spin in the direction of B,. The component of the spin in the direction of B,, the magnetic momentp, and the energy for each nuclear-spinstate (cf., Appendix C) all are indexed by the nuclear magnetic quantum number m, for that spin state. The component of the nuclear magnetic moment in the z direction pz is expressed in terms of the nuclear magnetic quantum number: Pz
= Ydh/2NmI
(16)
yN is the nuclear gyromagnetic ratio, and the component of the nuclear angular momentum in the direction of B, is: (h/2a)m,[cf., Eq. (4)]. The nuclear magnetic moment z component pz,however, is seldom written in this form. The nuclear magneton pN and the nuclear g-factorg, are defined so that Eq. ( 16) can be written in a form that is the same for all of the quantized magnetic moments (including electron-orbitaland electron-spin moments): pz = (gIpN)mI
pN = eh/4am, = 5.0493(
(17)
J T-*
(18)
The pairing up of protons and neutrons in the nucleus, each with a spin of f, determines the nuclear-spin quantum number I. If the spins of all of the nuclear particles are paired then Z = 0 and the nucleus has no magnetic moment. If one spin is unpaired then I = f and the nucleus will have a magnetic moment. The allowed values for m, are: Z, ( I - I), . . . ,0, . . . , (-I l), - I. In the absence of a magnetic field B, the energies of the various nuclear spin states are independent of ml. When all nuclear spin states have the same energy they are said to be degenerate. This degeneracy is removed by B, (Fig. 5). The energy of each nuclear spin state is analogous to the
+
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
99
Fig. 5. Angular momentum and nuclear-spin alignment (fine arrows)of spin states for a spin-2 nucleus in a static magnetic field Bo.
classicalenergy associated with a magnetic moment m in a magnetic field Bo (cf., Appendix C):
E = -p.Bo
= -p
B0 - Yd~/2n)m,Bo
2
(19)
E = g,P,m,Bo
(20)
Each nucleus with nonzero spin has its characteristicgvalue,the values for protons and neutrons are: gI,,Roton = 2(+ 2.79270)
(21)
*9 30)
(22)
gI,Neutron = 2(-
The pairing of protons and neutrons in the nucleus not only determinesthe total spin of a particular nucleus, it also uniquely determines the nuclear g value g, for that isotope. 5. Quantum Physics: Nuclear Magnetic Resonance
Nuclear spin states, and hence the magnetic moments of nuclei, are quantized and subject to physical restrictions which do not apply to classical magnetic moments. Foremost among these is a “selection rule” which restricts allowed transitions between spin states when energy is absorbed or
100
WILLIAM F. BLEAM
emitted. The selection rule determines the quantized energy of allowed transitions: mI,initial - mI,w
= AmI = zt
1
(23)
A = gIpNBO (24) Nuclear magnetic resonance in the quantum model occurs when a nucleus absorbs a photon of radiation equal to that for an allowed transition between spin states. Resonance can occur only if spin-state degeneracies are removed by a static magnetic field B,. Classical resonance occurs when a magnetic momentp, precessing in a static magnetic field B,, is subjectedto a time-varying magnetic field B , ( t ) whose frequency of oscillation equals the Larmor frequency of p (i.e., oo= q). A semiclassical picture of nuclear magnetic resonance represents the nuclear magnetic moment p as if it were a classical magnetic moment whose magnitude is a fundamental property of the nucleus. The time-varying magnetic field B , ( t ) that couples with p as it precesses in the static B, field is produced by rf-radiation circularly polarized in the plane normal to Bo. Resonance occurs when the frequency of the oscillating B ,(t)field equals the quantized frequency restricting allowed transitions between nuclear spin states: 0 1
= YNBO = (2ngIpN/h)B0
(25)
B. STATISTICAL MODELOF MAGNETIC RESONANCE 1 . Statistical Mechanics: Nonequilibrium Relaxation
A characteristic of NMR is the vast numbers of nuclei undergoing resonance during a measurement. These nuclei must obey statistical thermodynamics appropriate to their nature. Physics recognizes three types of statistics depending on the nature of the particles. If the particlesare classical,they obey Maxwell - Boltzmann statistics. Quantum particles follow either Fermi -Dirac or Bose - Einstein statistics depending on whether they are fermions (half-integer spins) or bosons (integer spins). Imagine a system I consisting of NI particles in thermal contact with a reservoir L. The reservoir L is assumed to be much larger than I and, as a consequence, perturbations of the energy distribution in I have a negligible effect on the energy distribution of L. If the energy of each state (E(I,j),E(Ij,, . . .) of system I is taken to be the total energy of all N spins, then system I is described by MaxwellBoltzmann statistics regardless of the nature of the particles. It is important
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
101
x
F 0
6 State Population Fig. 6. Maxwell- Boltzmann probability distribution of state populations.
to understand exactly what is meant by an energy state E(I,i) in system I. Each of the NI particles in system I may occupy a well-defined set of spin states (vide supra, Section II,A,4). Imagine a particular arrangement of the NI particles, each in its own spin state. This arrangement has a well-defined energy E(I,i) equalling the sum over all N, particles in their respective particle states. Each arrangement of the NI particles represents a state of system I, hence E(I,i)is called the state energy. System I occupies these energy states with the probabilities: PIEtI,,!], PIE(IJ,], . . . . The relation between these state-occupation probabilities in a system governed by Maxwell - Boltzmann statistics is: (26) P[E(I,i)l/f'tE(I, j)l = exp(-E(I,ijkT)/exP(- E(I,jjkT) The sum, over all states, of the probabilities PIE(I,i,]must equal unity with the result (Zemansky, 1968):
P[E(I,~)I=exp(-E(I,ijkT)/Z
i
(exfi-E(I,jjkT))
(27)
It is possible to define a temperature T for any system I obeying Eq. (26), even if system I is not in thermal equilibrium with the reservoir L. The distribution of energy states in such a system Zis illustrated in Fig. 6. If system I consists of nuclear spins, then the temperature is referred to as the spin temperature T, . Imagine a circumstance wherein system Z and reservoir L, though in thermal contact, are not at thermal equilibrium. The reservoir L is assumed to be sufficiently large that it is at a constant temperature. The rate of transition between states of system I depends not only on the transition probability connecting individual particle states, but also on the probability that the reservoir L is in a state permitting a transition because it must absorb energy released by system I.
WILLIAM F. BLEAM
102
“z 1 E (LJ)
E(1.j)
Reservoir, L
E (Lm)
Fig. 7. Most probable transitions between energy states of a system 1, W&,,j) coupled to a reservoir L, qLnhLIY Heaviness of the lines representing the various energy states qualitatively indicate the relative populations of the states.
The energy states of the combined system, system I and reservoir L, is illustrated in Fig. 7. Reservoir L is always at thermal equilibrium and the most probable state is the lower state m.The rate of transition into the lower energy state j of system I is:
dN(,j)ldt = (N(&’(Lm)
fli,m-rj,lJ
- ( 5 I , j ) 4 & 1 ) Wu.1 +i,m)J
(28)
where and N(L,m,are the populations in states i and m of I and L and W(i,m4j,l)is the transitionprobability between statesof the combined system. At steady state the rate of transition is zero and, assuming the principle of detailed balancing is valid (i.e., W’ = W(i,m+j,,) = W&.i,m)), (N(l,ijN(,,j)) = (N(L,/jNtLm))
(29)
The energy distribution in the system I must equal that in the reservoir L, a condition known as thermal equilibrium. The transition rate [Eq. (28)] can also be formulatedin an alternative form that focuses more directly on the energy distribution in system I. The transition probabilities in I are defined in terms of the state populations in L and the combined transition probabilities:
Either of these expressionscan be recast using Eq.(27) and the conservation
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
103
of energy to show the relation between the transition probabilities and state energies for system I similar to Eq. (26):
Because the reservoir L is large and at thermal equilibrium at all times, then N(Lm)will be much larger than with the implication that W(+,) is also much larger than Wu+). 2. Statistical Mechanics: Spin Temperature and Spin-Lattice Relaxation
Suppose the particles in system I are nuclei with nonzero nuclear spins. Transitions in the reservoir L need not be explicitly represented since these are accounted for in Eq. (32). The statistics of a simple “two spin-state” system illustrate all of the features of more complicated systems. The energies of the parallel and antiparallel spin states are -pB, and +pBo, respectively.The ratio of W(--+)to W(++) has the same form as Eq. (32): W&.++W(+,-, = exp(-E(-, - E(+,)/kT,,
= exp(2pBdkT,,).
(33)
The temperature is denoted Tm to explicitly indicate that Eq. (33) defines the spin temperature. Even though system I may not be in thermal equilibrium with reservoir L, a spin temperature is defined provided the energy follows a Maxwell - Boltzmann distribution [Eq. (26)]. At thermal equilibrium T, equals the reservoir temperature T, and, since the ratio ‘‘pB,,/ kT,” is typically on the order of exp(2pBo/kT,) = 1
N(-jN(+, = 1
+ 2pBo/kT,
+ 2pBo/kT,
(34) (35)
It is important to recognize that the transition probabilities in Eq. (33) determine the rate at which an unmagnetized material develops a magnetic moment p when placed in a field B,. Initially the populations in the two states are equal, implying that T, is infinite. As spins undergo transitions from higher to lower energy states, the population difference increases as heat flows from I to L and T, approaches T,. The “cooling” of system Iis equivalent to its becoming “more polarized.” The lowest possible spin temperature T, occurs when all ofthe spins are in the lowest spin state, a state that can never be achieved. The highest spin temperature T,. however, is not positive infinity. Population inversion, when the population in the higher energy state N(-)exceedsthe population of
104
WILLIAM F. BLEAM
the lower energy state N(+!,causes the spin temperature to be negative. The highest spin temperature 1s defined to be a complete population inversion. Transitions between the energy states of system I can also be induced by magnetic resonance. The rate of magnetic resonance transitions also obey detailed balancing: W(-++) - W(++-) = w (36) N and n symbolize the total nuclear-spin population and the population difference between the two states:
The thermal transition probabilities [Eq. (33)]can be expressed as: W(++-)= W W(++) = W{1
+ 2pBo/kTm)
(39) (40)
The rate of transition due to both thermal and magnetic resonance takes the form:
dN(+)/dt= N(-){ y-++) + W ) - N(+){w(++-) + W) dN(-)/dt = NO{ W(++-)+ W ) - N(-){ W(-++)+ w) dnfdt = -2n{ W + w) 2NW{pBo/kT-)
+
(41)
(42) (43)
The final result [Eq. (43)]is obtained by subtracting Eq. (42)from Eq. (4 I), substituting Eqs. (37)-(40), and recognizing that under typical conditions: n < N. In the absence of rf-induced transitions the equilibrium excess of spins is defined to be:
no = NpBdkT, (44) This definition further simplifies Eq. (43) to yield the final rate expression: dnldt = - 2 W { n
+ no) - 2wn
(45)
Setting w = 0 and solving Eq. (45) yields: n(t) = no
+ {n(O)- no) exp(- 2 W t )
(46)
The population difference, which is n(0) initially, approaches thermal equilibrium with a relaxation time equal to ( 2 W - l . This is known as the spinlattice relaxation time and is symbolized: T, .Spin-latticerelaxation involves the nonradiative transfer of energy from the spins of system I to the “lattice” L, which is what the reservoir is typically called.
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
105
3. Statistical Mechanics: Summary
Nuclear magnetic resonance exhibits several characteristics that distinguish it from classical magnetic resonance. One of these characteristicsis the quantum nature of nuclear particles which imposes certain restrictions on the magnitude of nuclear magnetic moments and the resonant condition. The second, and sometimes overlooked, characteristic is the vast number of nuclear particles observed in NMR experiments. The statistics of large populations of nuclear spins has several unique features. Thermal contact between nuclear spins and a reservoir, generally called the lattice L, provides the mechanism whereby energy absorbed by the nuclear spins can be transferred by nonradiativeprocesses. Without this flow of heat NMR would be impossible. The rate at which the lattice can absorb heat from the nuclear-spin system determinesthe rate nuclear spin polarize. The recycle time in a pulsed NMR experiment,the delay time before another rf-pulse can be sent to the sample, is determined by the rate the nuclear spins “cool” by releasing heat to the lattice. Several of the basic NMR experiments can only be understood within the framework of a statistical model of nuclear spins. While it is possible to represent relaxation processes by the semiclassicalvector model of magnetic resonance, the fundamental physics of relaxation demands a statistical model. We will see that the statistical model is useful in understanding cross-polarization NMR experiments, wherein polarization is transferred from one nuclear-spin population to another. In these experiments (vide infra, Section 111,C,5), one of the nuclear-spin populations is made to mimic the thermal characteristics of the lattice L.
C . RELMCATION MECHANISMS 1. General Considerations
Relaxation is the process by which a system reaches equilibrium. Relaxation processes are important because they set practical constraints on the design and execution of any NMR experiment. Though arising from a variety of sources, some relaxation processes can be linked to fluctuations such as vibrations, rotations, diffusion, or chemical exchange. This section will explore the mechanisms underlying relaxation and the effects which relaxation have on magnetic resonance. Relaxation processes can be pictured in a variety of ways. One way of imagining what happens during relaxation is to represent the behavior of the
106
WILLIAM F. BLEAM
Fig. 8. Precession of a nuclear magnetic moment p (cross-hatched arrow)about B, (solid arrow)in a rotating frame of reference ( w , = a,). Elapsed time indicated by arc is n/2yB,.
net nuclear magnetic momentp in the framework of the vector model (vide supra, Section II,A, 1). This dynamic picture, however, fails to capture the mechanisms that trigger relaxation. The statistical model (vide supra, Section II,B,2) represents the effect which one form of relaxation, spin-lattice relaxation, has on the spin temperature T-. Relaxation acts through the thermal contact between the nuclear spin system Z and the lattice L. The mechanisms underlying thermal contact are not contained within the statistical model either.
2. Vector Model of Relaxation
Although the semiclassical vector model (vide supra, Section II,A,S) has the individual nuclear magnetic moments confined to the surface of a cone whose axis is the static magnetic field Boythe total nuclear magnetic moment p is aligned with Bo. If magnetic field B,(r) satisfying the resonant condition is turned on, p will begin to precess about B I in the rotating frame. The time it takes to rotate p through an angle a is: t =a/yB,
(47)
If B I(t) is turned off after a time t equal to 7c/2yB 1 , p will have been rotated into the xy plane (Fig. 8). This orientation of p is not at equilibrium and relaxation will eventually bring p back into alignment with B,. In general, the rates at which the z component ofp grows and the compo-
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
107
Fig. 9. Relaxation of magnetic moment p in a static magnetic field B, when there is no transverse relaxation. Dashed arc marks the path followedbyp (solid arrow)and cross-hatched arrow indicates the xy projection ofp.
nent aligned alongy' axis (rotatingframe) decays are not equal. Bloch ( 1946) proposed first-order rate expressions to describe the relaxation of p to its equilibrium alignment
First-order rate constants T;' and T;' are termed the longitudinal and transverse relaxation times, respectively. Longitudinal relaxation [ Eq. (48)] involves a change in the energy of the nuclear spins comprising p (Section II,C,3), but transverse relaxation [ Eq. (49)] does not. Longitudinal relaxation is spin-lattice relaxation (vide supra, Section II,B,2). Suppose all nuclear spins in the system have precisely the same Larmor frequency o,.The relaxation dynamics of the net magnetization p (in a for such a case would proceed as frame of reference rotating at o, = 0,) shown in Fig. 9. Now, suppose instead that longitudinal relaxation is zero and the nuclear spins in the system do not have the same Larmor frequency 0,. In a frame of reference rotating at o , = (w,), the mean Larmor frequency of the system, the individual nuclear magnetic moments comprising p will spread out in the xy plane (Fig. 10)as each precesses at its own Larmor frequency.This loss of phase coherencecauses a decrease in the net magnetic moment p in the xy plane without an increase in the z component. The energy of the magnetic moment does not change as phase coherence is lost. This demonstrates that the component of p in the xy plane clearly depends, in part, on the longitudinal rate constant T y l because the projection
108
WILLIAM F. BLEAM
Fig. 10. Relaxation of magnetic moment p in a static magnetic field Bowhen there is no longitudinal relaxation. Cross-hatched section illustrates the dephasing ofp (solid arrow).
of p onto the xy plane is still time dependent, even if T;l is zero. This suggests that T;l has the form: The magnitude of the net nuclear magnetic moment p along B, depends only on spin-lattice relaxation whereby the energy of the nuclear-spin system flows into the lattice. Conversely, the magnitude of the nuclear magnetic moment in the plane normal to Bo depends both on spin-lattice relaxation and loss ofp phase coherence [Eq. (SO)]. The rate constant T,*-‘containsthe effect of all processes other than spin-latticerelaxation and for that reason is called the efective transverse relaxation time. 3. Physical Mechanisms of Relaxation and the Correlation Time
The fluctuations that trigger relaxation are characterized by a correlation time z, which can be thought of as the time scale of the motion producing the fluctuation. The correlation time z, is the average time over which an internuclear vector connecting two magnetic dipoles maintains its orientation. For instance, small molecules in low viscosity liquids will take 10-lo seconds to tumble through an angle of about one radian. The vibration of bonds in molecules and solids span a range of correlation times 7, from 1O-I4 to seconds. These motions represent the lower limit for correlation times z,. In the frequency domain (o= z);’, these fluctuations cover a range from zero to 1014Hz. There is always some frequency component of fluctuationsat the Larmor frequency a, of any nucleus at any conceivable B,. Thermal contact between the nuclear-spin system Z and the lattice L, leading to what is called spin-lattice relaxation (vide supra, Section 11,BY2),arises from the “z; l” frequency components at the Larmor frequency 06. Fluctuating motions
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
109
lend a time dependence to dipolar interactions and the local magnetic field B, which excite transitions between spin states. If a molecule is rapidly tumbling on the NMR time scale, as in liquids or solutions, the local magnetic field B,, is averaged to zero and dipolar interactions have no net effect on the Larmor frequency m, because the effective magnetic field B,,
+
B,= Bo B, (51) approximately equals B,. Spin-latticerelaxation times T, are on the order of 1O-, - 1O2 secondsfor small molecules in low viscosity liquids. Paramagnetic species are a source of spin-lattice relaxation and can reduce T , to as low as seconds. The lifetime of the excited state contributesto the linewidth in any form of spectroscopy. The Heisenberg Uncertainty Principle states that product of the uncertainty in the transition energy 6E and the lifetime of the excited state 6t is a constant: (52) 6E 6t = h/2a Neglecting inhomogeneousbroadening of the signal due to spacial variability in the static magnetic field B,, the effective transverse relaxation time T f is equivalent to the lifetime of the excited nuclear-spin state 6t. All relaxation processes that effect linewidth contribute to the transverse relaxation time T, . The full width at half-height Am,,,of a (Lorentzian) NMR peak without inhomogeneous broadening is:
Am,,, = (2/T2)-1 (53) Inhomogeneous line broadening is always a factor, and T, is commonly not determined from [Eq. (53)]. The actual NMR experiments measuring spinlattice and transverse relaxation times are described in a latter section. Quadrupolar interactions, dipolar interactions (both nuclear- nuclear and nuclear-electron) and spin-spin exchange all contribute to Tf.All nuclei with spin I > 3 have an electric quadrupole moment which can couple with the electrostatic field gradient at the nucleus. This coupling decreases the lifetime of the excited state and decreases Tf(Slichter, 1978). When the internuclear vector connecting nuclear dipoles is stationary (i.e., the fluctuation frequency 7;' is zero) dipolar interactions are no longer motion-averaged and contribute to B,. Imagine a series of nuclear dipoles in a molecule or crystal environment located at different parts of the sample. In one molecule or crystal environment, the spin states may be arranged as - while in another part of the sample the nuclei of the molecule or crystal environment are in the spin states - -." At an internuclear distance of a typical bond, B,, may vary as much as lo-' T and
"+ + +,"
"+ + +
WILLIAM F. BLEAM
110
P
Q
P
Q
Fig. 11. Spin-spin exchange between coupled spins P and Q, arrowsindicating the transitions between spin states.
produce broadening of lo5 Hz in the NMR spectrum. This is static dipolar coupling. Consider the process of spin -spin exchange. Imagine two nuclei P and Q aligned antiparallel. The precession of nuclear magnetic dipole P in B, creates a time-varying magnetic field B at the position of nuclei Q and vice versa. These time-varying magnetic fields are, by definition, at the Larmor frequency o,and excite the two nuclei to exchange spin states (Fig. 11). Though there is no net change in energy when spin states are exchanged, it decreases the lifetime of the excited state for both nuclei involved. Because spin - spin exchange does not require a net motion of the positions of either nuclei it does not depend on correlation time z., In summary, longitudinal relaxation provides thermal contact between the nuclear - spin system Z and the lattice L. Nuclear motions fluctuatingat a frequency 7'; equal to the Larmor frequency a,trigger spin-lattice relaxation. Longitudinal relaxation also contributes to the transverse relaxation time [Eq. (50)] as do other relaxation mechanisms which have either no frequency dependence or result from static dipolar interactions. 4. Frequency Dependence of Relaxation Processes
The relation between the correlation time z, and the frequency cornponents is expressed in the spectral density function (Kittel, 1958; Slichter, 1978; Fukushima, 1981)
G ( o )= 4
Lm
(C(t) cos[wt]) dt
(54)
The function C(t)is the rate expression for the relaxation process. All relaxation processes influencing NMR are first order so that C(t)and G ( o )take the form: C(7)= exp(-f/z,)
G ( o )= 4zJ( 1
+o*z~)
(55)
(56)
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
111
A
B C COO
Frequency, o B
Correlation Time, 7, (s) Fig. 12. (A) Spectral density functions G ( u )for three different correlation times zc: r: > > ~ 2 Notice . the relative densities of each G(w)at the representative Larmor frequency w,,. (B) Longitudinal T, and transvem T, relaxation times for a magnetic moment with Larmor frequency w,, as a function of correlation time 7,. Minimum T, occurs at 7, = w, I . The relative relaxation times corresponding to the spectral density curves G(w)appearing in Fig. 12A are indicated by the correlation times (r: ,r: ,r s dehning the G@). T:
This spectral density function G(o)has the characteristic that: G(7;') = G(0)/2
(57)
The spectral density function G(w)is plotted in Fig. 12A for three selected correlation times 7,. The relation between the correlation time z, and the intensitiesof fluctua-
112
WILLIAM F. BLEAM
tions at zero frequency and the Larmor frequency woare important because spin-lattice relaxation is triggered by fluctuations at the Larmor frequency wo while the effective transverse (often called spin-spin) relaxation processes are triggered mainly by static interactions. As the correlation time zc increases the spin-lattice relaxation time TI should first decrease and then increase. The minimum spin-lattice relaxation time would occur when the correlation time zc equals the inverse of the Larmor frequency wcl. The transverse relaxation time T2should decrease monotonically as the correlation time zCincreases. Low viscosity liquids are characterized by short correlation times 7,.Molecules in such materials would exhibit spin-lattice relaxation times TI approximately equal to their transverse relaxation times T, . The correlation times zc in solids and viscous liquids are very long. At this limit, spin-lattice relaxation times TIare much longer than the transverse relaxation times T, . The relation between relaxation times and correlation times are illustratedin Fig. 12B.
111. N M R SPECTROSCOPY AS A N EXPERIMENT: T H E DESIGN OF N M R STUDIES A. RADIO-FREQUENCYPULSES A pulse spectrometer produces rf radiation by passing an alternating current through a coil whose axis is aligned normal to the static magnetic field Bo (Fig. 13). As the current in the coil oscillates, a time-varying magnetic field B,(t) develops along the axis of the coil (Fig. 13). The sinusoidal oscillation of B ,(t)can be represented as the sum of two circularly polarized magnetic fields rotating in opposite directions. One of these two circularly polarized magnetic fields couples with the precessing nuclear magnetic moments of the sample to cause magnetic resonance. In pulse NMR spectrometersthe current in the transmitter coil is on for a very brief time ( 1 0-5 - 1 0-6 sec) and the radiation emitted is of very high power (10,- lo3 W). During the brief time the transmitter is on (i.e., the pulsewidth T-), the magnetic field B,(t) oscillates at only one frequency, known as the carrier frequency, a .Imagine a seriesof pulses, each with a width of z- and separated by intervals of equal width zoduring which the transmitter is off (Fig. 14A). The square-wavescreated by pulses of radiation with frequency w,, can be represented as a Fourier transform (FT). The pulse train is in the time is in the frequency domain (Fig. 14B). That is to domain (Fig. 14A), its
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
113
Fig. 13. The static magnetic field B, between the poles of an electromagnet and the radiofrequency field 9, generated by an oscillatingcurrent in a coil.
say, although the transmitter is tuned to a single frequency cod, transmitting the radiation as a series of pulses is equivalentto transmitting a range of frequencies centered about the camer frequency mhm. The range of frequencies available from a pulsewidth 7- is:
a , f 4K/ZpulSc (58) Pulsing at some camer frequency coda actually delivers a range of frequencies of sufficiently broad bandwidth to excite nuclei in all potential chemical environments. This same pulse also switches the B field on long enough to cause p to precess about the x’ axis in the rotating frame (Fig. 8). Replacing B, in Eq. (7) and integrating over the precession angle and time yields Eq. (47). A “90” pulse,” which rotates p onto the y’ axis (Fig. 8), requires a T , , equal ~ to 7r/2yB,.
,
B. FREEINDUCTION DECAY AND DETECTION The final stage in any pulse NMR sequence is detection. Whatever may have happened during previous stages, the detection stage begins after the nuclear magnetic momentp has been tilted into the xyplane by the last pulse of the previous stage and the rftransmitter turned off. In this configurationp continuesto precess about B, as it relaxes to equilibrium (videsupra, Section II,A,2). The free precession of p in the xy plane creates an oscillating mag-
114
1
WILLIAM F. BLEAM
, ,, t I
Frequency, w Fig. 14. (A) Timedomain series of radio-frequency pulses, each with the duration rPh separated by a delay r,. The frequency of the radio-frequency radiation (illustrated in the expansion of a pulse)is mdm. (B) Frequency-domainof a series of radio-frequencypulses, each with the duration rpw separated by adelayr, .The frequency ofthe radio-frequencyradiationis ~cdllier~
netic field B, in the xyplane (Fig. 15). This oscillatingB, is the signalwhich is detected and from which the NMR spectrum is derived. The detector in pulse NMR spectrometers is a coil whose axis also lies normal to Bo. As p precesses in the xy plane under the influence of B,, the oscillating B, induces an alternating current in the detector coil. The detector coil measures the magnitude ofp over time. Relaxation processes, both longitudinal and transverse, are responsible for the decrease in the magnitude ofp over time so that the curve shown in Fig. 16is commonly referred to as the free induction decay (FID). The FID is the time-domain NMR spectrum ofp.The FT of the FID is its frequency-domain NMR spectrum. Two characteristicsof the FID are notable. The first of these are the oscillations in the FID. If the FID consists of only a single frequency, its FT spectrum will contain a single peak. On the other hand, if there are one or more beat patterns superimposedon the basic frequency the FT spectrum will contain several peaks. Oscillations in the FID are comprised of the frequency components that make up the signal. The oscillations of the FID are bounded by an exponentially decaying envelope: I(t) = exp{-t/T,)
(59)
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
115
Fig. 15. Cut-away view of a coil between the poles of an electromagnetshowingthe p r e e sion of a magnetic momentp at an angularvelocity u,under the influenceof the static field B, . The illustration portrays the instant after p has been rotated into the xy plane by a pulse of radio-frequencyradiation: 7pw = n/2yB,.
The relation between the rate constant for the FID and the linewidth is illustrated in Fig. 17. Inhomogeneity in the static magnetic field B, also causes line broadening, but neglectingthis there is a simple relation between linewidth Ao,,~ and the transverse relaxation time T2[Eq.(53)].
C. ELEMENTARY PULSENMR EXPERIMENTS All pulse NMR sequences consist of three stages (Fig. 18): preparation, evolution, and detection. The final stage, detection by acquisition of the FID, was described in the previous section. The length of the detection stage is determinedby the transverserelaxation time T,. Notice that acquisitionof the initial portion of the FID is critical for the detection of extremely broad signals (Fig. 17). The first stage, preparation, is also very simple to understand. Preparation, like detection, demonstratesthe critical role which relaxation processesplay in setting the parameters in pulse NMR experiments. The length of the preparation stage is determined by the spin-lattice relaxation time T ,. If the nuclei are not given sufficient time to reach thermal equilibrium before
WILLIAM F. BLEAM
116
Fig. 16. Timedomain free-induction decay (FID) of magnetic moment p in the xy plane showing oscillations bounded by an exponentially decaying envelope (--).
being excited by another resonant pulse of rfradiation, the spin temperature T* will increase with each cycle. Once the populations of the excited and ground state equalize the system is saturated and no further radiation will be absorbed. The length of the preparation state, therefore, must be long enough for thermal equilibrium between the nuclear spin system Zand the lattice L to be achieved. This generally takes about 5T,.Preparation is simply the delay time following detection before another rf pulse is emitted by the transmitter.
Time, t
Frequency, w
a
b
Fig. 17. Free-induction decay (FID)curves illustrating long (-) relaxation times T2.(a) Time domain; (b) frequency domain.
and short (-) transverse
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
117
Fig. 18. Elementary stages of a NMR pulse sequence. Horizontal axis portrays time while vertical axis portrays the magnitude of the radio-frequency field B, .
Preparation is essential because the basic pulse sequence (Fig. 18) is repeated many times. The rational behind this is that the S/N grows at the rate of N1I2,where N is the number of FIDs acquired. The total duration of a pulse NMR experiment is a product ofthe time required to cycle through the basic stages (preparation- evolution -detection) and the number of acquisitions. The spin-lattice relaxation time TI in solids and solutions of large polymers can be on the order of seconds, making the total duration of some NMR experiments prohibitivelylong since so much time is spent waiting for the spin system to reach thermal equilibrium. The distinguishing features of any NMR experiment are the number, length, and timing of the rf-pulses that comprise the evolution stage. Five basic NMR experiments are described in the followingsectionsthat illustrate the nature of the evolution stage. Describing these elementary NMR experiments will serve other purposes as well. They will illustrate how the “language” of pulse-sequences are linked to the vector and statistical models. They also incorporate basic pulse-sequences that are elaborated and combined to form much more complex experiments. 1. Single-Pulse Excitation
This is the simplest of all pulse NMR experiments. The evolution stage consists of a single 90; pulse that tilts p onto the rotating-frame y’ axis (Fig. 8). The symbol “90;’ indicates the B, field is aligned along the x’ axis. The pulsewidthr,, equals ?&!yB, .The lengthof eachcycle is:ST,+ n/2yB,+T2.
118
WILLIAM F. BLEAM
Fig. 19. Inversion-recoverypulse sequence: 180;-r-90;.
Of course, there will be some instrument “dead time” following the excitation pulse and preceding the acquisition of the FID because the transmitter coil is doing double duty as the detector. 2. Inversion-Recovery: Measurement of TI
The longitudinal relaxation time TIcan be measured in a variety of experiments, each with certain advantagesand disadvantages. Though a review Tl experiments is beyond the scope of this article, the reader should be aware other techniques exist that may be better suited for their particular situation. The evolution stage of an inversion-recovery Tl experiment consists of two pulses separatedby an interval z (Fig. 19).The initial 180;pulse invertsp (Fig. 20). Because there is no xy component, the relaxation of p is determined solely by spin-lattice processes. Magnetization p is allowed to decay for a time z before a 90; pulse carries p into the xy plane. Recall, evolution must always end with a component of p lying in the xy plane. The pulse sequence 180;- z- 90; is equivalent to a single 270; pulse when z is zero. This would align p with the negative y’ axis in the rotating frame. The effect of p alignment at the beginning of detection is illustrated in Fig. 2 1A. When z is short relative to Tl the peak in the FT spectrum appearsto be inverted. As Tl relaxation proceeds, following the 180; pulse, p decays to zero then grows along the z axis. A plot of the peak intensities would appear as in Fig. 2 1B. The point where the signal has zero apparent intensity is the half-life of the decay zlI2:
W)
(60) Each spectrum appearing in Fig. 2 1A is a separate 180;- z- 90; experiment. 71/2
= Tl
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
t=O
t = nIyB,
t
+7
1 19
= n/yB,
t=xlyB,+r
T < 0.693 T,
T > 0.693 T,
Fig. 20. Rotating-framealignment of magnetic momentp (cross-hatchedarrow)at various times in the inversion-recovery pulse sequence. (a)p prior to 180;-pulse. (b)p following 180;pulse. (c) Relaxation ofp after a time r < TIln(2) 0.693 T I .(d) Relaxation ofp after a time T > TI ln(2) = 0.693 TI. i=
Once the requisite number of scans have been acquired and Fourier transformed, the experiment is repeated with another 7. 3. Spin-Echo: Measurement of T,
The underlying mechanisms and their effect on p distinguish longitudinal and transverse relaxation one from another. There is another distinguishing property that is used to measure T2.Spin-latticerelaxation involves the flow of heat from the nuclear spin system to the lattice and is an irreversible process. T2 relaxation involves the increase in entropy without the flow of heat from the nuclear spin system into the lattice and is reversible. The pulse sequenceappears in Fig. 22 and the dynamicsofp at each state is illustrated in Fig. 23. Evolution begins with a 90; pulse which carriesp onto the y’ axis. Immediately, T2 processes cause p to dephase. Some nuclei precess faster than the average and some precess more slowly. After an interval 7, a 180; pulse inverts the phase of all nuclei by 180” (Fig. 23). Following a second interval of length 7,p will refocus (this is possible because T2processes are reversible). Detection begins at the point when p refocuses. The evolution pulse sequence is: 90;-7- 180;-7. The refocusing of p is called a spin echo, hence the “90;-7- 180;-7” experiment is called a spin-echoexperiment. The decrease in signal intensity
120
WILLIAM F. BLEAM
A T=O
A
T
Fig. 21. (A) Fourier transform spectra showing the phase and magnitude of the signal for various relaxationtimes 7 in a series of inversion-recoveryexperiments. (B) Signal intensity as a function of relaxation time 7 for a series of inversion-recovery experiments.
as a function of z is due to T, processes which, as we have seen [Eq. (50)], include spin-lattice relaxation and other effects known collectively as spin spin relaxation processes. The refocusing is used to eliminate the effects of static field B, inhomogeneities. 4. Spin-Lock: Measurement of T,,
Rotating frame spin-lattice relaxation differs from spin-lattice relaxation in the laboratory frame. The difference between the two relates to the correlation times T, probed by the two processes. Recall that spin-lattice relaxation is triggered by fluctuations at the Larmor frequency o,,which is in turn determined by the strength of the magnetic field B, [cf., Eq. (7)]. Although the B field is much weaker than the Bo field (the former arises from rfradiation, the latter from a superconductingmagnet), in the rotating frame p appears to respond to B as if B, did not exist. If, somehow, p were polarized along B, the rotating-frame resonance frequency o,= yBl it would be much smallerthan in the laboratoryframe so that fluctuationswith a much longer correlation time z, (or a much shorter frequency zL1) stimulate rotating-frame spin-lattice relaxation. The motions probed by rotatingframe spin-lattice relaxation are typically in the kilohertz region compared
,
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
12 1
-T+ Fig. 22. Spinecho pulse sequence: 90;-r- 180;-r.
to megahertz fluctuations probed by laboratory-frame spin-lattice relaxation. The nuclear magnetizationp can be polarized (aligned) along the B field by a technique called spin-locking. The spin-lock experiment consists of two pulses (Fig. 24). The first pulse is a 90;. At the end of the 90; pulse, the
t = nl2yB,
t =0
C
L
t = n/2yB, + T
d
t=
z
3n/2yB, + T
Fig. 23. Rotating-framealignment of magnetic momentp (cross-hatchedarrow)at various times in the spin-echo pulse sequence. (a) p prior to 90;-pulse. (b) p following 90l-pulse. (c) Dephasing ofp after a time T. (d) Dephasedpfollowing 18O;-pulse (pwill refocus after a time T following this instant).
WILLIAM F. BLEAM
122
I
I
7-i
I
I
Fig. 24. Spin-lock sequence: 90:-~,+~~. The label “y” above the spin-lock interval indicates B has been aligned with the y‘-axis of the rotating frame following a 90”phase shift of the radio-frequency pulse.
,
,
magnetization vector p lies along the y’ axis and the rf field B lies along the x’ axis (Fig. 25a). The next pulse is a very long pulse, transmittedby the same coil but phase shifted 90”.By shifting the phase of the rf radiation, the B field is made to lie along the y’ axis (Fig. 25b). Nuclear magnetizationp no longer precesses in the rotating frame because p and B, are aligned following the 90”phase shift p is locked along the y’ axis. Fluctuations at the frequency w1= T;, trigger relaxation of p, which decays exponentially. This decay rate in the spin-lock experiment is measured by the rotating-frame spin-lattice relaxation time TI P.TIPis deter-
,
a
b
Fig. 25. Rotating-frame alignment of the magnetic moment jt (cross-hatched arrow)and the radio-frequency field B, (solid arrow).(a) Following 90:-pulse; and (b) following 90” phase shift of the radio-frequency field B,bringing it into alignment with jt along they‘ axis. When B, and p are both aligned along the y’ axis, jt is said to be “spin-locked.”
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
123
mined by varying the length of the spin-lock pulse and plotting the FT spectrum signal intensity as a function of spin-lock pulse. 5 . Cross-Polarization: Enhancement by Polarization Transfer
Imagine two spin systems Z and S that are coupled via direct or “throughspace” dipolar coupling; i.e., the magnetic field from a nuclear magnetic moment of spin system Z acts “through space” on a nuclear moment of spin system S. Direct dipolar coupling is extremely sensitive to the internuclear separation and NMR experiments that depend on such interactionsare very useful, as we will see. The coupled spin systems Z and S are assumed to be in thermal contact, but not necessarily at equilibrium, with the lattice heat-reservoir L. The number of S spins in the higher energy (antiparallel,-) state and in the lower energy (parallel, +)state are Nfand NZ,respectively. The populations of the two states in the I spin system are N: and N L Using Eq. (26), the relative populations in the two systems are: N$/NL = exp(+ pIBo/kT)/exp(-pIBdkT)
(61)
Ns,Ip= exp(+ ksBo/kT)/ex~-p~Bo/kT)
(62)
Suppose IpII > [psi, then the polarization of the spin system Z will be greater than the spin system S. This is important because sensitivity is directly related to the nuclear magnetic moment (or gyromagnetic ratio) of a nucleus. If thermal contact between the spins in I and the spins in S is very good, and if the thermal contact between either spin system and the lattice L is relatively poor, then the polarization of Iwould determinethe polarization of S [cf., Eq. (29)]. Spin-lattice relaxation, the thermal contact between a spin system and the lattice heat-reservoir, is relatively slow in solids because the correlationtimes z, in the solid state are so long (Fig. 12B). This being the case, provided the cross-relaxation time TIs linking systems Z and S is short relative to Ti and Tf then the two spin systems would equilibrate with one another before either would equilibrate with the lattice L. The first step in a cross-polarization (CP) NMR experiment is to “cool” the spins of system I. Spin-lockingthe nuclear magnetization p I (videsupra, Section III,C,4) accomplishesthis by polarizing the spins of system Z in the rotating frame. Once the spins of system I are “cool,” efficient thermal contact must be established so that the spins of system Swill relax rapidly to the spin temperature TI of system I . At that point, the polarization of system Swill equal that of system I. Thermal contact for the “cooling” of the spins in system S by the spins of
WILLIAM F. BLEAM
124
b
a
Fig. 26. Double rotating frames during cross-polwkation showing the alignment of magnetic moments (@ andp') and radio-frequency fields (B,, and B,,). (a) I-spin rotating frame (oi,= y,B,) during spin-lock. I-spins precess about B,, (note precession cone) with angular = y,B,,. (b) S-spin rotating frame (US,= ysBo) during contact. S-spins precess frequency about B,, (note precession cone) with angular frequency Q f =ysB,,. Hartmann-Hahn matching occurs when = as.
system Z is achieved through the Hartmann- Hahn ( 1962)condition. Spinlocking the spins of system I occurs under the influence of the rf field B1I (Fig. 26a). The resonance frequency wll of these spins in the rotating frame is: "11 = YIBU (63) A second transmitter coil emitting a rf field B ls such that: YSBIS = "1s = "II
(64)
will satisfy the Hartmann- Hahn condition (Fig. 26b). Alternatively, the Hartmann - Hahn condition can be written Efficient transfer of energy requires that spin system Z be able to accept a specific quantum of energy from the spins of system S as they "cool." The Hartmann-Hahn condition ensures that system Z is prepared to accept quanta from system S. This must be accomplished in the rotating frame where the strengths of the two fields B lI and B Is can be controlled simultaneously. The pulse sequence for a CP NMR experiment appears in Fig. 27. There are two transmitters in this experiment. One transmitter is for the spins in system Zand emits a pulse sequence identical to the conventional spin-locking experiment. Followinga 90ipulse, the rfradiation from the I-transmitter is phase shifted 90" and maintained for a time zd-locL. The S-transmitter
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
:-T5
125
-;
Fig. 27. Cross-polarization pulse sequence showing the timing and duration of pulses generated by the I- and S-coils. The B,, (=y,B,/y,) field is switched off after the contact time and the B,, field is increased to produce high-field decoupling of I and S spins following spin-lock. Decoupling is spin-locking at a higher B,, field.
emits a single pulse that is initiated once the I-transmitter has phase-shifted and maintained as long as the I-transmitter is spin-locking. Once the Stransmitter is turned off, S-magnetizationpsis free to precess in the xy plane and the FID is acquired. The I-transmitter continues to emit rf radiation after the S-transmitter is switched off in most CP experiments. Continuous spin-locking the I-spins effectively decouples them from the S-spins. Dipolar decoupling will be discussed further in Section I11 (NMR experimental design). The CP NMR experiment and its variations were originally designed to enhance the sensitivity of “low gyromagneticratio,” low natural-abundance nuclides. Sensitivity is directly proportional to the nuclear - gyromagnetic ratio yN .Cross-polarizationenhancementf, is greatest when the I-spins are “abundant” and have a high nuclear - gyromagnetic ratio yN (Pines et al., 1973):
fcp = V,(I, + 1)%,ImIs(Is+ 1)%sYi)
(66)
The symbols I,, Yo,, and y, represent the nuclear spin quantum number, percent sample abundance, and nuclear - gyromagnetic ratio (respectively) for the I-nuclide. With few notable exceptions, ‘H is the nuclide chosen for spin-locking. Recycle time in pulse NMR experiments is determined by the spin-lattice
126
WILLIAM F. BLEAM
relaxation time Tl .Spin-latticerelaxation Tf of a “dilute” spin system Scan be very long in solids (vide supra, Section II,C,4), but in a CP experiment spin-lattice relaxation becomes TB. Because TB tends to be much shorter the recycle time is reduced and more FIDs can be acquired in the than Tf, same amount of time. A shortened recycle time permitting more acquisitions reduces the S/N providing another boost in sensitivity.
D. EXPERIMENTAL CONSIDERATIONS 1. Factors InAuencing Sensitivity
There are two factors that determine the sensitivity of a nuclide: percent , , ,% (natural abundance or enriched) and the nuclear- gyroabundance magnetic ratio yN. Some nuclides of interest to soil science (e.g., ‘H, IIB, 19F, 23Na,27Al,31P, 39K)have natural abundances %equal to or close to 100%, others can be quite rare (e.g., I3C: 1.1%, I5N:0.4%, 29Si:4.7%, 33S: 0.7%). Low natural abundance combined with low sensitivity means that NMR experiments at natural abundance may be virtually impossible for some nuclides (e.g., 15N:0.4%, 33S:0.7%),especiallyif the element is already a trace constituent. Enrichment may improve the chances of a successful NMR experiment ifat natural abundance sensitivity is close to the detection limit. The inherent sensitivity of the nuclide is a more unforgiving constraint. Sensitivity is commonly tabulated at both constant field and constant frequency relative to ‘H for equal number of nuclei. If the experiment is at natural abundance,comparisonsmust be made using the natural abundance sensitivity Sd :
S,,,
= ((%mJIOO)
X relative sensitivity}
(67) Modem pulse NMR instruments with a field strength of 1 1.7 T can detect on the order of lOI7 ‘H in a sample. Sample sizes in solid-stateand solution NMR instruments is on the order of a milliliter. Proton detection limits in a single-pulseexcitationexperiment would be about 1 mg/kg. Detection limits of other nuclides for a given percent abundance %-& can be estimated using Eq. (68): Detection Limit = ((%-p,,JIOO)
X relative sensitivity) X (1 mg/kg} (68)
Sensitivityenhancement can be achieved in some cases using polarization transfer techniques. Cross-polarizationis based on a direct dipolar coupling mechanism and can realize a maximum sensitivity enhancementfcpcomputed using Eq. (66). There are other means of enhancingsensitivity that are
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
127
similar to CP but involve different mechanisms and which offer different degrees of enhancement (King and Williams, 1989b; Williams and King, 1990a). 2. Sources of Line-Broadening
There are numerous sources of line-broadening in solid-state spectra. Inhomogeneity in the static magnetic field B, and, in sample spinningNMR experiments (videinfra, Section III,D,4),the settingof the rotor spinningaxis represent instrumental sources. The range of chemical environments in the sample also influenceslinewidth (Clayden, 1988).Chemical heterogeneityis classified as a source of inhomogeneous line-broadening and can be quite significant in amorphous solids and natural polymers. Relaxation processes (spin - spin exchange, chemical exchange, etc.) which decrease the lifetime of the excited state increase line width. When a substance is placed in a static magnetic field B, the magnetic moments align with the field and the substance is said to magnetize. In addition to this effect, which we have already examined, Boinduces currents in the paired electrons of the molecular or crystal orbitals which modifies the local magnetic field B,m at the nucleus of each atom in the substance: BIm= { 1 - A}Bo.
(69) The quantity A is a symmetric second-rank tensor called the shielding or chemical shift tensor and it representsthe deviation ofthe Larmor frequency w, from that of the isolated nucleus. The chemical shift, being a symmetric second-rank tensor, is diagonal in the coordinate system known as the principal axis system
Molecules in liquids and solutionstumble at such a high frequency that the tensor elements are averaged to what is called the isotropic or scalar chemical
shift: Incidentally, the perturbation creating the local field B, is quite small and AB, is on the order of 10-6B,- 10-5B,. The orientation dependence of the chemical shift tensor A is expressed in the powder pattern of finely divided solids (Fig. 28). The chemical shift anisotropy Ad is defined in terms of the principle elements of the chemical
128
WILLIAM F. BLEAM
C
6,162, 633 Fig. 28. Solid-state chemical shift powder patterns showing the principal components of the chemical-shift tensor S, and the chemical-shif? anisotropy AS. (a) Isotropic, (b) axial, (c) anisotropic.
shift tensor: (72) Ad = (633 - 4,), 8 3 3 2 82, 2 4, Three types of local symmetry are distinguishedby solid-state NMR powder patterns: isotropic (point groups: T,and Oh),axial (point groups: C, n 2 3) and anisotropic (all other point groups). Direct (through-space)and indirect (electron-coupledor J-coupled)dipolar interactions are another source of line-broadening. J-coupling is generally small in polar inorganic solids and can be neglected but it can be very important in organic solids and polymers. “Through-space” dipolar coup ling results from the magnetic interactions between the nucleus being observed and surrounding magnetic dipoles. The dipole field BD depends on the relative orientation Oij of the two moments and the internucleardistance R, separating the nuclei:
BD = gI&rn,(,4,/4n)( 1 - 3 cos2 ejj)rjj/lrjjl4
(73) Rapid tumbling of a small molecule in a low-viscosity liquid averages BD to zero on the NMR time scale. However, if the correlation time z, for the motion becomes very long; as in solids, viscous liquids, or solutions of large
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
129
polymers; BD is not averaged to zero and it makes an orientation-dependent contribution to the local field B,:
B,
= (1
- A)BO
+ BD
(74) The nuclei of quadrupolar nuclides ( I > f)possess an electric quadrupole moment which couples with the electrostatic field gradient at the nucleus. Quadrupolar coupling can, and often is, the primary source of line-broadening. Quadrupolar broadening is zero in chemical environments with cubic and certain other special symmelocal symmetry (point groups: Tdand 0,) tries (Knop et al., 1975; Knop, 1976; Akitt and McDonald, 1984). Most coordination environments in minerals have symmetries lower than cubic. Severe line-broadening results when the sample contain paramagnetic centers. Magnetic dipole coupling between nuclear magnetic moments and the magnetic moments of unpaired electrons is a comparatively long-range interaction. The effects of this sort of line-broadening interaction can be misleading since the effect on solid-state spectra is difficult to distinguish from chemical shift anistropy. Very small quantities of transition metal cations in mineral or organic matter samples (as low as 3% wt/wt) will broaden lines to the point where much of the spectrum is completely lost in the background (cf. Oldfield et al., 1982b;Vassallo et al., 1987;Kirkpatrick, 1988). 3. High-Power Decoupling
Most modern pulse NMR spectrometers are equipped with two coils, one (the observe coil) is tuneable over a wide range of carrier frequencies ohm and serves both to transmit and detect. The second is the decoupler coil and is tuned to the ‘H frequency. The decoupler is used mainly to decouple protons (‘H) from the observe, although it is used to polarize protons in CP experiments. Effectivedipolar decoupling requires very high power. The decoupling field B I H must cause the protons to precess at a frequency comparableto the dipolar linewidth (Fukushima, 1981). Given a precession rate for protons of = 43 MHz/T and dipolar broadening is on the order of 3 - 5 kHz,the decoupling field BlH must be on the order of 10 mT. This translates into power requirements for solid-state decoupling in the kilowatt range (Clayden, 1988). Decoupling is achieved by continuously irradiating the protons at their i.e., spin-lockingthe protons. The I-transmitter in resonant frequency of, the CP experiment is the decoupler. The decoupler is transmitting continuously throughout “spin-lock‘’ (during which the observe coil transmits a Hartmann - Hahn-matching B Isfield) and throughout acquisition (during
WILLIAM F. BLEAM
130
Fig. 29. Alignment of rotor-spinningaxis (-.-.-) angle ewc equals cos-~(~/Js) = 54.740.
relative to the static Bofield. The magic
which the observe coil detects the FID).If the decoupler were switched off at the end of the contact time, the S-spins would relax much more rapidly than if decoupler transmitted continuously through FID acquisition.
4. Sample Spinning
Modem high-resolution solid-state N M R is possible largely because of the discovery by Andrew and co-workers (1958a,b) that spinning a sample at a high frequency (3 - 10 kHz) about an axis inclined at a special angle to the static field Bo (Fig. 29) could dramatically decrease several important sources of line-broadening. The angle is known as the magic angle:
ed
= cos-yi/d3) = 54.74"
(75)
The technique is described in a variety of excellent reviews (cf. Andrew, 1981; Fyfe et al., 1982, 1983; Kirkpatrick et al., 1985; Klinowski, 1985; Oldfield and Kirkpatrick, 1985)to which the reader is referred. Andrew and co-workers found that MAS could virtually eliminate chemical shift and dipolar broadening and reduce quadrupolar broadening provided the spinning frequency ,v was sufficiently high. The technological advance came when rotors were constructed which would withstand the extremely high spinning frequencies (v, = 3 - 5 kHz or 180,000- 300,000 rpm). The chemical shift tensor A is a second-ranktensor. When a static array of nuclei are rotated about an axis inclined at an angle,6 to the static magnetic
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
131
field B,, the time-average of the principle tensor elements Sii have the form:
(8,) = +(a,
sin2 8)
+ ){3 cos28 - 1) 2 {Sii cos2Ai) i
(76)
where i iare the angles between the principle axes of the chemical shift tensor A and the rotation axis. The net effect of spinning at the magic angle 8-c is , . to average the chemical shift tensor A to its scalar value S The spinningfrequency ,v must be on the order of Ad [ Eq.(72), expressed in frequency units] to eliminate broadening. The static powder pattern will break into satellite peaks positioned at multiples of the spinning frequency ,v at lower frequencies. These satellites are called spinning sidebands (ssb) and are rotational echoes resulting from sample spinning. Their intensities are determined by the principle components of the chemical shift tensor (Maricq and Waugh, 1979; Henfeld and Berger, 1980). Direct, “through-space” dipolar interactions (both homonuclear and heteronuclear)contain the term (3 cos20, - 1) [c.f. Eq. (73)]. The mean value of such terms are averaged to zero when the sample is rotated about the magic angle 8-= (Andrew, 1981). The net result of MAS on direct dipolar interactions is to average interactions to zero. The spinning frequency v, must be on the order of the dipolar broadeningexpressed in frequency units. First-order quadrupolar broadening of the central magnetic resonance rn, = +.t) can be reduced by sample spinning, but transition (m,= -4 second-order quadrupolar broadening cannot be completely removed by spinning about a single axis. Oldfield and co-workers (Meadowset al., 1982; Ganapathy et al., 1982)studied the effects of variable-angle sample spinning (VAS) on the resolution of quadrupolar nuclei. They found that for quadrupolar nuclei the angle yielding the highest resolution was not the magic angle. Spinningat an angle other than the magic angle is generally restricted to quadrupolar nuclei and limited to those situations where the experimentalist wishes to resolve overlapping peaks. Combined CP and MAS was proposed by Pines et al. (1973) and first demonstrated by Schaefer and Stejskal (1976). Stejskal et al. (1977) and Sardashti and Maciel ( 1987)studied the effectswhich MAS had on CP. They showed that the cross-polarizationcross-relaxationtime T,, was highly sensitive to the strength of both the B,, and the B,, fields under MAS. This result has important implications when assignments and interpretations of signalsin a CP-MAS spectra are based on the relative cross-relaxation times TIs of the peaks. Magic-angle spin is used in virtually all high-resolution solid-state NMR studies. There has been a progression over the last ten years to higher static fields B,. While this has increases sensitivity, higher spinning frequencies have been necessary to reduce spinning side-band intensities which increase
-
132
WILLIAM F. BLEAM
with field strength. Spinning frequencies above 20 kHz (Dec et al., 1986) have been reported, but commercial rotors manufactured from silicon nitride have a maximum rating of = 17 kHz.
E. SPECTRAL-EDITING AND SELECTIVE-EXCITATION 1. Variable Contact-Time CP - MAS
S-spin (“observe”) signal intensity in a CP experiment is proportional to its magnetic moment at the beginning ofacquisitionpswhich is a function of four parameters: the I-spin magnetic moment at the beginning of spin-locking the cross-polarization cross-relaxation time T, and the rotatingframe longitudinal relaxation times for both the I-spins and S-spin ( TiPand TSp):
@ = {k1/k2}{AH1 - ~xP(-7-&& kl = vls}-l
(77) (78)
k, = (T{,,}-l+ {TSp}-l (79) Maciel and Sindorf (1980) measured the 29Si signal intensity of three resonances in silicagel as a function of contact time ran- and assigned these to chemical environments with different levels of hydrolysis. The assignment was based on analysis of the cross-polarization cross-relaxation times Q k,) [Eq. T, of the three resonances. At very short contact times (r-(77)] simplifies to linear function whose slope is proportional to the crosspolarization cross-relaxation time T,:
PS = { A ~ ~ ~ a J I l d T l S }
(80) T,estimation is sensitive the initial part of contact time [Eq. (77)]. Resonances with low S/N will be very difficult to fit either in the linearized form [Eq. (go)] or the full exponential form [Eq. (77)]. 2. INTERRUPTED-DECOUPLING CP - MAS
The typical CP experiment involves a transfer of polarization from ‘H to the “observe” nucleus. For instance, CP to I3Cis denoted 13C(lH}.Spinspin relaxation of the observe results from dephasing of the protons in the rotating frame. Spin-lockingthe protons prevents dephasing and eliminates their contribution to spin - spin relaxation. Alla and Lippmaa (1976) and Hester et al. (1976) described CP experi-
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY 903,
1 33
m
Y
decouple
I I I I
I I I
1
I
S
LLr.5-L-i Fig. 30. Intermpteddecoupling cross-polarization pulse sequence showing the inte mp tion of decoupling following spin-lock and prior to acquisition under high-powerdecoupling.
ments in which the decoupler was momentarily switched off (Fig. 30). Switching off the decoupler means that the protons are no longer being spin-locked and they begin to dephase (i.e., they undergo spin - spin relaxation). Alla and Lippmaa (1976) showed that dephasing could be used to measure the spin-spin relaxation time Tf of the observe nucleus in a CP experiment. Opella and Frey (1979) realized the spin-spin relaxation time Tf could vary with the chemical environment and demonstrated how interrupted-decoupling CP could select nonprotonated resonances from the observe spectrum. The notion behind this application of interrupted-decoupling is that protonated nuclei would be selectively suppressed because spin - spin relaxation would be more rapid for them than for nonprotonated nuclei. Observe signal intensity in a CPexperiment is proportionalto its magnetic moment at the beginning of acquisitionkS. The signal intensity, before the firstrotational echo, in an interrupted-decoupling CP- MAS experiment has the form: PS = Pi?@P(-
7indT2))
(81)
Rapid motion, rotations, or tumbling reduces the strength of static dipolar coupling. For instance, the carbons of rapidly rotating methyl groups behave as if they are not protonated. Magic-angle spinning also weakens static dipolar interactions (Andrews, 1981) with the result that as the spinning
134
WILLIAM F. BLEAM
frequency, ,Y increases longer interrupt intervals are required to achieve the same degree of suppression (Newman, 1990). Alemany et af.( 1983) have reviewed the numerous modifications of the initial interrupted-decoupling CP - MAS experiments. One problem that arises in this and other pulse NMR experiments is phase error. This is typically corrected by a 180; pulse that refocuses the S-spins as in the spinecho experiment (vide supra, Section III,C,3). Murphy (1983) demonstrate the effectiveness of a 180; refocusing pulse as a means of eliminating phase error in the interrupted-decoupling CP - MAS experiment. CP studies of Although interrupted-decoupling is widely used in 13C(*H} polymers, there are surprisingly few reports of interrupted-decoupling CP studies of inorganic materials. Griffin and co-workers (Aue et ul., 1984; Roufosse et al., 1984; Beshah et af., 1990) used interrupted decoupling 31P(1H}and 13C(IH) CP in studies of various calcium phosphates in bone and dental material. Opella and Frey (1 979) recommended an interrupt equal to one or two rotor periods (i.e., l/vw or 2/vEAAS)to interval rintnrUpt suppress protonated carbons. Bleam et al. ( 1989a,b)used interrupted-decoupling H) CP- MAS to study aluminum phosphates. They found discrimination between protonated and nonprotonated phosphates is optimum when the interrupt interval equals one-half a rotor period (i.e., fv-) rather than the much longer period recommended by Opella and Frey (1979) and used by others (Aue et al., 1984; Roufosse et al., 1984).When a new material is being studied, a variety of interrupt intervals should be tested using solids of known structure to optimize the conditions for effective suppression. Interrupted-decoupling need not be confined to CP experiments. Wilson et af. (1986) used 27Aland 29Sisingle-pulse excitation, proton-decoupled experiments to study proto-imogolite and allophane. They found that in most of their samples proton-induced spin -spin relaxation was independent of A1 coordination, but that one of the samples could be resolved into two relaxation rates. Dipolar dephasing of 31Pin vuriscite and waveffitealso appears to follow two rates (Bleam et al., 1989a). Interrupted-decoupling CP is a technique that has not yet been applied to its full potential in the study of hydrous inorganic materials. It has been used to study the hydrolysis of phosphate ions adsorbed on aluminum oxide (Bleam et af.,199l), but it suffers from a serious flaw as a tool for studying speciesat mineral surfaces. The NMR signals of surface speciesare generally characterized by relatively low S/N. Dephasing of the “observe” signal necessarily reduces the S/N. If the S/N is low to begin with, it may be impossible to accurately determine the amount of residual intensity remaining after dephasing.
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
135
S
Fig. 31. Inversion-recoverycross-polarizationpulse sequence showing a 180”phase shift of B I,during spin-lock and prior to acquisition under high-power decoupling. At the beginning of spin-lock, BI, is aligned along the positive y’ axis (indicated by “y” in top, 1 - 4 sequence). Following the 180”phase shift, B,, is aligned along the negative y’ axis (indicated by “-y” in top, I-coil sequence).
3. Inversion-RecoveryCP - MAS
An experimentcalled both selective-phase-inversionor inversion-recovery CP(Zumbalyadis, 1987;Cory, 1988;Xiaoling et al., 1988)distinguishesthe same chemical environments as the interrupted-decoupling CP. The advantage of inversion-recoveryCP is that the S/N is not lowered to any significant degree in the course of the experiment, making it a much more suitable technique when the “observe” is present at trace levels. Inversion-recovery CP inserts a 180” phase shift in the decoupler at the end of the contact time (Fig. 3 1) which causes the B IH to flip from alignment along the y’ axis to alignment along the -y’ axis. At the beginning to the contact time the protons are polarized dong& (ie., the protons are “cool”) and the observe are unpolarized (i.e., the observe are “hot”) (Fig. 32a). At the end of the contact time the protons and the observe are at thermal equilibrium (Fig. 32b). A 180” phase shift in the decoupler at this point causes temperature inversion in the rotating frame (Fig. 32c) because the proton magnetizationpHare rendered instantaneously antiparallel to the spin-locking B ,H field. The conditions in a CP experiment are such that the observe nuclei will preferentially come to thermal equilibrium the protons rather than with the
+
136
WILLIAM F. BLEAM
a
b
C
d
e Fig. 32. Excited (+) and ground (-) spin-state populations of I-and S-spins in an inversion-recovery cross-polarization pulse sequence. The negciptemp t) (Zemansky, 1968) is de< 0 and (fls&,,) = 0. (b) fined to be equal to - T-I. (a) At the start of spin-lock --m < Following spin-locking for a time equal to the cross-relaxation time T,: --m < (0:~ = < 0. (c) Immediately following the 180" phase shift of B,,: 0 < ( O h ) <+-m (rotatingframe spin-temperature inversion) and - < (fls&,,) < 0. (d) Incomplete rotating-framespinand (t)&) = 0. (e) Complete lattice relaxation following the B,I phase shift:0 < (t)'&,,) < rotating-frame spin-lattice relaxation following the B,,phase shift:0 < (t)& = 0&) < +-m.
-
+-
lattice. Hence, at some time after the 180" B 1, phase shift, the ground and excited state populations of the observe will first equalize (Fig. 32d), then finally match the population distribution of the protons (Fig. 32e). Hypothetical spectra matching each population distribution of Fig. 32 appear in Fig. 33. There is no "observe" signal intensity in spectra Figs. 32a,d because the excited and ground state populations are equal (Fig. 33a). The signal of the observe is above the line before and instantly after the 180" B,, phase shift (Fig. 33b) because temperature inversion of the observe has not yet occurred (Figs. 32b,c). Once thermal equilibrium has been reestablished following the 180" B phase shift (Fig. 32e), the observe signal has also changed phase and appears below the baseline (Fig. 33c). By varying the ,,,-,z ,the phases of the observe resonances length of the inversion interval can be sequentially inverted with those having the shortest cross-relaxation time T,, inverting first. Proton decoupling(spin-locking of the protons by the B,, field) is contin-
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
13 7
Fig. 33. Fourier transform spectra showing the phase and magnitude of the signal for various relaxation times T in inversion-recoverycross-polarizationexperiments. (a) (fis&,) = 0; (b) - m < (fi&} < O;and(c)O <(I.?&,) < +=.Thenegciptempt?(Zemansky,1968)isdefined to be equal to - T’.
uous throughout the inversion-recovery CP experiment. Furthermore, phase inversion is stimulated after maximum observe signal intensity (and therefore maximal S/N) has been achieved. These two factors make inversion-recovery CP a much more desirable technique for discriminatingprotonation than either interrupted-decoupling or variable-contact-time CP.
4. Reverse-ElectronicsCP - MAS A major justification for using CP was to take advantage of the enhancement attendant upon the transfer of polarization from “abundant” high-y to “dilute” low-y nuclei [cf., Eq. (66); Pines et al., 19731. Polarization transfer in the opposite direction (low-y to high-y nuclei), though seemingly sacrificing sensitivity,has been shown (Crosby et al., 1988) to be an excellent means of obtaining selective high-resolution ‘H NMR spectra of solids. The role of the decoupler and the transmit-detect coils are reversed, hence reverse electronics CP- MAS. The transmit -observe coil spin-locks the low-y nuclei, while the decoupler brings the protons into HartmannHahn contact in this counter-intuitive CP - MAS experiment. It is a selective excitation since only those protons near the nuclei being spin-locked will have polarization transferred to them. Crosby et al. (1988) used this technique to obtain ‘H(”P) CP-MAS spectra of two phosphate compounds. There is some loss in sensitivity;31Phas a sensitivity about 7% of ‘H. However, the number ofprotons in the sample are sufficientlylarge to counter-balance it. This technique should prove very useful in studying protons in hydrous minerals.
138
WILLIAM F. BLEAM
F. EXPERIMENTAL CONSIDERATIONS: SUMMARY Modem pulse NMR experiments are designed around pulse sequences that prepare spins, allow them to evolve under the influence of various components of the chemical environment, and detect the free-induction decay of the nuclear magnetization p in the plane perpendicular to the static magnetic field B,. The nuclear magnetizationp is manipulated by rf pulses that create one or more time-varying magnetic fields B I in the plane normal to the static field B,. The duration of the rfpulses determine boih the range of Larmor frequencies w, being excited and the number of degrees through which the nuclear magnetizationp is rotated. Complex pulse experimentsare built from much simpler pulse sequences (single-pulse excitation, inversion-recovery, spin-echo, spin-lock). Some pulse experiments are designed to measure dipolar interactions while others are designed to increase sensitivity or improve resolution. Spectral editing uses the evolution of the nuclear magnetizationp in responseto variations in the chemical environment to distinguish important properties of the environment. For instance, the structural role of protons influences the static “through-space” dipole coupling between protons and many nuclei. Variations in this coupling can be revealed by several CP experiments: variablecontact-time, interrupted-decoupling or inversion-recovery. The experimentalist has literally scores of multiple pulse experiments to choose from, each with its particular strengthsand weaknesses, each revealing some prop erty of the chemical environment.
IV. CURRENT DEVELOPMENTS AND THEIR FUTURE IMPLICATIONS A.
QUADRUP~LAR NUCLEI
1. Symmetric Second-Rank Tensors
Chemical shift, the nuclear electric quadrupole moment, and the electrostatic field gradient (EFG) are all represented by symmetric second-rank tensors. A symmetric second-ranktensor E has the property that a transformation R always exists such that:
E’ = RTE!R
(82) where 2’ is diagonal. The tensor elements (i = I, 2, 3) are called the principle elements. Although these three tensors differ in their properties,
r;j
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
139
convention ranks the principle elements: I<;3I
> I<;zI > IGiI
(83)
and defines an asymmetry parameter q": q" = {<; 1 - <$2)/
eQ eQzz- 42, (85) The parameter eQ appearing in Eq. (82) is called the nuclear electric-quadrupole moment. It is actually a scalar value expressing the difference between the two unique principle elements of the quadrupole moment tensor eQ. Charge neutrality in crystals, as a direct consequenceof the Laplace Equation, requires the EFG tensor to be traceless:
vzz+ vyy+v,=o
(86) This condition means that only two parameters are required to describe the EFG tensor: qEm and V,, .The largest principle element of the diagonal EFG tensor V, is called thefieldgradient. It is a scalar quantity typically symbolized: eqzz
= vzz
(87)
2. Quadrupolar Coupling Constants and the Spectra of Quadrupolar Nuclei
All nuclei with spin I > f have a nonspherical nuclear-charge distribution. The energy of a nonspherical nuclear charge surrounded by a nonspherical electronic charge distribution depends on the relative orientation of the nucleus in the electronic charge distribution. The classical nuclear-quadrupole-coupling energy EQ,
EQ = - { e 2 q ~ ~ Q > / 6
(88) is defined using the scalar field gradient eq, and the scalar quadrupole moment eQ. This leads to the natural definition of what is commonly called the quadrupole coupling constant x = e2qzzQ/h(division by Planck's Constant converts the coupling constant to frequency units). Coupling between the nuclear quadrupole moment and the EFG at the nucleus perturbs the energies of the nuclear spin states. This is illustrated in Fig. 34 for the case I = 3. The quantum quadrupole interaction is typically
WILLIAM F. BLEAM
140
+3/2
+ 112 - 112
... c= L ...
= . .. ...
-
...
-
-312
i = ...
= ...
Quadrupolar Coupling
Zeeman
Fig. 34. Zeeman and quadrupole-couplingperturbationsof spin states for a spin3 nucleus. Crosshatched arrows indicate allowed transitions, vertical scale is in energy units.
-
represented using perturbation theory. The “central” transition (m,= m,= ++) of nonintegrd quadrupolar nuclei (I= 3, 4, 3, 3) is not affected “to first-order” in perturbation theory by quadrupolar coupling between the nucleus and the EFG. This is not the case for integral quadrupolar nuclei (I= 1, 2, 3) so that quadrupolar interactions are much more important for these nuclei. With the exception of I4N(I= l), all quadrupolar nuclei of interest to soil scientists are half-integral. Second-order quadrupole coupling affects the central-transition peak in two ways (Kundla et al., 1981;Behrens and Schnabel, 1982;Meadows et al., 1982; Samoson et al., 1982; Freude et al., 1985).The center ofgravity of the peak wcog,in general, no longer corresponds to the Larmor frequency 0,. The magnitude of the quadrupolar induced shift for half-integer nuclei is directly proportional to the square of quadrupolarfrequency,
-+
0Q = { 3 ~ ~ 2 1 ( 2 1 1)) -
(89)
and inversely proportional to the field strength Bo = y-’oo,
- 0-
+
+ (t1*/3)>
= ( & ) ~ ~ ~ / ~ O ) V 1) ( ~- a>c1
(90) The reader is cautioned that not all quadrupolar resonance frequencies appearing in the scientificliterature have been corrected for quadrupolar-induced shifts. Second-orderquadrupolarinteractionsalso distorts the central-transition peak shape (Fig. 35). When the quadrupolar coupling constant x is on the order of 3 - 10 MHz, peak-shape distortion becomes the dominant source of line-broadening. There are two peaks, two “flanks” (i.e., the maximum and minimum frequencies of the powder pattern), and two shoulderspotentially visible in the central-transition peak whose frequencies determine the asym0 0
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
141
0
25
20
15
10
5
0
5
10
15
PPm Fig. 35. Solid-state single-pulseexcitationmagic-angle spinning N M R spectra.(A) 130.32 MHz *'Al solid-state N M R spectra ( I M AIClp external reference) of augelite ("H # I 37305) at 13.7 kHz spinningfrequency:200 k scans, 1 k data points, 125 kHz spectral width, 0.004 msec delay between repetitions, 0.5 psec pulse width (Bleam et al., 1989~).(B) 191.24 MHz IlB solid-state N M R spectra (boron trifluoride etherate external reference) borax [NMNH#I505651 single-pulse excitation, magic-angle spin at 14.0 kHz spinning frequency: 10 k scans, 2 k data points, 125 kHz spectral width, 0.02 msec delay between repetitions, 2.0 psec pulse width.
metry parameter q, the Larmor frequency w,, and the quadrupolar frequency wQ(Kundla et al., 198 1 ;Behrens and Schnabel, 1982; Samoson et al., 1982). It should be apparent from Fig. 35 that a large quadrupolar coupling constant x can cause extreme broadening. Unlike line-broadeningcaused by chemical shift anisotropy and dipolar interactions, it cannot be completely removed by MAS (Engelhardt and Michel, 1987). The best prospect for increasing resolution is to obtain the spectra at the highest possible field strengths B,. There are circumstances, however, when this is not the best strategy. If there are two peaks with nearly the same Larmor frequency a,, ,
142
WILLIAM F. BLEAM
but significantly different quadrupolar coupling constantsx, then obtaining spectra at lower fields could improve resolution by exploiting the quadrupolar induced shift [Eq.(90)]to separate the resonances. Quadrupolar parameters become more difficult to measure at higher field strengths because the quadrupolar coupling is weaker. Oldfield and co-workers (Meadows et al., 1982; Oldfield et al., 1982a; Ganapathy et al., 1982; Schramm and Oldfield, 1982) found that rotating about an axis inclined at an angle other than the magic angle could lead to significant line-narrowing. Because satellite transitions are affected by quadrupolar interactionsto first-order, spinning slightly off the magic angle can dramaticallyreduce the interferenceof these satellite transitions with the central transition. The optimum spinning angle 0, to obtain maximal narrowing depends on the asymmetry parameter q. Variable-angle spinning has disadvantages. Chemical shift and dipolar broadening are removed only by spinning at the magic angle. If chemical shift and dipolar broadening are significant, VAS may not significantly improve resolution. Variable-angle spinning severely broadens the centraltransition signal of integral-spin nuclei (m,= - 1 * m,= 1) where the quadrupolar coupling interaction is first-order. It is often tempting to assume that the chemical environments represented by the signals in an NMR spectrum are all that are present in a sample. Similarly, signal intensity, as measured by peak area, might be thought to reflect the relative proportions of nuclei in the various chemical environments. These assumptions are generally safe, provided the sample is relatively free of paramagnetic impurities (vide supra, Section III,D,3) and the nuclei being observed are spin+ Quantitative analysis of quadrupolar NMR spectra, however, should be approached with caution. 27A1signal losses as high as 95%of the total aluminum in the sample have been reported (Akittand Farthing, 1978;Resingand Rubinstein, 1978; De Jong et al., 1983; MacKenzie et al., 1985), mainly in amorphousand poorly crystalline solids. Signal loss is generally attributed to a combination of chemical inhomogeneityand chemical environments with large quadrupolar coupling constants, viz., x > 20 MHz (Bosachek et al., 1982;Alemany and Kirker, 1986;Alemany et al., 1988). Coordination sites with noncubic symmetry have the potential of large quadrupolar coupling constants x. Quadrupolar coupling provides a mechanism for exciting satellite transitions through the absorption of energy in the central transition (Samoson and Lippmaa, 1983b; Fenzke et al., 1984).The peak area is a function of the duration of the excitation pulse and the quadrupolar coupling parameters (viz.,x and q). The central-transition signal intensity is independent of
+
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
14 3
cop Eq. (86)only ifp is rotated through a small angle (Samoson and Lippmaa,
1983a),
3. Polarization Transfer to Quadrupolar Nuclei
A development of considerable interest is CP-MAS from protons to quadrupolar nuclei. The feasibility of performing this experiment was demonstrated in the early 1960s by Redfield (1963)and Laurie and Slichter (1964).The first report of CP-MAS of quadrupolar nuclei seems to have been the 27Al(1H)spectra of aluminophosphate zeolites by Blackwell and Patton ( 1984).Other published accounts include: IIB(lH) spectra of kernite (Woessner, 1987), 23Na{’H) spectra of borohydride, borax and kenyaite (Hams and Nesbitt, 1988),27Al{1H)spectra of aluminum ions on the surface of hydrated alumina (Moms and Ellis, 1989)and 27Al(1H)spectra of hydrothermally weathered albite (Yang and Kirkpatrick, 1989). Three points concerning CP from spin? nuclei (e.g., ‘H) to half-integral quadrupolar nuclei deserve special comment. The Hartmann - Hahn ( 1962) matching condition takes a more general form:
For the special case where the “abundant” nuclei (I-system) are spin?, the “dilute” nuclei (S-system) are half-integral quadrupolar, and CP transfers polarization into the spin states associated with the central transition (rn, = -3 * rn, = +&): a,= 1,
as= [I(I - 1)
+
(93) (94)
A major reason CP-MAS was not attempted sooner was the widely held perception that T f Pwould be too short for efficient polarization transfer [c.f., Eqs. (79)and (8l)].It is clear that the quadrupolar coupling constant x does affect the CP peak intensity (Woessner, 1987;Hams and Nessbitt, 1988; Moms and Ellis, 1989)with those nuclei characterized by larger quadrupolar coupling constants x cross-polarizing less efficiently. Perhaps the major significanceof these experiments is that selective exci-
144
WILLIAM F. BLEAM
tation of nuclei near protons has been extended to quadrupolar nuclei. Every one of the studies cited in this section observed differences between singlepulse excitation MAS and CP-MAS spectra. Moms and Ellis (1989) performed the first interrupted-decoupling z7Al(1H)CP- MAS experiments.
B. NUCLEAR MAGNETIC RESONANCEOF INTERFACESAND ADSORBATES The first, and therefore the most important, issue an experimentalistmust face when studying chemistry at an interface is sensitivity: will the experimental technique detect a signal from a speciesthat is, by definition, present at very low concentrations.Equations (66) through (68) will aid in answering this question. As a general rule, the surfacearea of the adsorbent should be at least loz mz g-l, the surface excess of the species being studied should be at [Eq.(67)] should least 100pmolm-2, and the sensitivity of the nucleus S,, no less than Although no one will observe NMR spectra of species at the surface of dN (N> 0) transition metal oxides, and although paramagnetic impuritiespresent a significant problem in most natural minerals (vide supra, Section 11I,Dy2), judicious use of paramagnetic impuritiescan be very useful. For those who are interested in pushing the limits of detection, the studies of ammonia and pyridine adsorption on y-alumina by Ellis and co-workers (Majors et a[., 1986; Majors and Ellis, 1987) are revealing. Both of these studies report 15Nspectra, whose sensitivityat 99%enrichmentis times that of protons, making it a notoriously insensitive nuclei, at 25% surface coverage on a 220 m2g- l adsorbent. The S/N after 25,000 cycles (taking = 7 hours) is = 20, meaning a S/N = 3 could be achieved after 14 hours (an overnight run) with a 2.5% surface coverage. This would represent about the lowest practical limit in I5N NMR studies of adsorption. Nuclear magnetic resonance studies of interfaces and adsorbates can be grouped into four topical categories: chemical species absorbed by zeolites, cations absorbed by phyllosilicates, surface alteration of minerals and oxides, and adsorbates bound to the surface of nonporous minerals and oxides. Zeolite studies far outnumber all other categories, but is a subject not directly pertaining to soil chemistry and will not be reviewed here. There are remarkably few high-resolution solid-state NMR studies of cations adsorbed by phyllosilicates. Bank et al. (1989) measured significant line-broadening in interlayer l13Cd and were able to demonstrate that this arose from interlayer dynamics (cations and water) rather than chemical
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
145
inhomogeneity of exchange sites. Two distinct chemical environments were observed in some cases. These two environments were attributed to interlayer and edge sites. Weiss et al. ( 1990)also observed two chemical environmentsin their 133Cs study of hectorite slurries, assigning these to lS3Csin the Stem- and Diffuselayer. Weiss et al. ( 1990)consideredinterlayer and edge sites but rejected this assignment because chemical exchange between the two environments was rapid. Weiss et al. (1990) also observed two interlayer chemical environments in the 133Csspectra of hectorite dried at 5OO0C, assigning these to 9-fold and 12-fold coordinate sites. Bank et al. (1989) were unable to cross-polarize interlayer Il3Cd. They reasoned that rapid tumbling of the interlayer l13Cd effectively decouples dipolar interactions with protons rendering cross-polarization impossible. Bleam et al. ( 1989b)reported a phosphate impurity in synthetic taranakite that did not cross-polarize. This suggests a criteria for identifying an adsorbate bound as an outer-sphere complex: the resonance appears in a singlepulse excitation spectra but is absent in the CP spectra. The studiesby Barron et al. (1985a,b) and Altaner et al. (1988)differ from those just discussed in that the interlayer cation is not directly observed. Adsorption sites are probed through the effects of which the adsorbed cation has on atoms in the phyllosilicate layer. Swelling (srnectite)and nonswelling (illite) interlayer sites were resolved by partially exchanging interlayer cations with Mn2+and editing the spectrum using an inversion-recovery experiment (vide supra, Section III,C,2). The underlying notion is that dipolar coupling between 29Siin the phyllosilicate layer lying near paramagnetic Mn2+causes a significant decrease in the spin lattice relaxation time T I . Studies of surface alteration exploit differences in the composition of the surface from the bulk to selectively excite surface species. The basic approach follows that of Maciel and Sindorf( 1980),who obtained 29Sispectra of surface silicate groups by dehydroxylating silica at high temperature, rehydrating the surface by exposure to a humid atmosphere and obtaining CP-MAS spectra of the material. Only those 29Siwithin = 5 A of a proton are observed in a CP experiment, which were confined to the surface of the silica. The CP - MAS investigations of the surfaces of alumina (Morris and Ellis, 1989)and hydrothermally altered albite (Yang and Kirkpatrick, 1989) both rely on the absence of protons in the materials studied to selectively excite surfaces species. An adsorbate may be in one of three major chemical environments at the mineral/aqueous-solution interface: outer-sphere surface complex, innersphere surface complex, and surface precipitate. The surface complexes are directly influenced by the pH of the suspension. This comes as no surprise
146
WILLIAM F. BLEAM
since we know that the surface charge and the affinity of the mineral surface for adsorbing protons and hydroxyl ions is continuously changing with pH. Nuclear magnetic resonance experiments can provide very useful information about these chemical environments and their response to solution pH changes. Hydrolysis in solution or at a mineral/solution interface represents a change in chemical environment observed as a pH-dependent chemical shift. If the hydrolysis products are in rapid chemical exchange on the N M R time scale, as is often the case, a singlepeak is observed that represents the weighted average of the chemical shifts of the species in exchange. Both inner- and outer-sphere surface complexes should exhibit pH-dependent chemical shifts, distinguishing these chemical environments from surface precipitates. The major property distinguishing inner- from outer-spherecomplexes is their dynamics. Outer-sphere surface complexes are expected to display “solution-like” mobility (i.e., short correlation times .r&, characterized by spin-lattice T, and spin -spin T2relaxation times that are nearly equal (Fig. 12b). A rapidly tumbling outer-sphere complex would also be difficult to cross-polarize and may not appear in a CP spectrum (Bank et al., 1989; Bleam et al., 1989b). An inner-sphere surface complex, on the other hand, would exhibit relaxation times more characteristic of solids (i.e., T, 4 T,) and ready cross-polarize. NMR study of phosphate adsorption Bleam et al. ( 1991) conducted a 31P on the surface of an aluminum oxyhydroxideand found two major chemical environments: a relatively pH-independent signal assigned to a surface precipitate, and a highly pH-dependent signal assigned to inner-sphere surface complexes (Fig. 36). Interrupted-decoupling CP -MAS experiments indicate that the pH-independent signal came from phosphates that were not directly protonated, typical of most aluminum phosphates where the phosphate oxygens coordinate aluminum ions and are unavailable for protonation (Bleam et al., 1989a). The pH-dependent signal probably came from directly protonated phosphates, as would be expected if these phosphates were inner-sphere surface complexes.
V. CONCLUSIONS Naturally occumng soil materials, both organic and inorganic, are generally poor specimens for direct NMR study for two reasons. First, transition metal ions (mostly Fe and Mn) are a ubiquitous interference. Even though one may see a signal in the NMR spectrum of a natural sample, many of the
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
147
B
Fig. 36. 12 1.49 MHz 3'P cross-polarkation solid-state N M R spectra (85%H3P0, external reference):256 scans, 256 data points zero-6lledto 8 k, 35.7 kHz spectral width, 5 sec recycle time, 600 psec contact time. Freeze-dried boehmite with surface-adsorbed phosphate from a pH = 9.0 suspension, delay between spin-lock and acquisition: 1 psec (A) and 250 psec (B) (Bleam et al., 1991).
chemical environments will not appear because of severe line-broadening. Second, natural materials are a mixture of many types of chemical environments. Line-broadeningdue to chemical heterogeneity severely reduces resolution and cannot be removed by MAS or technologically accessible field strengths. Materials suitable for NMR study must be either naturally low in transition metals and organic free radicals or materials from which these could be extracted in some way. Nuclear magnetic resonance studies of soil organic matter that have not rigorously removed transition metal impurities and verified the absence of organic free radicals give an incomplete picture of the polymers. Many NMR studies of smectites have used the hectorite from Hector (California), famous for its low content of paramagnetic impurities. Most NMR studieswill be confined to syntheticmaterials carefully prepared in the laboratory. A misconception that often arises is that the peak position in an NMR spectrum has a significance comparable to peak positions in other types of
148
WILLIAM F. BLEAM
spectroscopy.Once a sufficient number ofspectra are tabulated from known compounds, so the misconception goes, one would need only to compare the peak positions in the spectrum of an unknown material to the reference collection to identifythe compound. The structure and chemical factorsthat influence the final Nh4R spectrum are so complex, it is unlikely this approach will ever be as useful as it has been in, say, infrared spectroscopy. Nuclear magnetic resonance spectroscopy, like other physical methods, has its strengths and weaknesses, and it will yield the greatest return when used within the constraintsthese impose. The structuralinformation derived from an NMR study is fundamentally different from that obtained in an x-ray diffraction study (cf. Bleam et al., 1989a,c). The power of NMR is its capacity to probe interactions. The extent to which we can translate our concept of the chemical environments existent in soil organic matter polymers, at mineral surfaces, in poorly crystalline solids or other natural materials into this language of interactions will largely determine the utility of modem pulse NMR spectroscopy in soil science.
VI. APPENDICES
A. CONVENTION FOR REPRESENTING THE MAGNETIC FIELD Maxwell defined four vector quantities to represent electromagnetic phenomena. Two of these, symbolized E and B by Maxwell, depend on the material medium and two, D and H, are independent of material media. These vector quantities, assigned a variety of names, are defined by the following equations:
E = D/E
B-pH B and H, though related to one another, symbolize two different physical variables. This article will adopt the convention used by most texts on electromagnetictheory and use the original Maxwell symbols. The vector quantity B will be called the magneticjield and is assigned units of Tesla. Many NMR references represent the magnetic field with the symbol H. The reader should be aware of these two convections for symbolizingthe magnetic field.
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
B. THEFORCEON A CHARGE MOVINGIN A
MAGNETIC
149
FIELD
A moment of force, or torque, N,about some point (Goldstein, 1980) is defined as vector product of the position r of the object relative to the point and the force F acting on the object:
N = dL/dt = r X F
(B1)
The force F acting on a charged particle in a static magnetic field is:
F = q(v X B)
( B2)
Bleaney and Bleaney ( 1976) show that (B 1) and (B2) can be combined to yield:
N = dL/dt = m
X B = yL X B
dmldt = ym X B
(B3) (B4)
C. THEPOTENTIAL ENERGY OF A MAGNETIC MOMENTIN A MAGNETIC FIELD The potential energy E of a magnetic moment m in a static magnetic field B, is derived from the torque N acting on m. By definition, the potential energy associated with a torque is:
E=
I
Ndcp
The magnitude of the torque on m resulting from B, is related to the angle 4 between them (B3): N = IN1 = mB, sin 4
(C2)
E=-mB,cos~=-m.B,
(C3)
WILLIAM F. BLEAM
150
SYMBOLS Symbol
Unit
Js
C m2 SKI,N m T-I N m T-' rad T-l s-' J T-I C kg J
J K-I K s-
I
s-
I
K K K rad
Quantity Magnetic induction vector Electric field intensity vector Permittivity of vacuum Relative permittivity Permittivity, E,+~ Electric displacement vector Permeability of vacuum Relative permeability Permeability,popr Magnetic field intensity vector Magnetic moment vector Position vector Velocity vector M a Charge Gyromagnetic ratio Angular momentum Moment of force or torque Time Larmor angular velocity vector Angular velocity vector, rotating frame of reference Rotating frame xy axes Nuclear angular momentum vector Nuclear spin vector Planck Constant Nuclear magnetic quantum number Nuclear magnetic moment vector Nuclear magnetic moment, z component Nuclear gyromagneticratio Nuclear magneton Nuclear g-factor Nuclear spin quantum number Elementary charge Proton rest mass Energy statej, particle system Probability of particle system occupying energy state j Boltzmann Constant Temperature Population of particle-system state j Population of reservoir state rn Transition probability, combined system and reservoir Transition probability, particle system only Spin temperature Reservoir temperature Lattice temperature Tilt angle
(continued)
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
151
Symbois (continued) Symbol
Unit
Quantity Longitudinal (spin-lattice) relaxation time, laboratory frame Transverse relaxation time Effective spin -spin relaxation time Full-width at half-height Correlation time Pulse width or length Dead time, interval between pulses Canier frequency Oscillating magnetic field caused by p Delay time in pulse NMR experiments Longitudinal (spin-lattice) relaxation time, rotating frame Larmor angular velocity vector, rotating frame Cross-polarization cross-relaxation time Rotating-frame resonance frequency, I-spins Rotating-frame resonance frequency, S-spins Percent abundance, I-nuclide Chemical shift tensor Principle component, chemical shift tensor; i = 1,2, 3 Isotropic (scalar) chemical shift Chemical shift anisotropy Internuclear vector Sample spinning frequency Sample spinning angle Angle between spinning axis and Bh principle axis of A Magnetic moment, “abundant” or “observe” nuclei Magnetic moment, “dilute” nuclei Negciptemp Asymmetry parameter, 0 5 q 5 1 Electric quadrupole moment tensor Scalar electric quadrupole moment Scalar electrostatic field merit Quadrupole coupling constant, e2qz,e/h Center of gravity frequency, quadrupolar nuclei
REFERENCES Akitt, J. W., and Farthing, A. 1978. New 27AlNMR studies of the hydrolysis of the aluminum(II1) cation. J. Magn. Reson. 32,345-352. Akitt, J. W., and McDonald, W. S. 1984. Arrangements of ligands giving low electric field gradients. J. Magn. Reson. 58,401 -412. Alemany, L. B., and Kirker, G. W. 1986.First observation of 5-coordinatealuminum by MAS 27AlNMR in well-characterized solids. J. Am. Chem. SOC.108,6158-6162.
152
WILLIAM F. BLEAM
Alemany, L. B., Grant, D. M., Alger, T. D., and Pugmire, R. J. 1983. Cross polarization and magic angle spinning NMR spectra of model organic compounds. 3. Effect of the ”C-IH dipolar interaction on cross polarization and carbon-proton dephasing. J. Am. Chem. SOC. 105,6697-6704. Alemany, L. B., Timken, H. K. C., and Johnson, I. D. 1988. Aluminum-27 NMR study of AlP04-21and andalusite. Advantagesof high-field and very fast MAS. J. Magn. Reson. 80, 421 -438. M a , M., and Lippmaa, E. 1976. High resolution broad line I3CNMR and relaxation in solid norbornadiene. Chem. Phys. Lett. 37,260-264. Altaner, S. P., Weiss, C. A., and Kirkpatrick, R. J. 1988. Evidence from 29SiNMR for the structure of mixed-layer illite/smectite clay minerals. Nature (London)331,699 -702. Andrew, E. R. 1981. Magic angle spinning. Int. Rev. Phys. Chem. 1, 195-224. Andrew, E. R., Bradbury,A., and Eades, R. G. 1958a.Nuclear magneticresonance spectra from a crystal rotated at high speed. Nature (London) 182, 1659. Andrew, E. R., Bradbury, A., and Eades, R. G. 1958b.Nuclear magnetic resonance spectra in solids: Invariance of the 2nd moment under molecular reorientation. Arch. Sci.11,223226. Aue, W. P., Roufosse, A. H., Glimcher, M. J., and Griffin, R. G. 1984. Solid-state phosphorus3 1 nuclear magnetic resonance studies of synthetic solid phases of calcium phosphate: Potential models of bone mineral. Biochemistry 23,6110-61 14. Bank, S.,Bank, J. F., and Ellis, P. D. 1989. Solid-state I ’ T d nuclear magnetic resonance study of exchanged montmorillonites. J. Phys. Chem. 93,4841 -4855. Barron, P. F., Slade, P., and Frost, R. L. 1985a. Solid-state silicon-29 spin-lattice relaxation in several 2 : 1 phyllosilicate minerals. J. Phys. Chem. 89,3305 - 33 10. Barron, P. F., Slade, P., and Frost, R. L. 1985b. Ordering of aluminum in tetrahedral sites in mixed-layer 2 : 1 phyllosilicates by solid-state high-resolution NMR. J. Phys. Chem. 89, 3880- 3885. Behrens, H.-J., and Schnabel, B. 1982. The second order influence of the nuclear quadrupole interaction on the central line in the NMR of quadrupolar nuclei using rapid sample spinning. Physica B 114, 185- 190. Beshah, K., Rey, C., Glimcher, M. J., Schimizu, M., and Griffin, R. G. 1990. Solid-state carbon-13 and proton NMR studies of carbonate-containing calcium phosphates and enamel. J. Solid State Chem. 84,7 1 - 8 1. Blackwell, C. S., and Patton, R. L. 1984. Aluminum-27 and phosphorus-31 nuclear magnetic resonance studies of aluminophosphate molecular sieves. J. Phys. Chem. 88,6 1356139. Bleam, W.F., Pfeffer, P. E., and Frye, J. S. 1989a. 31Psolid-state nuclear magnetic resonance spectroscopy of aluminum phosphate minerals. Phys. Chem. Miner. 16,455-464. Bleam, W. F., Pfeffer, P. E., and Frye, J. S. 1989b. 31Pand 27Alsolid-state nuclear magnetic resonance study oftaranakite. Phys. Chem. Miner. 16,809-816. Bleam, W .F., Pfeffer, P. E., and Frye, J. S. 1989c. *’A1 solid-state nuclear magnetic resonance study of five-coordinate aluminum in augelite and senegalite. Phys. Chem. Miner. 16, 817-820. Bleam, W. F., Pfeffer, P. E., Goldberg, S.,Taylor, R. W., and Dudley, R. 1991. solid-state nuclear magnetic resonance study of phosphate adsorption at the boehmite/aqueous-sohtion interface. Submitted. Bieaney, B. I., and Bleaney, B. 1976. “Electricity and Magnetism.”Oxford Univ. Press, Oxford. Bloch, F. 1946. Nuclear induction. Phys. Rev. 70,460-474. Bliimich, B., and Spiess, H. W. 1988. Two-dimensional solid-state NMR spectroscopy: New possibilities for the investigation of the structure and dynamics of solid polymers. Angew. Chem., Int. Ed. Engl. 27, 1655-1672.
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
153
Bosachek, V., Freude, D., Fr6hlich, T., Pfeifer, H., and Schmiedel, H. 1982. NMR studies of 27Alin decationated Y zeolite. J. Colloid InterfaeSci. 85,502-507. Clayden, N. J. 1988. Solid state nuclear magnetic resonance spectroscopy in inorganic chemistry. Chem. Scr. 28,211-221. Cory, D. G. 1988. Separation of non-protonated from protonated carbon NMR resonances in solids by inversion-recovery cross polarization. Chem. Phys. Lett. 152,430-434. Crosby, R. C., Reese, R. L., and Haw, J. F. 1988. Cross polarization magic angle spinning proton NMR spectroscopy ofsolids. J. Am. Chem. SOC.110,8550-8551. Dec, S. F., Wind, R. A., and Maciel, G. E. 1986. High-speed magic-angle spinning. J. Magn. Reson. 70, 355 -359. De Jong, B. H. W. S., Schramm, C. M., and Parziale, V. E. 1983. Polymerization of silicate and aluminate tetrahedra in glasses, melts, and aqueous solutions-IV. Aluminum coordination in glasses and aqueous solutionsand comments on the aluminum avoidanceprinciple. Geochim. Cosmochim. Acta 47, 1223- 1236. Engelhardt, G., and Michel, D. 1987. “High-Resolution Solid-State NMR of Silicates and Zeolites.” Wiley, Chichester, England. Fmar, T. C. 1987. “An Introduction to Pulse NMR Spectroscopy.” Farragut Press, Chicago, Illinois. Fenzke, D., Freude, D., Frohlich, T., and Haase, J. 1984. NMR intensity measurements of half-integer quadrapolar nuclei. Chem. Phys. Lett. 11 1, 17 1 - 175. Freude, D., Haase, J., Klinowski, J., Carpenter, T. A., and Ronikier, G. 1985. NMR line shifts caused by the second-order quadrapolar interaction. Chem. Phys. Lett. 119,365-367. Fukushima, E. 198 I. “Experimental Pulse NMR: A Nuts and Bolts Approach.” Addison-Wesley, Reading, Massachusetts. Fyfe, C. A. 1983. “Solid State NMR for Chemists.” C.F.C. Press, Guelph, Ontario. Fyfe, C. A., Bemi, L., Clark, H. C., Curtin, D., Davies, J., Drexler, D., Dudley, R. L., Gobbi, G. C., Hartmen, J. S., Hayes, P., Klinowski, J., Lenkinski, R. E., Lock, C. J. L., Paul, I. C., Rudin, A., Tchir, W., Thomas, J. M., Wasylishen, F. R. S., and Wasylishen, R. E. 1982. Analytical chemical applications of high-resolution nuclear magnetic resonance spectroscopy of solids. Philos. Trans. R. Soc. London, Ser. A 305,591 -607. Fyfe, C. A., Bemi, L., Clark, H. C., Davies, J. A., Gobbi, G. C., Hartmen, J. S., Hayes, P. J., and Wasylishen, R. E. 1983. High-resolution magic angle spinning and cross-polarization magic angle spinning solid-state NMR spectroscopy. I n “Inorganic Chemistry: Toward the 21st Century” (M. H. Chisholm, ed.),pp. 405-430. Am. Chem. Soc.,New York. Ganapathy, S., Schramm, S., and Oldfield, E. 1982. Variable-angle sample spinning high resolution NMR of solids. J. Chem. Phys. 77,4360-4365. Goldstein, H. 1980. “Classical Mechanics,” 2nd Ed. Addison-Wesley,Reading, Massachusetts. Harris, R. K., and Nesbitt, G. J. 1988. Cross polarization for quadrupolar nuclei-Proton to sodium-23.J. Magn.Reson. 78,245-256. Hams, R. K., Jackson, P., Menvin, L. H., Say, B. J., and Hiigele, G. 1988. Perspectives in high-resolution solid-state nuclear magnetic resonance, with emphasis on combined rotation and multipule-pulse spectroscopy. J.C.S. Faraday 184,3649-3672. Hartmann, S. R., and Hahn, E. L. 1962. Nuclear double resonance in the rotating frame. Phys. Rev. 128,2042-2053. Herzfeld, J., and Berger, A. E. 1980.Sideband intensitiesin NMR spectraofsamplesspinningat the magic angle. J. Chem. Phys. 73,602 1 -6030. Hester, R. K., Ackerman, J. L., Neff, B. L., and Waugh, J. S. 1976. Separated local field spectra in NMR. Determination of structure in solids. Phys. Rev. Left. 36, 108 1 - 1083. KessIer, H., Gehrke, M., and Griesinger, C. 1988. Two-dimensionalNMR spectroscopy:Background and overview of the experiments. Angew. Chem., Int. Ed. (Engl. Trans.)27,490536.
154
WILLIAM F. BLEAM
King, R. W., and Williams, K. R. 1989a. The Fourier Transform in chemistry. Part 1. Nuclear magnetic resonance: Introduction. J. Chem. Ed. 66, A2 13-A2 19. King, R. W., and Williams, K. R. 1989b.The Fourier Transform in chemistry. Part 2. Nuclear magnetic resonance: The single pulse experiment. J. Chem. Ed. 66, A243-A248. Kirkpatrick, R. J. 1988. MAS NMR spectroscopy of minerals and glasses. In “Spectroscopic Methods in Mineralogy and Geology” (I. C. Hawthorne, ed.),Vol. 18, pp. 341-403. Bookcrafters Inc., Chelsea, MI. Kirkpatrick, R. J., Smith, K. A., Schramm, S., Turner, G., and Yang, W. H. 1985. Solid-state nuclear magnetic resonance spectroscopy of minerals. Annu. Rev. Earth Planet. Sci. 13, 29-47. Kittel, C. 1958. “Elementary Statistical Physics.” John Wiley & Sons, New York, NY. Klinowski, J. 1985. Magic-angle-spinningNMR. Solid State Ionics, 16, 3- 14. Knop, 0.1976. Arrangementsofpoint chargeshavingzero electric-fieldgradient.11.Acta Cryst. A32, 147- 149. Knop, O., Palmer, E. M., and Robinson, R. W. 1975. Arrangements of point charges having zero electric-fieldgradient. Acta Cryst. A31, 19-31. Kundla, E., Samoson, A., Lippmaa, E. 1981. High-resolution NMR of quadrapolar nuclei in rotating solids. Chem. Phys. Lett. 83,229-232. Laurie, F. M., and Slichter, C. P. 1964. Spin temperature in nuclear double resonance. Phys. Rev. A 133, 1108- 1122. Maciel, G. E., and Sindorf, D. W. 1980. Silicon-29 nuclear magnetic resonance study of the surfaceof silicagel by cross polarization and magic-angle spinning. J. Am. Chem. SOC.102, 7606- 7607. MacKenzie, K. J. D., Brown, I. W. M., Meinhold, R. H., and Bowden, M. E. 1985. Thermal reactions of pyrophyllite studied by high-resolution solid-state 29Siand 27Alnuclear magnetic resonance spectroscopy. J. Am. Ceram. SOC.68,266-272. Majors, P. D., and Ellis, P. D. 1987. Surface site distributions by solid-state multinuclear NMR spectroscopy. Pyridine binding to y-alumina by I5Nand 2HNMR. J. Am. Chem. Soc. 108, 8 123-8129. Majors, P. D., Raidy, T. E., and Ellis, P. D. 1986.A multinuclearsolid-state NMR investigation of the chemisorption of ammonia on y-alumina. J. Am. Chem. SOC.108,8 123- 8 129. Maricq, M. M., and Waugh, J. S. 1979.NMR in rotating solids.J. Phys. Chem. 70,3300- 33 16. Meadows, M. D., Smith, K. A., Kinsey, R. A., Rothgeb, T. M., Skarjune, R. P., and Oldfield,E. 1982. High-resolution solid-state NMR of quadrupolar nuclei. Proc. Natl. Acad. Sci. U.S.A.79, 1351-1355. Morris, H. D., and Ellis, P. D. 1989. 27Alcross polarization of aluminas. The NMR spectroscopy of surface aluminum atoms. J. Am. Chem. SOC.111,6045-6049. Murphy, P. D. 1983. Improvement of the cross-polarization NMR experiment for suppression of rigid protonated carbons. J. Magn. Reson. 52,343 - 345. Newman, R. H. 1990. Analysis of results from intermpteddecoupling NMR pulse sequences combined with high-speed magic-angle spinning. J. Magn. Reson. 86, 176- 179. Oldfield, E.,and Kirkpatrick, R. J. 1985. High-resolution nuclear magnetic resonance of inorganic solids. Science 227, 1537- 1544. Oldfield, E., Kinsey, R. A., Smith, K. A., Nichols, J. A., and Kirkpatrick, R. J. 1982a. Highresolution NMR of inorganic solids. Influenceof magnetic centers on magic-angle samplespinning lineshapes in some natural aluminosilicates. J. Magn. Reson. 51,325-329. Oldfieid, E., Schramm, S., Meadows, M. D., Smith, K. A., Kinsey, R. A., and Ackerman, J. 1982b. High-resolution NMR spectroscopy of quadrupolar nuclei in solids: Sodium salts. J. Am. Chem. Soc. 104,919-920. Opella, S. J., and Frey, M. H. 1979. Selection of nonprotonated carbon resonances in solid-state nuclear magnetic resonance. J. Am. Chem. Soc. 101,5854-5856.
SOIL SCIENCE APPLICATIONS OF NMR SPECTROSCOPY
155
Pines,A.,Gibby,M.G.,and Waugh, J. S. 1973. ProtonenhancedNMRofdilutespinsinsolids. J. Chem. Phys. 59,569-590. Redfield, A. G. 1963. Pure nuclear electricquadrupole resonance in impure copper. Phys. Rev. 130,589-595. Resing, H. A., and Rubinstein, M. 1978. Hydrolysisofcrystalline aluminosilicatecatalysts. 27A1 NMR of hydrated zeolite 13-X. J. Colloid Interface Sci. 64,48-56. Roufosse, A. H., Aue, W. P., Robert, J. E., Glimche, M. J., and Griffin, R. G. 1984. Investigations of the mineral phases of bone by solid-state phosphorus3 1 magic angle spinning nuclear magnetic resonance. Biochemistry 23,6 1 15-6 120. Samoson, A., and Lippmaa, E. 1983a. Excitation phenomena and line intensities in high-resolution NMR powder spectra of half-integer quadrupolar nuclei. Phys. Rev. B 28,65676570. Samoson, A., and Lippmaa, E. 1983b. Central transition excitation spectra of half-integer quadrupolar nuclei. Chem. Phys. Left. 100,205-207. Samoson, A., Kundla, E., and Lippmaa, E. 1982. High resolution MASNMR ofquadrupolar nuclei in powders. J. Magn. Reson. 49,350-357. Sardashti, M., and Maciel, G. E. 1987. Effects of sample spinning on cross polarization. J. Magn. Reson. 72,467-474. Schaefer, J., and Stejskal, E. 0. 1976. Carbon-13 nuclear magnetic resonance of polymers spinning at the magic angle. J. Am. Chem. SOC.98, 1031 - 1032. Schramm, S., and Oldfield, E. 1982. High resolution solid-state NMR studies of quadrupolar nuclei: Quadrupolar-induced shih in variable-angle sample-spinning of a borosilicate glass. Chem. Commun. pp. 980-98 1. Slichter, C. P. 1978. “Principles of Magnetic Resonance.” Springer-Verlag, Berlin. Stejskal, E. O., Schaefer, J., and Waugh, J. S. 1977. Magic-angle spinning and polarization transfer in proton-enhanced NMR. J. Magn. Reson. 28, 105- 112. Turner, G. L., Kirkpatrick, R. J., Risbud, S. H., and Oldfield, E. 1987. Multinuclear magic-angle sample-spinning nuclear magnetic resonance spectroscopicstudies of crystalline and amorphous ceramic materials. Am. Ceram. SOC.Bull. 66,656-663. Vassallo, A. M., Wilson, M. A., Collin, P. J., Oades, J. M., Waters, A. G., and Malcolm, R. L. 1987. Structural analysis of geochemical samples by solid-state nuclear magnetic resonance spectrometry. Role of paramagnetic material. Anal. Chem. 59,558-562. Weiss, C. A., Kirkpatrick, R. J., and Altaner, S.P. 1990. The structural environments ofcations adsorbed onto clays: ”’Cs variable-temperatureMAS NMR spectroscopicstudy of hectorite. Geochim. Cosmochim. Ada 54, 1655- 1669. Williams, K. R., and King, R. W. 199Oa. The Fourier Transform in chemistry-NMR. Part 3. Multiple-pulse experiments. J. Chem. Educ. 67, A93-A99. Williams, K. R., and King, R. W. 1990b. The Fourier Transform in chemistry-NMR. Part 4. Two-dimensional methods. J. Chem. Educ. 67, A125-A137. Wilson, M. A., McCarthy, S. A., and Fredericks, P. M. 1986. Structure of poorly-ordered aluminosilicates. Clay Miner. 21,879-897. Woessner, D. E. 1987.Observation of effects of hydration on CP/MAS NMR measurements on quadrupolar nuclei. Z. Phys. Chem., N.F. (Munich) 152, 51-58. Xiaoling, W., Shanmin, Z., and Xuewen, W. 1988. Selective polarization inversion in solid state high-resolution CP MAS NMR. J. Magn. Reson. 77, 343-347. Yang, W.-H. A., and Kirkpatrick, R. J. 1989. Hydrothermal reaction of albite and a sodium aluminosilicateglass: A solid-state NMR study. Geochim. Cosmochim.Acfa53,805 - 8 19. Zemansky, M. W. 1968. “Heat and Thermodynamics.” McGraw-Hill, New York. Zumbalyadis, N. 1987. 1H/29Sicross-polarization dynamics in amorphous hydrogenated silicon. J. Chem. Phys. 86, 1162- 1166.
This Page Intentionally Left Blank
POTENTIAL OF BROWN-MIDRIB, LOW-LIGNINMUTANTSFOR IMPROVING FORAGEQUALITY J. H. Cherney', D. J. R. Cherney2, D. E. Akin), and J. D. Axtel14 'Department of Soil, Crop, and Atmospheric Sciences and 2Department of Animal Science, Cornell University Ithaca, New York 14853 'Richard B. Russell Agricultural Research Center Agricultural Research Service U.S. Department of Agriculture Athens, Georgia 306 1 3 'Depanment of Agronomy Purdue University West Lafayette, Indiana 47907
I. Introduction 11. History
111. Genetics IV. Lignin Biosynthesis A. Biochemistry B. Modifications in Enzyme Activity V. Forage Quality- Plant Response A. Lignin and Total Fiber B. Phenolics C. Neutral Cell Wall Sugars D. Nitrogen Concentration E. Digestibility F. In Vitro Rate of Digestion VI. Phenolic - Carbohydrate Complexes A. Isolation and Characterization B. Influence on Rumen Microorganisms VII. Plant Morphology and Anatomy A. Leaves of Normal and Brown-Mideb,, Sorghum B. Stems of Normal and Brown-Midrib Pearl Millet VIII. Forage Quality -Animal Response A. Intake B. In Vivo Digestibility C. Performance D. Palatability I57 ~ Cin &znmny9 I Val. 46 Copyright 0 1991 by Academic Press, Inc. AU rights of reproduction in any form reserved A
158
J. H. CHERNEY et d. IX. Biotechnology Potential X. Summary and Conclusions References
I. INTRODUCTION Two major concerns in our society today are maintaining a sustainable agriculture and reducing our dependence on imported energy sources. One common link between these problems is forage crops. Forage crops are seen as one of the better methods of reducing soil erosion and improving sustainability of agricultural lands. Forage crops also are being investigated by the U.S. Department of Energy as a potential alternative energy supply (J. H. Cherney et ul., 199I). Brazil is attempting to become energy independent with the assistance of forage crops for biomass conversion (Schaffert and Gourley, 1982). The prospect of supplementing our energy needs while at the same time maintaining our soil resources make forage crops an attractive option. A universal component of forage plants is lignin. The usefulness of forage incorporated into diets of ruminants, and forage used as biomass for biological conversion to liquid fuels, is limited by the quantity of Lignin present in the forage. There is worldwide interest in improving the quality of forages through breeding, and one of the most effective ways of improving energy availability is to reduce or alter lignin content of plants. Brown-midrib (bmr) mutations are seen as one way to modify Lignin quality and quantity in a favorable manner (Cherney, 1990). A phenotypic characteristic of this mutation is a reddish-brown pigmentation, which is particularly noticeable in the leaf midrib, that becomes visible in plants at the four- to six-leaf stage. As the stem lignifies, the pigmentation becomes conspicuous, especially in cross-section.The pigmentation is visually associated with lignified rind and vascular bundles. Grass internodes develop basipitally, such that pigmentation initially appears at the top of a developing internode, and moves downward as secondary cell walls are formed and lignin is deposited. The coloration on exposed plant surfaces gradually fades with maturation and can be difficult to see in a mature plant. Pigmentation in unexposed areas is maintained. The pigmentation is associated with lignin, and it persists in the cell wall residue after hemicellulose and/or cellulose are removed from the cell wall.
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
159
11. HISTORY
Brown-midrib(bmr)mutations have been found or induced in maize (Zea mays L.), sorghum [Sorghumbicofor(L.) Moench.], and pearl millet [Pennisetum gfaucum (L.) R. Br.]. Jorgenson (1931) noted that a bmr maize phenotype was first observed and maintained at the University of Minnesota’s University Farm, St. Paul, MN in 1924. This apparently spontaneous mutation segregated as a simple recessive. Jorgenson also noticed that the pigmentation was associated with lignified tissues, and lignin content was reduced in near-isogenic lines. The bmr genotype in maize has since been studied under a variety of environments in several countries (Miku, 1972; Sheldrick, 1979; Weller et af., 1984). Unfortunately, the bmr trait appeared to be linked to yield, which resulted in reduced grain yield as well as stover yield, compared to its normal counterpart (Gallais et af., 1980; Lee and Brewbaker, 1984; Miller et af., 1983; Weller et af., 1985). A trait which reduces the grain component of a silage maize will not be given serious attention, at least not in the United States. No attempt has been made to separate the bmr trait from the grain yield trait, since there has been little interest in developing maize varieties for silage in the United States. Whether or not maize will be utilized for silage or for a grain crop is typically not known at planting in the United States. In Europe, where maize for grain is less probable, however, much of the maize production is for silage purposes. Recently, there has been interest in incorporating the bmr trait into maize silage hybrids in Europe. Potential for stalk lodging has been linked to lignin concentration. The lower lignin in bmr genotypes indicates that they could be more susceptible to stalk lodging than their normal counterparts. Two measures of lodging potential are crushing strength and stalk-section weight. Crushing strength was observed to be 17- 26%less and stalk-sectionweight was 17- 26% less in bm, maize crosses than in normal maize crosses (Zuber et af., 1977). The authors suggested that bmr maize should be harvested at or near physiological maturity to avoid lodging problems. Lodging problems have not been observed with the bmr trait in sorghum or sorghum-sudangras (J. D. Axtell, unpublished observations). Chemical mutagenesis was used by Porter et af. (1975), which resulted in bmr mutants in two lines of sorghum. The sorghum bmr trait was backcrossed into sudangrass (Fritz et af., 1981) in 1978. Cherney et af. (1988) produced a bmr genotype in pearl millet through the use of chemical mutagens. Since that time, naturally occurring bmr mutants in pearl millet have been identified in Africa (S. C. Gupta, personal communication). Throughout this chapter, the acronym “bmr”wil1be used to refer to the
160
J. H. CHERNEY et ul.
brown-midrib trait in general, and where a specific bmr gene is discussed, its acronym will be used if it has one (e.g., bm, maize).
111. GENETICS
Jorgenson ( I 93 1) was the first to study the genetics of the bmr trait. Besides showing that the trait segregated as a simple recessive in maize, he determined the bmr locus was part of the Pr-V, linkage group of chromosome five. Jorgenson (1 93 1) also studied other bmrlines and noted that at least two of them were associated with different bmrgenes. Four distinct bmr loci have been identified in maize (Kuc’ and Nelson, 1964) and are labeled bm, to bm,, with bm, the first gene described by Jorgenson (1931). Patterns of genetic variation in one species are likely to occur in closely related species.This principle provided a basis for a search for bmrgenotypes in species other than maize. One method of producing genetic alterations is through mutation breeding (Auerbach, 1976). Effective mutagens must produce a large number of mutations that are nonlethal, and monofunctional alkylating agents are particularly effective in producing nonlethal mutations. Both diethyl sulfate (DES) and ethylmethanesulfate (EMS)have been used. These compounds transfer one of their alkyl groups to N-7 of guanine and this results in some substitution of AT for GC base pairs during replication. A seed treatment producing a rate of 50% seed kill is generally considered to produce a maximum number of desirable mutations. Porter et al. ( 1975)treated seed from two grain sorghum lines with DES by soaking 75 g of seed in 1 or 2 ml DES 1-I. Plants were then screened for bmr mutants after two generations of selfing. They noted that pigmentation was most conspicuousin the midrib on the abaxial side ofthe leafand in the stem where it was covered by the leaf sheath. Nineteen bmr phenotypes were identified, and there was a range in color intensity of the bmr phenotype. Six phenotypes were quickly eliminated due to sterility or poor expression. Three promising phenotypes, bmr,, bmr,, ,and bmr,, ,were backcrossed into both grain sorghum and sudangrass [Sorghum bicolor (L.) Moench.] (Fritz et al., 198 1). The bmr trait was expressed in both grain sorghum and sudangrass types after only one backcross. Allelism tests of the bmr sorghum mutants produced inconsistent results (Bittinger et al., 1981). Twelve bmr phenotypes were backcrossed with nuclear male sterile germplasm from a broad-based, random mating population. Some crosses between bmr,, and bmr,, indicated that the genes were allelic, because progeny were all bmr phenotypes. Other crosses indicated that bmr, was located in a different place on the genome.
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
161
Cherney et al. ( 1988)treated pearl millet seed with DES or EMS at several rates. Seed treated with EMS had essentially 100%emergence, indicating that a small number of favorable mutations occurred. Seed treated with DES, however, ranged from 5 to 50% emergence, depending on the rate of DES applied. After selfing plants, one bmr phenotype was observed in one DES-treated head row in the M, generation. As with maize and sorghum, it appeared to be inherited as a simple recessive. Development of bmr mutants through the use of mutagens has been attempted in several other species. These attempts have proven unsuccessful in cool-season grasses and cereal grains (J. H. Cherney, 1984 unpublished observations). The bmr mutation may be specific to C4 species, since it has not been observed in any C, species. Hundreds of induced mutations have been described in crop species other than maize, sorghum, or pearl millet, but to our knowledge no bmr phenotype has been observed.
IV. LIGNIN BIOSYNTHESIS
A. BIOCHEMISTRY Historically, emphasis on understandingsynthesis and degradation of cell walls was primarily restricted to woody species. As a result of interest in sustainable animal agriculture and in herbaceous biomass conversion to fuels, more effort is being expended to understand lignification in herbaceous species. Lignin develops as an amorphous matrix in the plant cell wall, and is believed to form covalent cross-linkages with hemicellulose, but not with cellulose. Lignin content of herbaceous plants is much lower than that of woody species, yet lignin still has a major impact on degradation of cell wall components of herbaceous plants. A better understanding of the lignification process will facilitate development of approaches to modify lignin content of tissues. The amino acid phenylalanine is converted to trans-cinnamic acid by phenylalanine ammonia-lyase (PAL).This is considered the first committed step toward synthesis of lignin (Hagerman, 1987). In herbaceous grasses, tyrosine is also utilized for synthesisof phenolic compounds. Phenylalanine, however, is considered the primary substrate for synthesis of phenolic compounds. Activity of PAL fluctuates over relatively short periods of time, and changes in activity are related to environmental factors (Camm and Towers, 1973). Activity of PAL has been closely correlated with synthesis of phenolics (Davies, 1972), and specifically with an increase in lignin accumulated
162
J. H. CHERNEY et ul.
(De Jaegher et al., 1985). Levels of PAL may be regulated by feedback control of PAL by cinnamic acid and its derivatives (Hahlbrock and Grisebach, I979), or it may be regulated by changes in the rate of degradation of PAL (Lawton et al., 1980). trans-Cinnamic acid is converted to pcoumaric acid (PCA) by adding a hydroxyl group to the para position of the aromatic ring. This reaction is catalyzed by cinnamic acid-4-hydroxylase(CAH) (Malmstrom, 1982). Following formation of PCA, a second hydroxylation involving conversion of PCA to caffeic acid (CA) is catalyzed by pcoumaric acid 3-hydroxylase (PAH). Methylation of CA yields ferulic acid (FA) and is catalyzed by catechol-0methyltransferase (OMT). The phenolic hydroxyl is transmethylated, with S-adenosyl L-methionine as the methyl donor (Gross, 1985). Ferulic acid is converted to 5-hydroxy FA by FA-5-hydroxylase (FAH). The final phenolic lignin precursor is formed by changing 5-hydroxy FA into sinapic acid (SA) through the action of OMT. pCoumaric, ferulic, and sinapic acids are converted to their coenzyme A (CoA) esters by hydroxycinnamicacid CoA ligase, a reaction requiring ATP as a cosubstrate and Mg2+as a cofactor. This class of enzyme is specific toward hydroxylated cinnamic acids and appears to have the highest affinity for PCA (Gross and Zenk, 1974). The CoA esters are then reduced to their respective aldehydes by cinnamoyl-CoA:NADP oxidoreductase. Cinnamyl aldehydes of pcoumaric, ferulic, and sinapic acids are reduced to pcoumaryl, coniferyl, and sinapyl alcohols (monolignols) by cinnamyl alcoho1:NADPH oxidoreductase (CAO). Monolignols may be converted to glycosides to protect the alcohol from free radical formation during transport to the cell wall (Hagerman, 1987). Cinnamyl alcohols(or glycosides)are oxidized to free radicals by cell wall bound peroxidasesin the presence of HzOz,and the free radicals condenseto form the lignin polymer. Coniferylalcohol has been suggested as the primary substrate for peroxidase-mediated lignification (Mader and Fussl, 1982).
B. MODIFICATIONS IN ENZYMEACTMTY Modification of the lignification process can be caused by a change in the activity of enzymes involved in lignification. Bucholtz et al. (1980) assayed bmr, and normal sorghum for lignin, using a phloroglucinol-HC1assay. The assay indicated that the bmr genotype accumulated three times more aldehyde groupsin its lignin than normal. The enzyme which catalyzesreduction of aldehyde intermediates, CAO, was much less active in bmr tissues than normal. G. A. Foxon (personal communication) also noted differences be-
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
163
tween normallbmr, sorghum in CAO activity, while Grand et al. (1985) found no differences in CAO activity between normallbm, maize. Preliminary evidence suggests that bm maize may be deficient in CAO, compared to normal maize (G. A. Foxon, personal communication). Activity of CAO was lower in bmr, sorghum and pearl millet internodes compared to normal (J. H. Cherney, 1990 unpublished observations), with no differences found in CAO activity between bm, and normal maize internodes, which agrees with the results of both Bucholtz et al. (1980) and Grand et al. ( 1985). Grand et al. ( 1985)assayed for six enzymes involved in lignin synthesisin bm, and normal maize. Of the six, only OMT activity differed between the two genotypes. Activity of OMT was 10-fold lower in bm, compared to normal. While the decrease in lignin concentration was associated with OMT activity, it was not determined that decreased OMT activity was responsible for decreased lignin concentration. The authors suggested that OMT activity could be used to aid in selection of bm, genotypes in breeding studies, because differences in this trait can be detected in very early plant growth stages well before differences in lignin concentration are apparent. Similar results with OMT were found in bm, maize, however, bmr,, and bmr,, sorghum were found to be OMT deficient, compared to normal (G. A. Foxon, personal communication). The presence of 5-hydroxyguaiacyl monomeric units in bm, maize and not in its normal counterpart suggested that the methylation of 5-hydroxyferulic acid to sinapic acid, catalyzed by OMT, is a primary step affected by the bm, mutation (Lapierre et al., 1988). Large differences in OMT activity between elongating internodes of bm, and normal maize were noted by J. H. Cherney et al. (1989b),agreeing with the results of Grand et al. (1985). In the same study (J. H. Cherney et al., 1989b),activity of OMT did not differ between bmr and normal pearl millet internodes. While these changes in enzyme activities can only be circumstantially associated with reduced lignin content, it is likely that the phenotypically similar brown-midrib types result from different modificationsof the lignification pathway.
,
V. FORAGE QUALITY -PLANT RESPONSE A. LIGNINAND TOTAL FIBER Lignin may help impart structural integrity to the plant, but because it is considered indigestible, it is an undesirable component of a feedstock. Lignin also limits the breakdown of cell wall polysaccharides (Jung and Fahey,
164
J. H. CHERNEY et
a/.
1983). Neutral detergent fiber (NDF), also reported as cell wall constituents (CWC),is a gravimetric measure of total cell wall minus pectin and biogenic silica (Goering and Van Soest, 1970). Kuc’ and Nelson (1964) found the bm, mutant of maize to have 86% as much lignin as its near-isogenic normal counterpart. They introduced the term core lignin to refer to the nonalkali-labilepolyphenoliccomplex in cell walls, and suggested that it exerted control over the incorporation of phenolics into the lignin polymer. Research by this same group (Kuc’ et al., 1968) concluded that double mutants (two maize mutant genes combined in a single genotype) resulted in lower lignin content than single mutants. When leafblade,leaf sheath, and stem of maize single and double mutants were examined, lignin concentration was lower in all component parts than in their normal counterparts (Muller et al., 1971). Neutral detergent fiber concentrations tended to be lower in bmr mutants compared to normal. There were no significant differences, however, in lignin concentration between the single and double maize mutants, disagreeing with Kuc’ et al. (1968). Lignin concentration in maize bmr mutants was found to be consistently lower than their normal counterpartswhen harvestedapproximately 3, 6, or 9 weeks after grain harvest (El-Tekriti et al., 1976). Stem lignin averaged approximately 62 g kg-’ in bm, maize, compared to 120 g kg-I in normal maize. Neutral detergent fiber concentrations were significantly lower for bm, compared to normal for leaf and stem at all three harvest dates. Depending on harvest date and location, however, the difference in NDF between bmrlnormal ranged from 26 to 91 g kg-’. Three sampling dates of bmrlnormal maize in England yielded results similar to previous trials (Sheldrick, 1979), and the authors suggested giving serious consideration to introducing bmr maize into breeding programs in Europe. Chemically induced bmr mutants in sorghum vaned in lignin concentration, but were typically lower than their normal counterparts (Porter et al., 1978). Reductionsin lignin concentration in mature plants ranged from 5 to 25% in leaves and 5 to 5 1% in stems. Of the 19 reported mutants, only two bmrlnormal pairs differed in NDF concentration of stems, with no differences between bmrlnormal leaves in NDF. Immature 4-, 8-, and 10-weekold sorghum forage (bmr,, and bmr,,) was analyzed for lignin concentration, and differencesin lignin concentration between bmr and normal leaves were apparent at 4 weeks (Hanna et al., 198 1). When bmr, was backcrossed into several sudangrass lines (Fritz et al., 198 1), bmr genotypes averaged 68 g kg-’ NDF less than normal genotypes backcrossed into sudangrass. As expected, lignin concentration was lower in bmr sudangrass compared to normal. Sorghum-sudangrass hybrids have been developed that express the bmr trait, and seed companies are developing bmr hybrids for possible commercial release (Kalton, 1988).
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
165
Maize and sorghum mutants and their normal counterparts were compared to bmr/normal pearl millet to find out if all three mutants were phenotypically similar from a compositional standpoint (Table I). Plants of all three species were harvested at anthesis. Lignin concentration in bmr pearl millet was 40 g kg-I compared to 50 g kg-I for the normal genotype, with similar differences between bmr and normal for maize and sorghum. Neutral detergent fiber was significantly lower in bmr genotypes of all three species compared to their normal counterparts. Lignin concentration in bmrlnormal roots followed the same pattern as the herbage for maize, sorghum, and pearl millet (J. H. Cherney, 1989 unpublished observations). Reduced lignin concentration in the roots may make bmr genotypes more susceptible to pathogens, such as Striga species on sorghum. Brown-midrib maize has been found to be more susceptible to Fusarium muniliforme Sheld. infection (Nicholson et al., 1976),although it may be possible to screen for resistance in bmr breeding stocks. Lignin accumulation in a single stem internode was followed from the beginning of internode elongation through to maturity in bmrlnormal sorghum (Cherney et al., 1985a). Elongation of a single sorghum internode (7) lasted 1 week and was complete 57 days after planting. Fiber composition of this internode was followed until 102 days after planting. As expected, lignin accumulation increased throughout maturation. When the lignin concentration data are transformed using the first derivative, the rate of lignin synthesis can be estimated, as well as the extent of synthesis (Fig. 1). Extent of lignin synthesis was relatively similar for both genotypes, but the bmr genotype had a much lower rate of lignin synthesis early in internode development. It also is possible to follow daily lignin accumulation in an elongating leaf blade or internode. Rate of elongation in developing internodes in bmrlnorma1 bm, maize was calculated based on daily length measurements taken for five days (Cherney et al., 1989b). Internodes were marked with spaced pins, and the distance between adjacent pin holes on the following morning was used to determine the length of the meristematic/elongating region of the internode. With data from rate of elongation and length of meristematic region, it was possible to isolate internode segments above the meristem/ elongation segment which correspond to the daily growth increment of the internode. Due to small sample size, procedures were modified for sequential, gravimetric analysis of fiber (Cherney et al., 1985b). Upon analysis, these segments revealed a similar composition in the internode meristem between bmr and normal. In the one-day-old segment just above the meristematic/elongating segment, however, normal maize was significantly higher in both NDF and lignin concentration compared to the bmr genotype. The characteristic brown coloration associated with the modified lig-
Table I
Leaf:Stem Ratio and Concentrations of Nitrogen, IVDMD, NDF, Celldose, Hemicellulose, and Permangaate Lignin in Total Forage of Normal and bmr Genotypes of Pearl Millet, Sorghum, and Maizeub*' Pearl millet Component
Leaf:Stem
Normal 2.2
bmp
Sorghum Significance
2.2
NS
Normal I .3
25 659 656 305 303 50
28 726 635 293 298
40
Significance
1.7
**
Normal 1.1
** ** ** * NS
**
19 568 704 311 337 55
21 642 678 307 325 43
Adapted from Cherney el a[. (1988). IVDMD,In vifrodry matter digestibility; NDF, neutral detergent fiber, bmr, brown-midrib. 'Significant difference between normal and bmr at the 0.05* and 0.01** probability levels, respectively, based on an F-test. NS = Not significant (P> 0.05).
bm#'
Significance
1.1
NS
g kg-I dry weight
g kg-I dry weight
g kg-' dry weight Nitrogen IVDMDb NDF6 Cellulose Hemicellulose Lignin
bmp
Maize
NS
** **
NS
** **
18 630 638 31 1 278 49
19 748 598 289 277 29
NS
**
#
** NS
**
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
G
;.
167
2
M
z 1
L5 0
6 0 7 0 SO 90 100 Age (days a f t e r planting)
Fig. 1. Rate and duration of Lignin synthesis of Internode 7 of normal and brnr sorghum lines. Lignin concentrationwas determined for Internode 7 at eight maturities, and the first derivative of the resulting relationship (lignin concentrationversus time) was calculated. (-), Normal; (----), brnr. (After Cherney er al., 1985a.)
nin did not appear until several days after a change in lignin content was measured between bmr and normal.
B. PHENOLICS There are generally only two major alkali-labile phenolic monomers in forage plants, PCA and FA, although exceptions occur (J. H. Cherney et af., 1990). These phenolics are possible cross-linkages between core lignin and hemicellulose (Hartley, 1972).Approximately20%of the 1.O Malkali-labile PCA and FA in grasses is actually water soluble, compared to about 80%for most legumes (Cherney et af., 1989a).The relatively high solubility of PCA and FA in legumes indicates that most of the alkali-labile PCA and FA in these species is not in the form of esters bound to hemicellulose or lignin. Along with the fact that legumes are very low in alkali-labile phenolics, the high solubility of the phenolics present explains why base treatment (eg., NaOH treatment) of legumes would not be expected to influence forage digestibility as it does in grasses. Grasses, particularly maize and sorghum, contain very high levels of alkali-labile PCA and FA, approaching 30 g kg-l of the dry weight in mature stems (Cherney et af., 1989a). Kuc’ and Nelson (1964) were the first to observe that brn maize contained significantlyless alkali-labile PCA than its normal counterpart. The same research group (Gee et af., 1968; Kuc’ et af., 1968)also noted that ratios of nitrobenzene oxidation products, syringaldehyde, vanillin, and phydroxybenzaldehyde, were considerably different in bm, core lignin compared to normal. Not only was the quantity of lignin reduced in bmr genotypes, but the quality of lignin also was affected. Recent research (Gaudillere and Monties, 1989) found syringkguaiacyl ratios for
,
168
J. H. CHERNEY et al.
bm, maize were significantly lower than those found by Kuc’ et al. ( 1968), and were similar to the ratios found in bm, . Changes in the ultraviolet characteristicsof bmr maize compared to normal supported the contention that phenolic composition was modified in bmr mutants (Yu et al., 1972). Results of Hartley and Jones (1978) dealing with alkali-labile phenolics in bm, maize agreed with Kuc’ and Nelson ( 1964). Normal and bmr maize, sorghum, and pearl millet were compared for concentrationsof alkali-labile phenolics (Cherney et al., 1988). A relatively consistent pattern was found for PCA, with a large reduction in PCA in the bmrgenotypes compared to normal for all three species (Table I). There was over a 60% reduction in PCA in bmr sorghum compared to normal. Genotypes did not differ in FA concentration in maize or pearl millet, but did differ in sorghum. The PCA :FA ratio was much smaller for the bmr genotypes compared to normal and agreed with similar results found in bmr sudangrass (Cherney et al., 1986). Preliminary results with an African bmr pearl millet also indicated a small PCA :FA ratio compared to normal genotypes (J. H. Cherney, 1990, unpublished). The PCA :FA ratio is considered to be negatively related to cell wall digestion (Burritt et al., 1984). Similar reductions in PCA concentration have been found in maize, sorghum, sorghum-sudangrass, and pearl millet bmr genotypes (Weller et al., 1984; Akin et al., 1986b; Fritz et al., 1990). Other minor alkali-labile phenolic monomers are not considered to crosslink hemicellulose and core lignin, but these are likely involved in phenylpropanoid metabolism. Differences between bmr and normal in these phenolics were significant, but were not consistent across the three species (Cherney et al., 1988). The pattern of minor phenolic differences in bmr/ normal sorghum was the same as was found in bmrlnormal sudangrass and sorghum-sudangrass hybrid (Cherney et al., 1986), indicating that this pattern may be consistent within bmr genes. The same gene (bmr,) was used in both experiments for the sorghum, sudangrass, and the hybrid between the two. The model system for isolating meristems and daily growth increments of internodes described previously (Cherney et al., 1989b) was used to determine the soluble portion of alkali-labile phenolic monomers in elongating bmJnormal maize internodes. The soluble proportion of alkali-labile PCA decreased from 46% in the bmr meristem to 1 1% in a three-day-old segment of the elongating internode. A similar decrease occurred in the normal genotype, Solubility of most of the phenolics was high in the meristem and then decreased as secondary cell wall formation and lignin deposition occurred. Solubility of one of the minor phenolics, vanillin, differed between bmr and normal genotypes. The bmr meristem contained 65%soluble vanillin, compared to 44% for normal, and decreased to 37% soluble vanillin in
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
169
three-day-old bmr segments, compared to 0%soluble vanillin in the normal genotype. A similar difference between bmrlnormal was found for phydroxybenzaldehyde, indicating that solubility of minor phenolics is influenced by the bmr trait. In a study with mature leaves and stems of bmr,/ normal sorghum, solubility of PCA and FA were less than 30% (D. J. R. Cherney et al., 1991).
C. NEUTRAL CELLWALLSUGARS Besides phenolic compounds, the plant cell wall is composed of the structural polysaccharides, cellulose and hemicellulose. Cellulose is composed of glucose, whereas hemicellulose contains a variety of 5- and 6-carbon neutral sugars, although most are xylans. Neutral sugars can be extracted from NDF residue of bmrlnormal genotypes using sulfuric acid. This results in only a partial removal of neutral sugars from the NDF. Also, it is not possible to determine what proportion of the isolated glucose is from hemicellulose as opposed to cellulose. Weller et al. (1984) measured the xylose concentration of bm,/normal maize and found similar concentrations, averaging 70-75% of the total hemicellulose. No other neutral sugars were measured. Concentrations of neutral sugars in the cell wall of digesting brnrlnormal sudangrass and sorghum-sudangrass did differ between genotypes (Cherney et al., 1986). Brown-midrib genotypes were higher in arabinose and xylose than their normal counterparts throughout digestion. Glucose concentrations were lower in the bmr genotype, with no difference between genotypes in galactose concentration. Fritz et al. (1990), however, found no differences in neutral sugars between bmr and normal sorghum-sudangrass. It is not clear if the bmr mutation has any effect on hemicellulose composition. Hemicellulose composition of bmrlnormal maize and pearl millet has not been studied.
D. NITROGEN CONCENTRATION Although nitrogen content does not directly relate to cell wall digestibility, it is nevertheless a critical forage component necessary for maximum animal performance. Few studies on bmr mutants have included an analysis of nitrogen content. Muller et al. (1974) found that brn , bm,, and the double maize mutant, bm, /bm, were not significantly different in nitrogen concentration from the normal genotype. Similar results were noted for the bm,
,
170
J. H. CHERNEY et al.
maize mutant and its normal counterpart (Weller et al., 1984). Maize (bm,) and sorghum (bmr,) also were found not to be different from normal genotypes (Cherney et al., 1988), but bmr pearl millet was significantly higher than normal in nitrogen concentration (28 g kg-l for bmr versus 26 g kg-' for normal). Before bmr varieties are released for public use, nitrogen content may need to be investigated more thoroughly.
E. DIGESTIBILITY Forage utilization is limited by the extent of digestion of forage fiber. In vitro dry matter digestibility(IVDMD)can be determined by digestingdried, ground forage in a buffer containing a batch culture of rumen microorganisms. The two-day digestion is typically followed by a one-day treatment in acid-pepsin to simulate the action of the ruminant's digestivesystem (Marten and Barnes, 1980). Since correlation coefficients up to -0.95 between lignin content and in vivo dry matter digestibility have been reported (Oh et al., 1966),digestibility would be expected to be improved by the bmr mutation. Barnes et al. (197 I) analyzed three harvests of bm,, bm,, bm, /bm,, and normal maize for IVDMD. The double mutant, along with bm,, had significantly higher IVDMD than the normal type, averaging up to 100 g kg-l higher IVDMD than normal. The bm mutant was not significantlydifferent from normal at the first harvest, and was only moderately better than normal at the other two harvests. The improvement in digestibility was found in stem, leaf blade, and leaf sheath of the bmr mutants. Other studies with the maize mutants confirmed the earlier digestibility findings (Lechtenberg et al., 1972; Colenbrander et al., 1973; Cymbaluk et al., 1973).Digestibility of bm, stems averaged 536 g kg-I compared to 444 g kg-l for the normal genotype (El-Tekriti et al., 1976). Digestibility using a commercial cellulasewas 74 1,721, and 70 1 g kg-l for leafblade, leaf sheath, and stem, respectively, of bm, maize, compared to 602,6 19, and 572 g kg-' for leaf blade, leaf sheath, and stem, respectively, of normal maize (Hartley and Jones, 1978). Hartley and Jones (1978) also found that digestibility of cell walls of the maize mutant was enhanced more by NaOH treatment than with the normal genotype, Porter et al. (1978) tested all possible bmr/normal sorghum pairs for IVDMD and found most bmr genotypes to be significantly higher in digestibility, compared to their near-isogenic normal counterpart. Increases in digestibility were as much as 33%and correlated well with lignin concentra-
,
BROWN-MIDRIB, LOW-LIGNIN M U T A N T S
171
' M 100 * M
60
J 0
10 20 3 0 40 5 0 60 7 0
Time, h Fig. 2. Apparent lignin digestionof Internode 7 normal and bmr sorghum lines from 0 to 72 hours of in vifro incubation. Bayes least significant difference (BLSD) approximately P = 0.05. (W), Normal; (O-O),bmr. (After Cherney ef al., 1986.)
tions. Similar increasesin IVDMD were found in bmr sorghum, sudangrass, and sorghum-sudangrass hybrids (Fritz et al., 1981; Akin et al., 1986b; Cherney et al., 1986).Hanna et al. ( 1981) determined that immature forage from bmr,, and bmr,,,were consistently higher in IVDMD over three years, but three other mutants were not consistently higher in IVDMD than their normal counterparts. These results indicate a possible environmental interaction, and point out the need for evaluation of forage quality over multiple environments. Digestibility of bmr pearl millet at anthesis was 726 g kg-', compared to 659 g kg-I for the normal genotype (Cherney et al., 1988). Digestibility responses of bmr sorghum and maize in the same study were consistent with previous studies. Similar differencesin digestibility were found between bmr and normal African pearl millet genotypes (J. H. Cherney, 1990, unpublished). Results obtained with bmrlnormal genotypes are consistent with the theory that a 10 g kg-I decrease in lignin concentration will result in about a 40 g kg-I increase in digestibility (Bula et al., 1981). When the amount of lignin present before and after forage digestion is measured, the amount is typically lower after digestion, even though lignin is generally considered to be indigestible. A portion of the measured lignin disappeared during digestion from both bmr, and normal sorghum (Cherney et al., 1986). The extent of this apparent lignin digestion was significantly lower in bmr compared to normal, such that the concentration of lignin in the cell wall of both genotypes merged at 138 g kg-I NDF as digestion progressed (Fig. 2). In another study, digestion of mature bmrlnormal sorghum internodes followed a similar pattern, with lignin concentration of both bmr and normal genotypes leveling off at about 140 g kg-I NDF after
172
J. H. CHERNEY et al.
96 hours of digestion. This implies that bmr lignin is more resistant to degradation.
F. In Vitro RATE OF DIGESTION Along with digestibility and forage intake by ruminants, the rate of cell wall digestion is one of the criticalparametersdeterminingthe suitability of a particular forage for ruminants. Rate of digestion is a function of cell wall composition and cell wall structure. Since the bmr trait significantly modifies both cell wall composition and structure, it is logical to assume that there is potential for the bmr trait to influence rate of digestion. In vitro rate of cell wall digestion is determined by incubating ground forage samplesat timed intervals in an in vitro system, and then determining the amount of NDF remaining after digestion. Assuming first-order digestion kinetics, a linear regression can be fit to the time-digestion relationship, with a digestion rate constant obtained from the slope of the regression line. Nonlinear regression also may be used (Moore and Cherney, 1986). The regression also can be used to estimate a lag time, which is the length of time between initiation of incubation and actual digestion of fiber. Results with maize stover silage indicated no significant difference in digestion rate constants between bm, and normal genotypes(Lechtenberget al., 1974). Stover of sudangrassand sorghum-sudangrassbmrlnormal pairs was harvested and analyzed for in vitro rate of digestion (Cherney et d., 1986).Rate constants for digestion of NDF, cellulose, and hemicellulose did not differ between plant materials, but rate constants for bmr genotypes averaged -0.083 hr',compared to -0.064 hr-' for normal genotypes. Digestion lag times did not differ among genotypesand averaged 4.0 hours. These results disagree with Fritz et al. ( 1990),who found no significant differencesin NDF rate constantsbetween bmr and normal sorghum-sudangrassmorphological components, although plants were sampled at a more immature stage in the Fritz et al. (1990) study. Rate of digestion was determined in normal and bmr, sorghum internodes harvested at four maturity stages from cessation of internode elongation to cessation of lignin accumulation in the internode (Cherney et al., 1985a). Digestion rate constants for NDF were significantly higher in the bmr genotype (-0.046 h i ' ) compared to normal (-0.035 hr-'). Rate of NDF digestion also was determined in bmrlnormal pearl millet stems harvested at anthesis at two locations (Cherney et al., 1988). The in vitro NDF rate constant, calculated using nonlinear procedures, was much larger for the bmr genotype than its normal counterpart.
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
173
A study by E. M. G. Thorstensson, D. R. Buxton, and J. H. Cherney ( 1990 unpublished observations) compared digestion rate constants of bmrlnormal maize at two locations, as well as bmr/normal sorghum and pearl millet. Lower internodeswere used as samples,since lignin differencesbetween bmr and normal are most pronounced in these tissues, and, therefore, should be the most likely plant component to exhibit a difference in rate of digestion,if it exists. In sorghum, the digestion rate constant of bmr, was significantly higher than normal, agreeing with previous results (Cherney et al., 1985a, 1986,1988).In the first reported study with bmr,, sorghum, however, bmr,, rate constants did not differ from normal. The bmr pearl millet rate constant was significantly different from its near-isogenic counterpart, agreeing with Cherney et al. ( 1988).Neither bm or bm, maize rate constantsdiffered from normal, agreeing with previous work with maize (Lechtenberget al., 1974). When a surface area fermentability index (Fisher et al., 1989) was used to adjust rate constants for extent of digestion, only bmr pearl millet was significantly different from its normal counterpart. It is not clear, however, whether or not this correction for rate constants is a reasonable one. In the same study, (E. M. G. Thorstensson, D. R. Buxton, and J. H. Cherney, 1990 unpublished observations) the indigestible residue :lignin ratio was higher for bmr than normal plants, indicating that bmr lignin was more inhibitory per unit of lignin to digestion than was normal lignin. This relates to early work by Cherney et al. (1985a, 1986), where it was noticed that bmr lignin was more resistant to “apparent” lignin digestion than the normal genotype. KUC’and Nelson ( 1964),found that bm, maize lignin was more resistant to nitrobenzene oxidation than normal maize.
,
VI. PHENOLIC-CARBOHYDRATE COMPLEXES Cell wall polysaccharidedigestibility in the rumen may be limited due to the association of aromatic compounds, such as phenolic acids, with the cell walls (Hartley, 1985). Cell wall phenolic-carbohydrate complexes may decrease utilization of forage fiber by limiting microbial digestion of bound structural carbohydrates. As cell walls are degraded by ruminal microorganisms, soluble phenolic - carbohydratecomplexes are released into solution (Gaillard and Richards, 1975; Neilsen and Richards, 1978; Jung et al., 1983a,b), and these complexes may affect digestion. Akin et al. (1988) suggested that information on the interactions of mixed ruminal micro-organisms is necessaryto clarifythe antiquality aspect of phenolic compounds in forages. The bmr trait in forages offers a unique model for studying this problem, without confounding influences of environment or maturity.
174
J. H. CHERNEY et al.
0 500
1000
1500
2000
ml e l u t e d Fig. 3. Profile of extracts from normal and bmr sorghum stems after partial digestion with cellulase and separation using size exclusion chromatography on Sepharose CG6B. Peaks were bmr. (From D. J. R. Cherney, detected using W 320-nm absorbance.(DCI),Normal;(0-O), J. H. Cherney, J. A. Patterson, and J. D. Axtell, 1990 unpublished observations.)
A. ISOLATIONAND CHARACTERIZATION Complexes consisting of FA or PCA bound to glucan or xylan groups have been isolated from forage cell walls (Hartley, 1972; Morrison, 1974). Evidence suggests that these phenolics are often esterified to carbohydrates within the cell wall (Mueller-Harvey et al., 1986). Jung (1988), using size exclusion chromatography, suggested that phenolic-carbohydrate complexes released from forage cell walls during fermentation contain some high-molecular-weight material, but that most of the phenoliccarbohydrate complexes are of low molecular weight. Normal and bmr, sorghum stems, leaf blades, and leaf sheaths were partially digested with highly purified cellulase to release complexes (D. J. R. Cherney, J. H. Cherney, J. A. Patterson, and J. D. Axtell, 1990 unpublished observations). Extracts were separated based on molecular size using size exclusion chromatography (Sepharose CG6B, Sigma Chemical Co., St. Louis, MO); fractions were eluted using 0.02 MHCl. Two major peaks were identified based on UV320-nm absorbance(Fig. 3). This was observed for all morphological components. The first peak contained compounds of high molecular weight, composed primarily of PCA, xylose, and arabinose. The second peak consisted of relatively smaller molecular weight compounds containing primarily FA and xylose. Increases in the xylose :arabinose ratio have been associated with an increase in linear xylan relative to branched xylan (Morrison, 1974). Elution profiles observed were different for bmr and normal genotypes, which indicated that the relative proportion of these complexes released from cell walls is affected by the bmr mutation. Data indicated compositional differences in hemicellulose between bmr and nor-
BROWN-MIDRIB, I,OW-I~l(;NlN MUTANTS
$
175
1.2
d d L
p 0.8 5.4
0
$ 0 4
-4
0.0 200
400
600
800 1 0 0 0
ml e l u t e d Fig. 4. Profile of extracts from normal and bmr sorghum stems after partial digestion wth cellulase and separation using size exclusion chromatographyon SepharoseCL-6B followed by further separation using Sephadex LH-20 Peaks were detected using UV 320-nm absorbance (UU),Normal, (O-O),hmr (From D. J R. Cherney. J H. Cherney, J A Patterson, and J D Axtell, 1990 unpublished observations.)
mal, among components, and in isolated peaks. Phenolic-carbohydrate complex absorbance values of the first peak were higher for hmr than the normal genotype. The second peak was further separated using Sephadex LH-20 (Sigma Chemical Co.); fractions were eluted using 0.5% acetic acid. Several major peaks were identified based on UV 320-nm absorbance (Fig. 4). Primary phenolics in peaks were PCA and FA, with small amounts of vanillin also detected. Further separation using semipreparative high performance liquid chromatography (HPLC) are required to further identify these phenolic - carbohydrate complexes.
B. INFLUENCE ON RWEN MICROORGANISMS Several studies have investigated the inhibitory potential of phenoliccarbohydrate complexes on fiber digestion. Jung ( 1988) reported that phenolic - carbohydrate complexes released during rumen fermentation of alfalfa (Medicago sativa L.) hay, smooth bromegrass (Bromus inerrnus Leyss.) hay, and maize silage inhibited fermentation of structural carbohydrates. Fermentability was correlated negatively with phenolics, but varied among various fractions collected. Jung ( 1988) indicated that alkaline-extractable phenolic esters that were most inhibitive of carbohydrate fermentation were the benzoic acids and aldehydes, rather than the cinnamic acids normally associated with reduced cell wall digestion (Burritt el al., 1984; Hartley, 1972). Theodorou et al. (1987) observed inhibition of dry matter disappearance with phenolics extracted from cell walls of maize stems.
176
J. H. CHERNEY et al.
Griggs et al. (1989) isolated phenolic- carbohydrate complexes from cell walls of maize stems, and observed suppression of bacterial growth with peaks high in FA. Akin et al. (1988) observed no to slight inhibition of bacterial growth with addition of phenolics extracted from cell walls of barley straw. In vitro esterification of forage cell walls with PCA or FA resulted in decreased digestibilityof structuralcarbohydrates(Sawai et al., 1983). These data indicate that low-molecular-weight phenolics may be a factor responsible for inefficient utilization of forage cell walls by limiting microbial degradation of bound structural carbohydrates (Akin et al., 1988). When isolated extracts of Sepharose CG6B fractionated phenolic carbohydrate complexes were added to digesting bmr, sorghum leaves in vitro, digestion was greatly inhibited by peak 2 extracts (D. J. R. Cherney, J. H. Cherney, J. A. Patterson, and J. D. Axtell, 1990 unpublished observations). Peak 2 extracts from the bmr sorghum reduced digestibilitymore than normal extracts, possibly because more cell wall was solubilized during digestion of bmr types than their normal counterparts (D. J. R. Cherney et a)., 1990 unpublished observations). Peak 2 extracts were further separated using Sephadex LH-20 and resulting fractionstested for digestion-inhibiting activity. Only one of these fractions(peak 1) appeared to significantly inhibit NDF digestion (Table 11).Further studies using HPLC are required to determine if the inhibitory effect is due to one or more phenolic-carbohydrate complexes.
VII. PLANT MORPHOLOGY AND ANATOMY A. LEAVESOF NORMAL AND Brmn-Midrb,, SORGHUM The leaf laminas, midribs, and sheaths in these samples all showed characteristics differentiating the bmr from the normal counterparts, i.e., lignin concentrationand in vitro digestibility(Table111). Percentagesof tissue types within cross sections of the plant parts were calculated to determine if differences in digestibilitywere influenced by variations in anatomy. Tissue types were not different for normal and bmr leaf laminas (Table IV); tissues in leaf sheaths and midribs were not calculated as the laminas, but microscopic observations did not indicate any substantial variations in the anatomies between normal and bmr parts. Percentages of tissue types in leaf laminas, midribs, and sheaths of normal and bmr pearl millet have recently been determined, and no differences occurred between plant types, except that parenchyma bundle sheaths of bmr plants occupied less cross-sectionalarea
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
177
Table I1 Inhibition of NDF Digestion by Sephadex LH-20Fractionated Extracts, 96 (N= 6)Qb
Fraction
IC 2 3 4 5
6 7 8 9 10
Mean
S.D.
5.8
5.6
4.0
2.8 20.8 15.2 1.9 2.0 2.8 1.9 2.9 2.3
112.4 25.4 -1.4 -5.1 -5.9 -0.6 -0.3 -1.0
'From D. J. R. Cherney, J. H. Cherney, J. A. Patterson, and J. D. Axtell (1990 unpublished observations). NDF, Neutral detergent fiber. The last peak to elute from a Sepharose CL6B fractionation was injected onto a Sephadex LH-20 column and fractions were collected sequentially from void to exclusion volumes. Fractions 3,5,and 8 corresponded to peaks I , 2, and 3, respectively (Fig. 4). Rate was 20 X that amount extracted from 0.25 g sorghum stem NDF. Values are means of bmr, and normal genotypes.
than that in normal plants (Akin, et al., 1991). All availabledata indicatethat plant anatomy does not differ substantially between leaves of normal and bmr plants and, therefore, does not contribute to the greater digestibility in bmr mutants. Histological staining for lignin and phenolics was carried out to determine if greater concentrations of phenolics could be localized within particular tissues. Reactions with acid phloroglucinol, chlorine sulfite, and a series of diazonium salts indicated some variations between normal and bmr plants, but differences were not consistent and, therefore, not indicative of any particular factor in these substrates (Akin et al., 1986a; Akin, unpublished observations). Degradation of individual tissue types was investigated by light and electron microscopy to identify particular cell types responsible for differences between normal and bmr plants. Variations in the digestibility of cell wall types occurred among leaves within normal and bmr types, but in general,
178
J. H. CHERNEY et al. Table I11 Lignin Concentrations and in Yitm Digestibility of Leaves of Normal and bmr,, Sorghum'
Plant type
Permanganate lignin (mg g-I)
In vitro digestibility (mg g-'1
Leaf blade' Normal bmrn Leaf sheath Normal bmr,2 Midrib Normal bmr12
37+ 21.
697+
29. 14*
757 789
NDe ND
146+
554+
771*
* Values in the columns within plant parts differ,
P I 0.05. 'From Akin et af.(1986a,b).
'Includes lamina and midrib. Not determined.
bmr leaves demonstrated a greater degree of degradation consistent with the greater dry weight loss (Table 111). For leaf laminas, a substantial loss of structural integrity occurred after 48 hours incubation with fiber-digesting, rumen bacteria in bmr plants. Such degradation indicates that the digestible tissues such as mesophyll and epidermis were more rapidly degraded, resulting in a loss ofintegrity of leaf structure. The slowly, partially digested tissues, such as parenchyma bundle sheath and adaxial sclerenchyma, showed a greater degradationin bmrplants (Figs. 5 and 6). A similar pattern was found in bm, maize (Grenat and Barry, 1991). In contrast, the highly lignified, rigid support tissues, i.e., xylem and mestome sheath cells, were not degraded in either normal or bmr mutant plants (Figs. 5 and 6). Similar phenomena occurred in leaf sheaths and midribs. In sheaths, digestible tissues, e.g., mesophyll, were totally removed in both plant types, with those in bmr mutants often more rapidly degraded lignified epidermis and vascular bundles were not degraded in either plant type (Figs. 7 and 8). In midribs (Figs. 9 and lo), the parenchyma, which was by far the most prevalent tissue type, was partially degraded in both plant types but more digestible in bmr mutants. The lignified epidermal and vascular tissues were
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
179
Table IV Tissue Types in Leaf Laminas of Normal and bmr,, Sorghumu Tissue Type
Plant type Normal
bmr,,
Adaxial epidermis
Abaxial epidermis
Sclerenchyma
(%I
(%I
12.1 f 1.7 13.1 f 2.4
8.2 f 1.5 9.1 & 2.0
Mesophyll
(%)
Vascular bundle (96)
3.9 f 1.0 3.9 lk 1.1
25.3 k 3.9 25.1 f 3.6
50.5 f 4.4 49.0 f 4.1
(96)
From Akin ef al. (1986a). Values for tissues between plant types were not different, P > 0.05.
not degraded in either normal or bmrplants, thus followingthe trend within other parts of the leaf.
B. STEMSOF NORMAL AND Brown-Midrib PEARLMILLET Differences between normal and bmr pearl millet stems in permanganate lignin were about 43% (75 g kg-I for normal and 43 g kg-' for bmr) and in vitro digestibility differences were about 2OYo (600 g kg-' for normal and
Fig. 5. Normal sorghum leaf lamina incubated 24 hours with Ruminococcurflavefaciens
FD-1 showing removal of mesophyll and epidermis resulting in a residue of cuticle(C),lignified vascular tissues (V), sclerenchyma (S), and small vascular bundles with mostly intact parenchyma bundle sheaths (arrowheads). Bar = 100 pm. (From Akin et al., 1986a.)
180
J. H. CHERNEY et al.
Fig. 6. bmr,, Sorghum leaf lamina, incubated as in Fig. 6, showing greater degradation of parenchyma bundle sheath (arrowheads) and adaxial sclerenchyma(double arrowhead). Cuticle (C) and lignified vascular tissues (V)remain as in normal leaves.Bar = 100 pm. (From Akin et al., 1986a.)
Fig. 7. Normal sorghum leafsheath incubated with rumen fluid inoculum for 48 hr showing degradation of mesophyll and a residue of epidermis (E),sclerenchyma (arrowhead), and lignified vascular tissues (V).Bar = 200 pm. (From Akin et al., 1986a.)
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
181
Fig. 8. brnr,, Sorghum leaf sheath, incubated as in Fig. 3, showing a similar residue to normal of epidermis (E), sclerenchyma (arrowhead), and lignified vascular tissue (V). Bar = 200pm. (Akin ef al., 1986a.)
Fig. 9. Normal sorghum leaf midrib, incubated as in Fig. 3, showing degradation at the periphery (double arrowhead) and to a small extent in the more central part (arrowhead)ofthe parenchyma (P). Epidermis (E), lignified vascular tissues (V), and sclerenchyma (S) were not degraded. Bar = 400 pm. (From Akin ef al., 1986a.)
182
J. H. CHERNEY et ul.
Fig. 10. brnr,, Sorghum leaf midrib, incubated as in Fig. 3, showing degradation at the periphery (double arrowhead) and greater tissue loss in the central areas (arrowheads) of the parenchyma (P) than in the normal midribs. Epidermis (E), lignised vascular tissues (V), and sclerenchyma (S) remain undegraded as in normal midribs. Bar = 400 pm.(From Akin et al., 1986a.)
720 g kg-l for bmr; Akin et al., 1991). Microscopic studies indicated that parenchyma was degraded to a greater extent in brnr stems, while the rind and vascular tissues were not observed to be degraded in either plant type (Figs. 1 1 and 12; Akin et al., 199 1). Parenchyma was the most prominent tissue, occupying about 60%of the cross-sectional area in both normal and bmr plants, and, therefore, was the cell type contributingmost to the differences in digestibility between normal and bmr stems. Investigations of the degradation of tissue types within various parts of normal and bmr plants indicate a similar phenomenon. The most digestible and the slowly, partially biodegradable tissues in both sorghum leaves and pearl millet leaves and stems are degraded by fiber-digesting rumen bacteria to a greater extent in bmr than normal plants. In contrast, the highly lignified,support tissues such as vascular tissues are not degraded in either plant type. The modifications in phenolic composition and concentrations that
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
183
Fig. 11. Normal pearl millet stem incubated 48 hr with Ruminococrur/lavelhciens FD-I showing a residue of rind (R) and most of the parenchyma(P) with embedded vascular bundles (arrowheads).Bar = 400 pm. (From Akin d a/.)
Fig. 12. bmr pearl millet stem, incubated as in Fig. 7, showing a residueof rind (R), isolated vascular bundles (arrowheads),and a small amount of parenchyma(double arrowhead).Bar = 400 pm. (From Akin et uf.).
184
J. H. CHERNEY et a l
influence cell wall digestibility, therefore, occur in the less lignified and at least partially biodegradable tissues. If modifications have occurred in the highly lignified tissues, they apparently are insufficient to influence the digestibility and loss of such tissues. The precise chemical modifications that result in enhanced digestibility require further study. The consistent finding among all normallbmr comparisonsof significantlylower concentrationsof PCA in bmr plants (Akin et al., 1986b) has not been directly related to improved digestibility. Future research should take into account the fact that variations in digestibility exist among tissue types within plant parts, and investigations should concentrate on specific areas (Lee,cell wall types) that demonstrate these differences in digestibility.
VIII. FORAGE QUALITY -ANIMAL RESPONSE Only those forage characteristicswhich interact with some aspect of ruminant nutrition or physiology determine forage quality (Moore, 198 1). Factors influencing voluntary intake and nutrient digestibility are particularly important (Raymond, 1969; Moore and Mott, 1973). Forage quality is controlled by voluntary intake, digestibility, and efficiency of utilization (Milford and Minson, 1965; Raymond, 1969; Moore and Mott, 1973; Minson, 1980). Voluntary intake accounts for much more of the variation in digestible organic matter intake than does digestibility (Raymond, 1969; Hutton, 197 l), and is considered to be the most important factor accounting for differences in forage quality (Moore and Mott, 1973). Numerous studies have been conducted investigatingthe influence of the bmr trait on animal performance, intake, and digestibility. Most of these investigations have been conducted with maize silage, with fewer studies involving sorghum or sorghum-sudangrassand one study to date involving pearl millet.
A. INTAKE Results of studies of the bmr trait on intake are mixed, although most studies indicate that intake of bmr genotypes is higher than that of normal counterparts. In one of the earlier studies investigatingintake of bmr genotypes, Muller et al. ( 1 972) reported that lambs fed bm, maize silage had a 29%greater voluntary dry matter consumption than lambs consuming the normal genotype. All ears were removed so that differing grain yields would
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
185
not mask differences between bmr and normal silages. Permanganate lignin values were 50 g kg-I for bm, and 88 g kg-' for normal. The authors reported a correlation of-0.83 between dry matter intake and lignin content. Rook et al. (1977) observed digestible dry matter and energy intakes by Holstein cows in early lactation were 19-27% greater for bm, maize silage rations than for the normal genotype. Increasedintakeswere consistent over varying forage to concentrate ratios (60 :40 and 85 : 15). Lignin was lower in bm, silage (49 g kg-') than the normal genotype (64 g kg-'). Sommerfeldt et al. ( 1979)fed bm, silage ad libitum to lactating cows 49 - 143 days postpartum supplementing with 1 kg concentrate per 3 kg of milk and only obtained small increases in silage intake. Yearling wethers fed bmr maize silage supplemented with 145 g concentrate per wether per day consumed more dry matter than those fed normal maize with supplementation (Stallings et al., 1982). Cows were supplemented with 1 kg concentrate per 3 kg of milk. Slightly higher intakes of bm, silage rations by cows also were reported by Stallings et al. ( 1977).Studieswith cows by Frenchick et al. ( 1976)and Keith et al. (1979) reported no increase in intake of bm, maize silage over the normal genotype. In the study of Frenchick et al. (1976), only 50Yo of the ration was silage; the remainder included 10% alfalfa silage, 30% ground shelled corn, 7% soybean meal and 4% supplement. Keith et al. (1979) included ear in the silages, resulting in a dilution of the lignin content due to ear:stover ratios of 1.07 and 1.12 for bm, and normal, and considerable amounts of energy in the diet. With supplemented diets, energy requirements of the animals may have been satisfied. Many of these studies have been conducted with diary cows. In a feeding study comparing bm, maize hybrids with their near-isogenic normal counterparts, Gallais et al. ( 1980) reported that all classes of livestock studied (sheep, steers, dairy cows, and dairy goats) had increased intake of bm, hybrids. Voluntary intake is controlled by one of two mechanisms in ruminants, metabolic or physical (Forbes, 1977).For grain and high-quality forage diets (digestible energy > 650-700 g kg-I), the intake mechanism appears to be metabolic (Blaxter et al., 1961; Conrad et al., 1964). Animals would have quit eating due to metabolic fill and not physical fill. Differences between bmr silage and normal silage may have been masked by supplementation, if this were the case. In those studies where bmr maize silage was the primary source of feed, intake of bmr silage appeared to be higher than the normal. Fewer studies have examined the influence of the bmr trait on intake of sorghum or sorghum-sudangrass.Intake of sorghum-sudangrasshay, fed as the sole source of feed, by wethers did not differ in a study reported by Wedig et al. ( 1988).It should be noted, however, that lignin concentrationsbetween the bmr and normal genotypes did not differ, possibly explaining the lack of differencesobserved in intake. Lusk et al. ( 1984)reported higher dry matter
186
J. H. CHERNEY et a/.
intake of bmr,, sorghum than normal sorghum fed as silage by lactating cows. Lignin content of bmr silage was 24% lower than normal sorghum, resulting in a 15% increase in digestibility of the bmr genotype (Lusk et al., 1984). This difference in digestibility was probably sufficient to account for the differences in intake reported. Only one study to date has evaluated voluntary intake of pearl millet (D. J. R. Cherney et al., 1990b).Results of that study, involvingwether lambs fed bmr or normal chopped hay as the sole source of feed are mixed. Hay was fed from two cuttings, one in midseason and another in late season. First-cutting hays did not vary in intake, while wethers fed second-cuttingbmrpearl millet hays consumed approximately 30% more hay than wethers fed second-cut normal hay. Intake of second-cutting hay was less than that of first-cutting hay for both genotypes, indicating environmental influence. The authors reported palatabilityproblems with second-cut normal hays, contributingto differences in intake. With most forage diets, physical control limits intake. With physical control, a ruminant’s intake is limited by digestive tract capacity and the fill effectof the diet (Mertens, 1986).Differencesin fill may be due to differences in rate of passage and/or to differences in rate of digestion. Ruminal digestion rates of NDF, acid detergent fiber (ADF), and cellulose have been reported to be higher in sheep fed bmrsorghum-sudangrass hay harvested at prehead stage than those fed normal hay (Wedig et al., 1988).No differences in rate of in situ cell wall digestion between all forage bmr and normal sorghum-Sudan diets were reported in another study (Fritz et al., 1988).
B. In I/iuo DIGESTIBILITY Nutrient utilization in ruminants is influenced greatly by the relationship between intake and digestibility. Most data indicate some depression in apparent digestibility as level of feeding increased, although this effect is somewhat dependent on the forage (D. J. R. Cherney et al., 1990a). If intake of bmr genotypes is greater than that of corresponding normal genotypes, as generally is reported, then differences in digestibility could be partially masked. There have been numerous reports of increased in vivo digestibility of the bmr genotype in maize over its normal counterpart, despite possible masking problems due to increased intake of the brnr genotype. Muller et al. (1972) reported greater apparent in vivo digestibility, in addition to larger intakes, for bm, maize silage fed to lambs. They indicated that decreased
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
187
lignin content in bmr genotypes was primarily responsible for the observed results. Gordon and Neudoerffer ( 1973) also reported higher in vivo digestibility of bmr versus normal maize silage. They attributed higher digestibility not only to lower lignin content, but also to structural differences in the lignin of bmr and normal genotypes. Nuclear magnetic resonance spectra indicated a higher degree of cross-linking of the propane side chain of the lignin molecule in bmr maize than the normal genotype (Gordon and Griffith, 1973).More cross-linking, however, would typically be associated with a lower potential for digestion. Several other investigators have reported increased digestibility of bm, maize silage, among them Sommerfeldt et al. (1979), Block et al. (1982), and Weller and Phipps (1986). Rook et al. (1977) reported no statistically significant differences in apparent digestibilities of any forage component, although digestibilities tended to be higher for all components except protein for bm, silage. Digestibility declinesat about 4% for each increment of intake above maintenance (Tyrrell and Moe, 1975). Correction of digestibility values for this would result in greater differences in digestibility (Rook er af., 1977). In vivo trials with sheep and dairy cows have been conducted with bmr and normal sorghum-sudangrass,both as silage and as hay. Two varieties of bmr sorghum-sudangrass silage fed to growing Holstein steers and heifers resulted in larger extents of digestion than corresponding normal genotypes, regardless of variety (Wedig et af.,1987). Digestibilities of organic matter, crude protein, NDF, ADF, and cellulose did not differ between variety or genotype for steers but were higher for dairy heifers fed bmr “Redlan X Greenleaf” versus heifers fed the normal genotype.Brown-midribsorghumsudangrasshay fed to Holstein cows and sheep also exhibited higher extents of digestion and digestibilitiesfor these componentscompared to the normal variety (Fritz et af.,1988;Wedig et af.,1988). Extent ofdigestion of hemicellulosic sugars and digestibilities of these sugars was higher in bmr sorghumsudangrass than normal genotypes (Wedig et af., 1989a,b). Genotypic differences observed for hemicellulosic sugar digestion may be because these hemicellulosic sugarsare capable of bond formation to phenolic compounds of lignin (Wedig et al., 1989a). Fewer bond formations in the bmrgenotypes would result in more digestible cell wall fractions. Lusk et af.( 1984)reported no differencesin dry matter digestibilitiesbetween bmr and normal sorghum silage fed to lactating cows. There was a trend, however, which indicated greater dry matter digestibilities and intake for the bmr genotype. Dry matter, NDF, and ADF digestibility of two cuttings of bmr pearl millet were consistently higher than that of normal in a wether intake and digestibility study (D. J. R. Cherney et al., 1990b).This wasdespite increased intake of bmr hays in second-cut hays. This is consistent with increased in
188
J. H. CHERNEY et
a[.
vitro dry matter digestibility of bmr pearl millet over normal (Cherney et al., 1988),and can probably be attributed to the lower lignin of bmr pearl millet.
C . PERFORMANCE Differences in intake and digestibility will generally affect animal performance. Intake and digestibility, provided fiber and lignin values are lower in bmr genotypes, have generally indicated superiority of the bmr trait. Animal performance has not always followed as consistently. Several studies have noted small increases in daily milk production. Keith et al. ( 1979)found significant increases in milk production for Holstein cows fed bmr maize silage compared to normal. They did not report any animal weight gains or losses due to treatment. Cows past peak lactation fed bmr maize silage produced 0.5 kg day-' more milk than cows on normal silage (Frenchick et al., 1976). Cows on bmr silage also gained slightly more weight than those fed normal silage (Frenchick et al., 1976). Several other studies, however, showed no significant differencesin actual milk production between bmr and normal maize silage, but weight gains for animals fed bmr silage (Rook et al., 1977;Sommerfeldt et al., 1979;Stallings et al., 1982).Cows on a 12-weekstudy fed bmror normal maize silagedid not differ in intake or milk production (Sommerfeldt et al., 1979). Digestibility was higher for bmr silage, however, and the cows on this diet gained more weight. Block et al. (1981) reported that milk production of cows fed a bmr maize silage with 3 1% soybeans over an 8-week period beginning 18 days postpartum were not different. When analyzed by weeks, however, cows produced more actual milk in weeks 5,6, and 8. Cows fed bmr silage gained weight, while those on the normal diet lost weight. Extra nutrients were apparently used for tissue deposition, rather than milk production in this and other studies (Block et al., 1981;Sommerfeldt et al., 1979). Weller and Phipps (1986) indicated that bmr maize silage fed to growing calvesresulted in greaterweight gains than for calves fed the normal diet, and animals on the bmr diet had better feed efficiency. This was despite the fact that lignin concentrationsof normal maize silage found in the United Kingdom are typically half those found in the United States. Results are in agreement with earlier studies indicating improved animal performance in terms of weight gain and feed efficiency (Colenbranderet al., 1975;Keith et al., 1981). Results are apparentlydue to changesin the stover, rather than the grain. The bm, genotype was observed to have no effect on grain utilization or body weight gain in growing rats when compared to the normal genotype (Keith et al., 1980). Very little data exists comparing bmr sorghum-sudangrass, sorghum, or
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
189
pearl millet to their normal genotypes in terms of animal production. Gain of wethers fed pearl millet hay closely mirrored differences in voluntary intake (D. J. R. Cherney et al., 1990b). Wethers on first-cuttinghays gained similarly,regardlessof genotype. For second-cuttinghays, bmr hays resulted in no wether weight gain, while wethers consuming normal hay lost weight. This was primarily due to low intake of second-cut normal hay, which may have had acceptability problems (D. J. R. Cherney et al., 1990b). More research on animal production related to the bmr trait clearly needs to be initiated. Sommerfeldt et al. ( 1979) observed that concentrations of total volatile fatty acids and butyrate were lower in rumen contents of cows fed bm, maize silage compared to normal. Reduced ruminal pH and acetate to propionate ratios have been reported in cows fed bm, maize silage compared to normal (Block et al., 1981). This lower acetateto propionate ratio resulted in a lower percent milkfat. Cows fed the normal genotype exhibited a high incidence of ketosis. Block et al. (198 1) attributed this possibly to the lower energy intake of the normal maize silage. Rook et al. (1977) reported that total and individual ruminal volatile fatty acids were higher and pH lower with time after feeding for the bm, ration. Wedig et al. (1987) observed no differences in acetate to propionate ratios in rumen fluid from Holstein steers fed bmr/ normal sorghum-sudangrasssilage. Acetate :propionate ratios were lower in sheep fed bmr “Redlan X Greenleaf” sorghum-sudangrasssilage (Wedig et al., 1988). Stallings et al. ( 1982)noted that cows past midlactation that are underfed in energy more likely would respond to bm, silage than those that are overfed slightly. Feeding cows to appetite may preclude obtaining small differences in milk production. They suggested slight underfeeding should be done to detect feeding differences of a particular dietary ingredient. Most studies in which forages are fed as the sole or primary source of feed indicate that the bmr trait will provide increased animal performance, primarily in terms of increased animal gains. This is apparentlydue to increases in intake and digestibility. Increases in milk production have not been consistent. More research on partitioning of energy from bmrgenotypes to milk or gain is indicated.
D. PAIATABILITY Forage palatability and acceptability can influence forage voluntary intake and utilization, thereby having a potentially large impact on forage quality. Forage palatability can be defined as a “plant characteristic(s)elicit-
190
J. H. CHERNEY et ul.
ing a proportional choice among two or more forages conditioned by plant, animal and environmental factors which stimulate a selective intake response by the animal; this characteristic(s)may also be described in terms of acceptability, preference, selective grazing, and relish conditioned by sensory impulse” (Marten, 1970). Acceptability is defined as the readinesswith which an animal will select and consume forage (Hughes ef al., 1962). Several investigators have reported differences in palatability or acceptability between bmrgenotypes and their normal counterparts.Stallings er al. (1982) noted that bmr maize silage was more palatable than normal silage, based on a higher dry matter intake for bmr silage. Higher intakes may have been related to increased digestibility of dry matter and ADF of the bmr genotype, however, and not to differencesin palatability per se. Although no quantitative measures were taken, steers grazing bmrlnormal sorghum-sudangrass regrowth were observed to prefer bmr over normal (V. L. Lechtenberg, personal communication). A replicated palatability study with lambs grazing bmr and normal pearl millet regrowth revealed that sheep spent 2.6 X more time grazing bmr compared to normal (D. J. R. Cherney et al., 1990b). Wethers in metabolism crates took longer to adjust to normal second-cut pearl millet hay and consumed less than wethers on bmr second-cut hays, also indicating possible acceptabilityproblems (D. J. R. Cherney et al., 1990b).There were differencesin lignin and fiber composition ofthese pearl millet hays, but it is unlikely that these differences would account for the wethers’ refusal to consume second-cut normal hays when they were first offered. Following prolonged drought conditions, cattle abruptly ceased grazing normal pearl millet pastures (Roquette ef al., 1980).This pearl millet forage was found to be relatively high in alkaloid content. This also was an unlikely reason for the differences in acceptability observed by D. J. R. Cherney ef al. (1990b), as alkaloids were nondetectable in both bmr and normal hays. Organoleptic factors may be involved. The factor influencing palatabjlity or acceptability is unknown and requires further investigation.
IX. BIOTECHNOLOGY POTENTIAL One of the potentially most effective methods for understanding and modifying lignin synthesis is through biotechnology. Once the molecular and genetic characteristicsof a given plant species are understood, it may be possible to modify processes such as lignin synthesis. At this time, transposon tagging appears to be the most promising method for molecular isolation
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
191
of plant genes associated with a phenotype, but not yet affiliated with a specific gene product (Bennetzen et al., 1987). The gene associated with the simple recessive bmr trait can be isolated using a controlling element system for transposon mutagenesis and tagging. The transposable element system in maize is relatively well understood, and this system can be utilized to isolate the bmr gene(s) of maize. The mutable “candy stripe” system in sorghum also can be used to isolate bmr genes. The candy stripe locus specifies a somatically and germinally unstable synthesis of anthocyanin pigments in the pericarp of sorghum. A second method of isolating the gene involves identification of the protein product from the bmrgene followed by DNA cloning. Identification of a single peptide differing in the mutant compared to normal lines would suggest that the peptide is related to the mutant gene. This relationship, however, must then be proven. Isolating bmr genes and understandingtheir pattern of expression could lead to plants with a modified cell wall more amenable to degradation.
X. SUMMARY AND CONCLUSIONS Plant fiber is a major constituent of forage crops, comprising 30 - 80%of their dry matter. The usefulness of forage for ruminant animals or for bioconversion processes is limited primarily by the degradability of the plant fiber. The most important constraint to digestion of plant cell walls in lignin. Brown-midrib mutants, differing in quality and quantity of lignin from normal genotypes, offer an opportunity to increase the overall digestion of plant fiber. Brown-midrib mutants were first discovered in 1924 in maize, and the trait has since been induced in sorghum and pearl millet. It is a simple recessive trait that phenotypically produces a reddish-brown pigmentation associated with lignified tissues. Several different genes have been identified in both maize and sorghum which produce the characteristic bmr phenotype. Lignin concentrations in bmr genotypes are consistently lower than their normal counterparts. Besides differences between bmr and normal in alkali-labile phenolics, the nitrobenzene oxidation of the two lignins yields different ratios of products. In vitro digestibility of bmr genotypes has been consistentlyhigher than normal, but in vitro rate ofdigestion does not appear to be consistently affected by the mutation. Animal performance on bmr forage has not always produced better results than with normal genotypes, but the tendency is for improved animal performance with bmr genotypes.
192
J. H. CHERNEY et uf.
Investigations of the degradation of tissue types within various parts of normal and bmr plants indicate that the most digestible and the slowly, partially digestible tissues in both sorghum leaves and pearl millet stems are degraded by fiber-digestingbacteria to a greater extent in brnr mutants than in the normal genotypes. Highly lignified support tissues in both types are not digested in either genotype. The precise chemical modificationsresulting in enhanced digestibility require further study and should concentrate on specific areas (i.e., cell wall types) that demonstrate differences in digestibility. Activities of several enzymes involved in lignification differed between bmr and normal genotypes. Differences in activities were not consistent across species, indicating that several different modificationsof the lignification pathway may result in a similar phenotypic bmr response. Geneticiststhroughout the world are now incorporatingthe brnr trait into a range of backgrounds in sorghum and pearl millet, and there is renewed interest in the maize bmr trait in Europe. Studiesare now underway to isolate the bmr gene and characterize it. A better understanding of this modified lignification process should lead to methods for transferringthis trait to other species. Brown-midrib, low-lignin mutants of herbaceous grass species provide an excellent system for examining and possibly modifyingthe lignification process in plants.
REFERENCES Akin, D. E., Hanna, W. W., and Rigsby, L. L. 1986a. Normal-12 and brown midrib12 sorghum. I. Variations in tissue digestibility.Agron. J. 78,827-832. Akin, D. E., Hanna, W. W., Snook, M. E., Himmelsbach,D. S., Barton, F. E., 11, and Windham, W. R. 1986b. Normal-I2 and brown midrib12 sorghum. 11. Chemical variations and digestibility. Agron. J. 78,832-837. Akin, D. E., Rigsby, L. L., Theodorou, M. K., and Hartley, R. D. 1988. Population changes of fibrolytic rumen bacteria in the presence of phenolic acids and plant extracts. Anim. Feed Sci. Technol. 19,26 1- 275. Akin, D. E., Rigsby, L. L., Hanna, W. W., Gates, R. N. 1991. Structure and digestibility of tissues in normal and brown midrib pearl millet. J. Sci. Food Agric. (in press). Auerbach, C. 1976. Chemical mutagens: alkylating agents. I. Genetical effects. I n “Mutation Research Problems, Results and Perspectives,” pp. 256-298. Chapman &Hall,London. Barnes, R. F., Muller, L. D., Bauman, L. F., and Colenbrander, V. F. 1971. In vitro dry matter disappearance of brown midrib mutants of maize (Zea mays L.). J. Anim. Sci. 33,88 1 884. Bennetzen, J. L., Cresse, A., Brown, W. E., and Lee, L. 1987. Molecular cloning of maize genes by transposon tagging with mufafor.In “Plant Gene Systemsand Their Biology” (J. L. Key and L. Mclntosh, eds.),pp. 183-204. Allan R. Liss, New York.
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
193
Bittinger, T. S., Cantrell, R. P., and Axtell, J. D. 1981. Allelism tests of the brown midrib mutants of sorghum. J. Hered. 72, 147- 148. Blaxter, K. L., Wainman, F. W., and Wilson, R. S. 1961. The regulation of food intake by sheep. Anim. Prod. 3, 5 1 - 6 1. Block, E., Muller, L. D., Griel, L. C., Jr., and Ganvood,D. L. 1981. Brown midrib3 corn silage and heat extruded soybeans for early lactating cows. J. Dairy Sci. 64, 1813- 1825. Block, E., Muller, L. D., and Kilmer, L. H. 1982. Brown midrib3 versus normal corn plants (Zea mays L.) harvested as whole plant or stover and frozen fresh or preserved as silage for sheep. Can. J. Anim. Sci. 62,487-498. Bucholtz, D. L., Cantrell, R. P., Axtell, J. D., and Lechtenberg, V. L. 1980.Lignin biochemistry of normal and brown midrib mutant sorghum. 1.Agric. Food Chem. 28,1239- 1241. Bula, R. J., Lechtenberg, V. L., and Holt, D. A. 1981. Potential of temperate zone cultivated forages for ruminant animal production. In “Potential of the World’s Forages for Ruminant Animal Production’’(R. D. Child and E. K. Byington, eds.),Winrock Rep., 2nd Ed., pp. 7-28. Winrock Int., Morrilton, Arkansas. Bumtt, E. A., Bittner, A. S., Street, J. C., and Anderson, M. J. 1984. Correlations of phenolic acids and xylose content of cell wall with in vitro dry matter digestibilityof three maturing grasses. J. Dairy Sci. 67, 1209- 1213. Camm, E. L., and Towers, G. H. N. 1973. Review article: Phenylalanine ammonium lyase. Phytochemistry 12,961 -973. Cherney, D. J. R., Mertens, D. R., and Moore, J. E. 1990a. Intake and digestibility by sheep as influenced by forage morphology at three levels of offer. J.Anim. Sci. 68,4387-4399. Cherney, D. J. R., Patterson, J. A., and Johnson, K. D. 1990b.Digestibilityand feedingvalue of pearl millet as influenced by the brown-midrib, low-lignin trait. J. Anim. Sci.68,4345 435 1. Cherney, D. J. R., Patterson, J. A., Cherney, J. H., and Axtell, J. D. 1991. Fibre and soluble phenolic monomer composition of morphologicalcomponents of sorghum stover. J. Sci. Food Agrk 54, (in press). Cherney, J. H. 1990. Normal and brown midrib mutations in relation to improved lignocellulose utilization. In “Microbial and Plant Opportunities to Improve LignocelluloseUtilization by Ruminants” (D. E. Akin and L. G . Lungdahl, eds.). pp. 205-214. Elsevier, Amsterdam. Cherney, J. H., Volenec, J. J., and Moore, K. J. 1985a. Cell wall composition and rate of digestion of brown midrib sorghum internodes as influenced by maturity. Proc. Int. Grassl. Congr. Nishi-Nasuno, lSth, Tochigiken, Jpn. pp. 953-954. Cherney, J. H., Volenec, J. J., and Nyquist, W. E. 1985b. Sequential fiber analysis of forage as influenced by sample weight. Crop Sci. 25, 1 1 13- 1 115. Cherney, J. H., Moore, K. J., Volenec, J. J., and Axtell, J. D. 1986. Rate and extent of digestion of cell wall components of brownmidrib sorghum species. Crop Sci. 26, 1055- 1059, Cherney, J. H., Axtell, J. D., Hassen, M. M., and Anliker, K. S. 1988. Forage quality characterization of a chemically induced brownmidrib mutant in pearl millet. Crop Sci. 28,783787. Cherney, J. H., Anliker, K. S., Albrecht, K. A., and Wood, K. V. 1989a. Soluble phenolic monomers in forage crops. J.Agric. Food Chem. 37, 345 - 350. Cherney, J. H., Volenec, J. J., and Brown, G. A. 1989b. Synthesis of cell wall components in maize internodes. Proc. Int. Grassl. Congr., 16th, Versaifles,Fr. pp. 759-760. Cherney, J. H., Cherney, D. J. R., Sollenberger,L. E., Patterson, J. A., and Wood, K. V. 1990. Identification of a quinic acid ester in limpograss and its influence on fiber digestion. J. Agric. Food Chem. 38,2 140- 2 143.
J. H. CHERNEY et ul.
194
Cherney, J. H., Johnson, K. D., Volenec, J. J., and Greene, D. K. 1991. Biomass potential of selected grass and legume crops. Energy Sources 13,283 - 292. Colenbrander, V. F., Bauman, L. F., and Lechtenberg, V. L. 1973. E5ect of brown midrib mutant genes on the nutritional quality of corn. Proc.Annu. Corn Sorghum Res. Conf:28, 92-97.
Colenbrander, V. F., Bauman, L. F., and Lechtenberg, V. L. 1975. Feeding value of low lignin corn silage. J. Anim. Sci.41,332-333. Conrad, H. R., Pratt, A. D., and Hibbs, J. W. 1964. Regulation of feed intake in dairy cows. I. Change in importance of physical and physiologicalfactors with increasing digestibility.J. Dairy Sci. 47, 54-62. Cymbaluk, N. F., Gordon, A. J., and Nuedoerffer, T. S. 1973. The effect of the chemical composition of maize plant lignin on the digestibilityof maize stalk in the rumen of cattle. Br. J. Nutr. 29, 1-12. Davies, M. E. 1972. Polyphenol synthesis in cell suspension cultures of Paul‘s Scarlet Rose. Planta 104,50-65. De Jaegher, G. N., Boyer, N., and Gaspar, T. 1985. Thigmorphogenesisin Bryonia dioica: changes in solubleand wall peroxidases, phenylalanineammonia-lyaseactivity, cellulose, lignin content and monomeric constituents. Plant Growth Regul. 3, 133- 148. El-Tekriti, R. A., Lechtenberg, V. L., Bauman, L. F.,and Colenbrander, V. F. 1976. Structural composition and in vitro dry matter disappearanceof brown midrib corn residue. Crop Sci. 16, 387-389.
Fisher, D. S., Bums, J. C., and Pond, K. R. 1989. Kineticsof in vitrocell-walldisappearanceand in vivo digestion. Agron. J. 81,25 - 33. Forbes, J. M. 1977. Interrelationships between physical and metabolic control of voluntary food intake in fattening, pregnant and lactating mature sheep: a model. Anim. Prod. 24, 91- 101.
Frenchick, G. E., Johnson, D. G., Murphy, J. M., and Otterby, D. E. 1976. Brown midrib corn silage in dauy cattle rations. J. Dairy Sci. 59,2 126 - 2 129. Fritz, J. O., Cantrell, R. P., Lechtenberg, V. L., Axtell, J. D., and Hertel, J. M. 1981. Brown midrib mutants in sudangrass and grain sorghum. Crop Sci. 21,706-709. Fritz, J. O., Moore, K. J., and Jaster, E. H.1988. In situ digestion kineticsand ruminal turnover rates of normal and brown midrib mutant sorghum x sudangrass hays fed to nonlactating Holstein cows. J. Dairy Sci. 71, 3345 - 335 1. Fritz, J. O., Moore, K. J., and Jaster, E. H. 1990. Digestion kinetics and cell wall composition of brown midrib sorghum x sudangrassmorphological components. Crop Sci. 30,2 13 219.
Gaillard, B. D. E., and Richards, G. N. 1975. Presence of soluble lignin-carbohydrate complexes in the bovine rumen. Carbohydr. Res. 42, 135- 145. Gallais, A., Huguet, L., Berthet, H., Berth, G., Broqua, B., Mourguet, A., and Traineau, R. 1980. Preliminary evaluation of brown midrib maize hybrids for their feeding and agronomic value in France. In “Improvement of Quality Traits of Maize for Grain and Silage Use” (W.G. Pollmer and R. H. Phipps, eds.),pp. 319-339. Nijhoff, The Hague. Gaudillere, M., and Monties, B. 1989. Biochemical and biosynthetic studies on lignification of Gramineae. In “Plant Cell Wall Polymers :Biogenesis and Biodegradation” (N. G. Lewis and M. G. Paice, eds.), ACS Symp.Ser., 399, 182- 192. American Chemical Society, Washington, D.C. Gee, M. S., Nelson, 0. E., and Kuc‘, J. 1968. Abnormal lignins produced by the brown-midrib mutants of maize. 11. Comparative studies on normal and brown-midrib1 dimethylformamide lignins. Arch. Biochem. Biophys. 123,403-408. Goering, H. K., and Van k s t , P. J. 1970. “Forage Fiber Analysis; Apparatus, Reagents,
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
195
Procedures, and Some Applications,” USDA Agric. Handb. No. 379. U.S. Gov. Print. Off., Washington, D.C. Gordon, A. J., and Griffith, P. R. 1973. Chemical and in vivo evaluation of a brown midrib mutant of Zea mays. 11. Nuclear magnetic resonance spectra of digested and undigested alkali lignins and undigested dimethyformamide lignins. J. Sci. FoodAgric. 24,579-587. Gordon, A. J., and NeudoerfFer, T. S. 1973.Chemical and in vivoevaluationofa brown midrib mutant of Zea mays. I. Fibre, lignin and amino acid composition and digestibility for sheep. J. Sci. Food Agric. 24,565-577. Grand, C. P., Parmentier, P., Boudet, A., and Boudet, A. M. 1985.Comparison of ligains and of enzymes involved in lignificationin normal and brown midrib (bm,) mutant corn seedlings. Physiol. Veg. 23,905 -9 1 1. Grenat, E., and Barry, P. 1991. Microbial degradation of normal maize and bm, make in the rumen observed by scanning electron microscopy. J. Sci. Food Agric. 54, 199-2 10. Grim, T. C., Cherney, J. H., Cherney, D. J. R., and Patterson, J. A. 1989. Phenolic-carbohydrate complexes in maize cell walls and influences on rumen microbes. Agron. Absfr. p. 169. Gross, G. G. 1985. Biosynthesis and metabolism of phenolic acids and monolignols. In “Biosynthesis and Biodegratation of Wood Components” (T. Higuchi, ed.), pp. 229-271. Academic Press, Orlando, Florida. Gross, G. G., and Zenk, M. H. 1974. Isolation and properties of hydroxycinnamate:CoA ligase. Eur. J. Biochem. 42,453-459. Hagerman, A. E. 1987. Phenolic biosynthesis: pathways and regulation. In “Models in Plant Physiology and Biochemistry,” (D. W. Newman and K. G. Wilson, eds.), Vol. 2, pp. 69 - 73. CRC Press, Boca Raton, Florida. Hahlbrock, K., and Grisebach, H. 1979. Enzymic control in the biosynthesis of lignins and flavonoids. Annu. Rev. Plant Physiol. 30, 105- 130. Hanna, W. W., Monson, W. G., and Gaines, T. P. 1981. IVDMD, total sugars, and lignin measurements on normal and brown midrib (bmr) sorghums at various stages of develop ment. Agron. J. 73, 1050- 1052. Hartley, R. D. 1972.pCoumaric and ferulic acid components ofcell walls of ryegrass and their relationships with lignin and digestibility. J. Sci. Food Agric. 23, 1347- 1354. Hartley, R. D. 1985. Chemistryof lignocellulosicplant materials in non-microbial processesfor increasing their food value for the ruminant. In “Improved Utilization of Lignocellulosic Materials in Animal Feed,” pp. 10-30. OECD, Paris. Hartley, R. D., and Jones, E. C. 1978. Phenoliccomponentsand degradability ofthe cell walls of the brown midrib mutant, bm,, of Zea mays. J. Sci. Food Agric. 29,777-782. Hughes, H. D., Heath, M. E., andMetcalfe, D. S., eds. 1962. “Forages: The ScienceofGrassland Agriculture.” Iowa State Univ. Press, Ames. Hutton, E. M. 1971. Plant improvement for increased animal production. J.Aust. Inst. Agric. Sci. 37,212-225. Jorgenson, L. R. 1931. Brown midrib in maize and its linkage relations. J. Am. SOC.Agron. 23, 549-557. Jung, H. G . 1988. Inhibitory potential of phenolic-carbohydrate complexes released during ruminal fermentation. J. Agric. Food Chem. 36,782-788. Jung, H. G., and Fahey, G. C., Jr. 1983. Nutritional implications of phenolic monomers and lignin: a review. J. Anim. Sci. 57,206-219. Jung, H. G., Fahey, G .C., Jr., and Garst, J. E. 1983a.Simplephenolic monomers of foragesand effects of in vitro fermentations on cell wall phenolics. J. Anim. Sci. 57, 1294- 1305. Jung, H. G., Fahey, G. C., Jr., and Merchen, M. R. 1983b. Effects of ruminant digestion and metabolism on phenolic monomers of forages. Br. J. Nutr. 50,637-65 1.
196
J. H. CHERNEY et d
Kalton, R. R. 1988.Overviewofthe forage sorghums.Proc. Annu. Corn SorghumRes. Cont43, 1-12. Keith, E. A., Colenbrander, V. F., Lechtenberg, V. L., and Bauman, L. F. 1979. Nutritional value of brown midrib corn silage for lactating dairy cows. J. Dairy Sci. 62,788 - 792. Keith, E. A., Colenbrander, V. F., Bauman, L. F., and Lechtenberg, V. L. 1980. Effect of the brown midribthree gene on corn grain utilization by growing rats. J. Anim. Sci. 51, 892- 895. Keith, E. A.,Colenbrander, V. F., Perry, T. W., andBauman, L. F. 1981. Performanceoffeedlot cattle fed brown midribthree or normal corn silage with various levels of additional corn grain. J, Anim. Sci. 52,8- 13. Kuc’, J., and Nelson, 0.E. 1964. The abnormal lignins produced by the brown midrib mutants of maize. I. The brown-midrib1 mutant. Arch. Biochem. Biophys. 105, 103- 1 13. Kuc‘, J., Nelson, 0. E., and Flanagan, P. 1968. Degradation of abnormal lignins in the brown midrib mutants and double mutants of maize. Phytochemistry 7 , 1435- 1436. Lapierre, C., Tollier, M. and Monties, B. 1988. Occurrence of additional monomeric units in the lignins from internodes of a brown-midrib mutant of Maize bm, .C. R. Acad. Sci. Paris 307 (series III), 723-728. Lawton, M. A., Dixon, R. A,, and Lamb, C. J. 1980. Elicitor modulation of the turnover of L-phenylalanine ammonia-lyase in French bean cell suspension cultures. Biochim. Biophys. Acta 633, 162- 175. Lechtenberg, V. L., Muller, L. D., Bauman, L. D., Rhykerd, C. L., and Barnes, R. F. 1972. Laboratory and in vitro evaluation of inbred and F2populations of brown midrib mutants of Zea mays. Agron. J. 64,657-660. Lechtenberg, V. L., Colenbrander, V. F., Bauman, L. F., and Rhykerd, C. L. 1974. Effect of lignin on rate of in vitro cell wall and cellulose disappearance in corn. J. Anim. Sci. 39, 1165- 1169. Lee, M. H., and Brewbaker, J. L. 1984. Effects of brown midrib3 on yields and yield components of maize. Crop Sci. 24, 105- 108. Lusk, J. W., Karau, P. K., Balogu, D. O., and Gourley, L. M. 1984. Brown midrib sorghum or corn silage for milk production. J. Dairy Sci. 67, 1739 - 1744. Mader, M., and Fussl, R. 1982. Role of peroxidax in Lignification of tobacco cells. 11. Regulation by phenolic compounds. Plant Physiof.70,1132 - 1 134. Malmstrom, B. G. 1982. Enzymology of oxygen. Annu. Rev.Biochem. 51,21-59. Marten, G. C. 1970. Measurement and significance of forage palatability. Proc. Natl. ConJ Forage Qual. Evuf. Util., Lincoln, Nebr. pp. D1-053. Marten, G. C., and Barnes, R. F. 1980. Prediction of energy digestibilityof forageswith in vitro rumen fermentation and fungal enzyme systems.In “Standardization of Analytical Methodology for Feeds” (W. J. Pigden, C. C. Balch, and M. Graham, eds.), pp. 6 1 -7 1. Int. Dev. Res. Cent., Ottawa. Mertens, D. R. 1986. Importance and characteristicsof functional fiber in dairy rations. Proc. Cali$ Anim. Nutr. Con5 pp. 38-55. Miku, V. E. 1972. Revealing and genetic analysis of brown-midrib maize mutants. Genetica 8, 157-158. Milford, R., andMinson, D. J. 1965.Intake oftropical pasture species. Proc.Int. Grassl. Congr., 9th, Sao Paulo, Braz. pp. 8 15 - 822. Miller, J. E., Geadelmann, J. L., and Marten, G. C. 1983. Effect of the brown midriballele on maize silage quality and yield. Crop Sci. 23,493-496. Minson, D. J. 1980. Relationshipsofconventional and preferred fractionsto determined energy values. In “Standardization of Analytical Methodology for Feeds” (W. J. Pigden, C. C. Balch, and M. Graham, eds.), pp. 72-78. Int. Dev. Res. Cent., Ottawa.
BROWN-MIDRIB, LOW-LIGNIN MUTANTS
197
Moore, J. E. 198 I. Principlesof forage quality evaluation. In “King V ng Scholar Lectures.” Spec. Rep.-Arkansas Agric. Exp. Stn. No. 93, pp. 66-87. Moore, J. E., and Mott, G. 0. 1973. Structural inhibitors of quality in tropical grasses. In “Anti-Quality ComponentsofForages”(A. G. Matches, ed.), pp. 53-98. CSSA Spec. Publ. No. 4. Crop Sci. Soc. Am., Madison, Wisconsin. Moore, K. J., and Cherney, J. H. 1986. Digestion kinetics of sequentially extracted cell wall components of forages. Crop Sci. 26, 1230- 1235. Momson, I. M. 1974. Structural investigationson the lignin-carbohydratecomplexes of Lolium perenne. Biorhem. J. 139, 197-204. Mueller-Harvey, I., Hartley, R. D., Harris, P. J., and Curzon, E. H. 1986. Linkage ofpcoumaroyl and feruloyl groups to cell wall polysaccharides of barley straw. Carbohydr. Res. 148, 71-85. Muller, L. D., Barnes, R. F., Bauman, L. F., and Colenbrander, V. F. 197 I. Variations in lignin and other structural components of brown midrib mutants of maize (Zea mays L.). Crop Sci. 11,413-415. Muller, L. D., Lechtenberg, V. L., Bauman, L. F., Barnes, R. F., and Rhykerd, C. L. 1972. In vivo evaluation of a brown midrib mutant of Zea mays L. J. Anim. Sci. 35,883-889. Neilsen, M. J., and Richards, G. N. 1978. The fate of the soluble lignin-carbohydratecomplex produced in the bovine rumen. J. Sci. Food Agric. 29,5 13- 5 19. Nicholson, R. L., Bauman, L. F., and Warren, H. L. 1976. Association of Fusarium moniliforme with brown midrib maize. Plant Dis. Rep. 60,908-910. Oh, H. K., Baumgardt, B. R., and Scholl, J. M. 1966. Evaluation of foragesin the laboratory. V. Comparison of chemicalanalyses, solubilitytest and in vitro fermentation. J. Dairy Sci. 48, 850-855. Porter, K. S., Axtell, J. D., Lechtenberg, V. L., andcolenbrander, V. F. 1975. Thebrown midrib gene (bmr) in sorghum and its affect on forage quality. Proc. Int. Sorghum Workshop, Mayaguez, P. R. pp. 47 1-482. Porter, K. S., Axtell, J. D., Lechtenberg, V. L., and Colenbrander, V. F. 1978. Phenotype, fiber composition, and in vitro dry matter disappearance of chemically induced brown midrib (bmr) mutants of sorghum. Crop Sci. 18,205 - 208. Raymond, W. F. 1969. The nutritive value of forage crops. Adv. Agron. 21, 1- 108. Rook, J. A., Muller, L. D., and Shank, D. B. 1977. Intake anddigestibility ofbrown-midribcorn silage by lactating dairy cows. J. Dairy Sci. 60, 1894- 1904. Roquette, F. M., Jr., Keisling, T. C., Camp, B. J., and Smith, K. L. 1980. Characteristicsof the Occurrence and some factors associated with reduced palatability of pearl millet. Agron. J. 12, 173- 174. Sawai, A., Kondo, T., and Ara, S. 1983. Inhibitory effects of phenolic acid esters on degra&bility of forage fibers. J. Jpn. Grassl. Sci. 29, 175 - 179. Schaflert, R. E., and Gourley, L. M. 1982. Sorghum as an energy source. Sorghum Eighties: Proc. Int. Symp. Sorghum, India (L. R. House, L. K. Mughogho, and J. M. Peacock, eds.), pp. 605-623. Pantancheru, A. P., India. Sheldrick, R. D. 1979. The quality of “brown midrib-3” mutant maize grown for forage under field conditions in southern England. Grass Forage Sci. 34,283 - 29 I. Sommerfeldt, J. L., Schingoethe, D. J., and Muller, L. D. 1979. Brown-midrib corn silage for lactating dairy cows. J. Dairy Sci. 62, 16 1 1 - 16 18. Stallings, C. C., Donaldson, B. M., Thomas, J. W., and Rossman, E. C. 1977. Brown-midrib corn silage as a substitute for regular corn silage. J. Anim. Sci. 45, Suppl. 1, 126. Stalling, C. C., Donaldson, B. M., Thomas, J. W., and Rossman, E. C. I982. In vivoevaiuation of brown-midrib corn sdage by sheep and lactating dany cows. J. Dairy Sci. 65, 19451949.
198
J. H. CHERNEY et al.
Theodorou, M. K., Gascope, D. J., Akin, D. E., and Hartley, R. D. 1987. Effect of phenolic acids and phenolics from plant cell walls on rumenlike fermentation in consecutivebatch culture. Appl. Environ. Microbiol. 53, 1046- 1050. Tyrell, J. F., and Moe, P. W. 1975. Effect of intake on digestive efficiency. J. Dairy Sci.58, 1151- 1163. Wedig, C. L., Jaster, E. H., Moore, K. J., and Merchen, N. R. 1987. Rumen turnover and digestion of normal and brown midrib sorghum x sudangrasshybridsdages in dairy cattle. J. Dairy Sci. 70, 1220- 1227. Wedig, C. L., Jaster, E. H., and Moore, K. J. 1988. Effect of brown midrib sorghum on ruminal fluid and particulate rate of passage and extent of digestion at various sites along the gastrointestinaltract in sheep. J. Anim. Sci. 66,559-565. Wedig, C . L., Jaster, E. H., and Moore, K. J. 1989a.Disappearanceof hemicellulosic monosaccharides and alkali-soluble phenolic compounds of normal and brown midrib sorghum x sudangrasses fed to heifers and sheep. J. Dairy Sci. 72, 104- 1 1 1. Wedig, C. L., Jaster, E. H., and Moore, K. J. 1989b. Disappearanceof hemicellulosic monosaccharides and alkali-soluble phenolic compounds of normal and brown midrib sorghum x sudangrass silages fed to Holstein steers. J. Dairy Sci. 72, 112- 122. Weller, R. F., and Phipps, R. H. 1986. The feeding value of normal and brown midrib3 maize silage. J. Agric. Sci. 106, 3 1- 35. Weller, R. F., Phipps, R. H., and Griffith, E. S. 1984. The nutritive value of normal and brown midrib3 maize. J. Agric. Sci. 103,223-227. Weller, R.F., Phipps, R. H., and Cooper, A. 1985. The effects ofthe brown midrib3 gene on the maturity and yield of forage maize. Grass Forage Sci. 40,335 - 339. Yu,Y., Thomas, J. W., and Muller, L. D. 1972. Ultraviolet characteristics of brown midrib corns. J. Anim. Sci. 35, 1 1 15. Zuber, M. S., Colbert, T. R., and Bauman, L. F. 1977. Effect of brown-midrib3 mutant in maize (Zea mays L.) on stalk strength. Z. Pflanzenzucht. 79,310-314.
M~ASUREMENT OF SURFACE CHARGE OF INORGANICGEOLOGIC &TERIALS: BCHNXQUES AND ~ E I CONSEQUENCES R Anne Lewis-Russ Department of Chemistry and Geochemistry Colorado School of Mines Golden. Colorado 80401
I. Introduction 11. Surface Charge Terminology 111. Surface Charge Development A. Permanent Charge B. Variable Charge C. Specifically Adsorbed and Indifferent Ions D. Problems with Ion Classifications IV. Measurement of Surface Charge A. Potentiometric Titration B. Electrokinetic Techniques C. Ion Retention D. Salt Titration E. Mineral Addition V. Separating Permanent and Variable Charge VI. Measurement Problems A. Altering the Solid Phase B. Interactions with the Electrolyte Solution VII. Applications and Predictions for Composite Materials VIII. Summary References
I. INTRODUCTION The importance of surface charge in controllingthe behavior of colloids is well known. An understanding of surface charge is basic to explaining adsorption in both natural and synthetic systems. Coagulation (Healy et d., 1966,1968; Parks, 1967;Breeuwsma and Lyklema, 1973; Hohl et al., 1980), 199 Adwnrr in Apmnny, Vol.46 Copyright 8 1991 by Academic Press, Inc. AU rights of repduction in any form I-CSCN~.
200
ANNE LEWIS-RUSS
heat of immersion (Healy and Fuerstenau, 1965),surface tension (Parks and de Bruyn, 1962),electrophoreticmobility (Healy et al., 1968;Murray, 1974; Houchin and Warren, 1984; Hansmann and Anderson, 1985; Bryant and Williams, 1987; Carroll-Webb and Walther, 1988; Schulthess and Sparks, 1988), solubility (Parks and de Bruyn, 1962; White and Zelazny, 1986; Carroll-Webb and Walther, 1988),and cation and anion exchange capacities (Schofield, 1949; van Raij and Peech, 1972; Gillman and Uehara, 1980)all depend on surface charge and the variation of surface charge with solution properties. Many techniques have been developed to measure surface charge, including potentiometric titration, electrophoretic mobility, ion retention, salt titration, and mineral addition. For pure oxides, applications of these techniques are well documented (Healy et al., 1966; Breeuwsma and Lyklema, 1971, 1973; Murray, 1974; James et al., 1981; Regazzoni et al., 1983). Applicationsto mixed systems have been more limited. Recent publications have described innovative applicationsof surface charge techniques to complex oxides (Kuo, 1987; Kuo and Yen, 1986, 1988; Carroll-Webb and Walther, 1988;Kuo et al., 1988).Schulthessand Sparks (1986, 1987, 1988, 1989) have critically examined measurement techniques and their unsuspected effects on the study materials. During recent attempts by the author to determine the surface charge of crushed volcanic rock, it became apparent that the measurement techniques were affecting the surface charge. Although reports of similar effects are found in the literature, there has been no comprehensive review of these problems. The purpose of this chapter is to examine some of the past work on surface charge measurement,with specific attention to the applicationof the measurement techniques to complex geologic materials. The attendant problems of these applicationsare investigated. The most widely used technique, potentiometric titration, is examined in the greatest detail. This chapter is a partial survey of surface charge literature.It is intended to provide a critical review of surface charge measurement for practitioners working with complex, reactive geologic materials, such as soil or rock. This review is also relevant for scientists using surface ionization and complexation adsorption models, because adsorption constants used in these models are derived, in part, from surface charge measurements. In the first section, surface charge terms commonly found in the literature are defined, and the development of surface charge is discussed. Next, various measurement techniques are reviewed and problems in applying these techniques are examined. In the third section, applicationsto mixed systems are presented. Organic matter and its effect on surface charge is not considered in this chapter.
MEASUREMENT O F SURFACE CHARGE
20 1
11. SURFACE CHARGE TERMINOLOGY The following terms are briefly defined here. More detailed explanations are provided in the text or in the listed references. Terms are listed in conceptual rather than alphabetical order. Surface charge: electrical charge on the surfaces of colloidsresulting from isomorphous substitutions in the crystallographic structure, from ion adsorption, or from both. Net proton charge (qJ:surface excess density of protons (H+)compared to hydroxide (OH-). Inner-sphere complexes: ion pairs that have direct, short-range bonding between the ion and the particle, and no interposed water molecules (Langmuir, 1979). net surface charge of ions (other Inner-sphere complex charge (qJ: than H+ or OH-) that have formed inner-sphere complexes with the surface (Sposito, 1981). Outer-sphere complexes: ion pairs that have at least one hydration sphere between the ion and the particle. Bonding generally is Coulombic and long range (Langmuir, 1979). Outer-sphere complex charge (cT-): net surface charge of ions (including H+ or OH-) that have formed outer-sphere complexes with the surface (Sposito, 1981). Permanent charge (OJ: surface charge inherent in a mineral, resulting from isomorphous substitutions in the crystallographicstructure (White and Zelazny, 1986). Variable charge: surface charge that varies depending on solution properties, such as ionic strength and pH (White and Zelazny, 1986). Potential-determining ions (PDI): ions that are common to both the mineral and the solution,and determine charge by adsorption and desorption (Arnold, 1978). Zero point of charge (ZPC): the suspension pH at which the net surface charge of a colloid is zero (Parks, 1967).The ZPC is dependent on surface properties and solution composition. Zeta potential (0:the average electric potential developed in the slipping plane when liquid and solid phases move with respect to each other (Hunter, 1981). Zero zeta potential: the suspension pH at which the net potential in the slipping plane is zero. Isoelectricpoint (IEP): ( 1) the zero zeta potential measured using an
202
ANNE LEWIS-RUSS
electrokinetictechnique;or (2) the ZPC resulting from the interaction of a colloid solely with H+ and OH- (Parks, 1967). Unless otherwise noted, the first definition is used in this chapter. Pristine point of zero charge (PPZC): another term used for the second definition of IEP. The PPZC is a unique property of the solid that is independent of the electrolyte (Bowden et al., 1980b). Point of zero net proton charge (PZNPC): the pH at which the net proton charge is zero (Sposito, 1981; Charlet and Sposito, 1987).The PZNPC is the same as the PPZC when charge is developed due solely to H+ and OH-. Potentiometric titration: a titration technique in which the pH of a suspension is modified in small steps by adding dilute acid or base. At each step, the pH is measured to determine the quantity of H+ and OH- remaining in solution. These amounts are subtracted from the total H+ or OH- added and the remainder is assumed to be adsorbed on the solid particles. Zero point of titration (ZPT): the equilibrium pH attained between a solid and solution before acid or base additions (Parker et al., 1979). Point of zero salt effect (PZSE): the pH at which two or more potentiometrictitration curves intersect (Parker et al., 1979; Sposito, 1981). This is also referred to as the common intersection point (CIP), or the crossover. Common intersection point (CIP): the pH at which two or more potentiometrictitration curves intersect (also referred to as the crossover). This is identical to the PZSE, defined above, and the same as the ZPC when there is no specific adsorption (Pyman et al., 1979a; Lyklema, 1984). Cation exchange capacity (CEC): the quantity of cations (except H+) in solution (expressed as milliequivalents (meq) per 100 grams dry material)that can be exchanged by a material for adsorbed cations (Robinson, 1962; van Olphen, 1963; Uehara and GiIlman, 1980). Because cation exchange capacity measurement techniques do not account for proton adsorption, CEC values vary as the measurement pH varies (C. P. Schulthess, 1990 personal communication). Anion exchange capacity (AEC): the positive charge of soils as measured by chloride (Cl-) or nitrate (NO,) retention. This value is small compared to CEC in the pH range typically encountered in nature (Parfitt, 1980). Because the anion exchange measurement techniques do not account for hydroxyl adsorption, AEC values vary as the measurement pH varies (C. P. Schulthess, 1990 personal communication).
MEASUREMENT OF SURFACE CHARGE
203
Point of zero net charge (PZNC): the pH at which the algebraic sum of the CEC and the AEC is zero (Parker et al., 1979). This pH is determined by measuring cation and anion retention (the CEC and AEC) at a range of pH values. Flotation: a separation process in which ground minerals are made hydrophobic so they attach to air bubbles and separate from the rest of the slurry as froth (Fuerstenau, 1970). Coagulation: aggregation of colloids in a suspension; occurs when van der Waals attraction is greater than electrostatic repulsion (van Olphen, 1963). Sedimentation: settling of colloidal particles in a suspension due to coagulation. Several of the terms defined above (ZPC, IEP, PPZC, PZNPC, ZPT, PZSE, and PZNC) have incorrectly been used interchangeably. Some of the confusion results because these terms derive in part from the techniquesused to measure the surface charge and in part from the theoretical models of the particle surface (Hohl et al., 1980).These models are mathematical descriptions of the boundary between the colloid with its adsorbed ions and the solution. In the simplest model, the double-layer model, all specifically or Coulombically adsorbed ions are in the inner (Stern)layer and all repelled ions are in the outer diffuse layer. There also are triple-layer models that allow H+ and OH- to be in a layer closer to the particle than other adsorbed ions (Davis et al., 1978;James, 1981;James and Parks, 1982),and a four-layer model that has an additional layer for the specifically adsorbed ions (Bowden et al., 1980a). (For detailed comparisons of models, see Westall and Hohl, 1980; Morel et al., 1981; Barrow, 1985.) Sposito ( 1981) clarified the interrelationship of various ZPC terms by using a mass balance approach that did not refer to models of surface chemistry. His “balance of surface charge” is based on the overall neutrality of an aqueous solution, in which the sum of charges associated with any particle is equal to zero. In addition to the previously defined permanent charge (op), net proton charge (oh), inner-sphere complex charge (oJ, and outer-sphere complex charge (om),Sposito (1 98 I ) added dissociated charge (odd). The balance of surface charge is:
The dissociated charge (od) is the neutralizingcharge of the unassociated ions
ANNE LEWIS-RUSS
204
Table I Some Definitions of Points of Zero Charge" Symbol
Name
ZPC PPZC
Zero point of charge Pristine point of zero charge
PZNPC PZSE PZNC
Point of zero net proton charge Point of zero salt effect Point of zero net charge
Conditionb a, = 0 a h =0 and up= a, = a , =0 Uh = 0 (du,,/dZ), = 0 + a , a, = 0
Used with permission, from Sposito (1981), Table 1. dZ = change in ionic strength; T = absolute temperature; other symbols defined in text.
of the diffuse layer. The more common definitions of points of zero charge can be understood in terms of the above definitions, as illustrated in Table I.
111. SURFACE CHARGE DEVELOPMENT The surface charge of a mineral can result from structural factors inherent in the mineral or from adsorption of potential-determining ions (PDI). When inherent structural factors cause development of surface charge, the surface charge is constant (permanent), and increases in solution ionic strength cause decreases in the surface potential (van Olphen, 1963). These decreases result from the increased positive and negative charge density in solution and in the diffise layer. A smaller volume of high ionic strength solution satisfies the permanent charge deficit because this smaller volume has a higher concentration of both counterions (ions having the opposite charge as the surface) and co-ions (ions having the same charge as the surface). The small solution volume is more charge-balanced very close to the surface than the same volume of low ionic strength solution, because there is proportionately less co-ion repulsion by the surface, resulting in a lower surface potential. This is illustrated schematically by van Olphen (1963). When surface charge results from adsorption of PDI, the surface potential is constant as long as the activities of the PDI in solution are constant. Increased ionic strength results in increased (variable) surface charge (van Olphen, 1963). The increased surface charge results from the increased charge density in solution and in the diffise layer. Solution chemistry is the
MEASUREMENT O F SURFACE CHARGE
205
predominant factor for development of variable charge, and ions in solution are classified according to their influence.
A. PERMANENTCHARGE Isomorphous substitution in clay layers by cations of lower positive charge results in a built-in (permanent) net negative charge. Aluminum (A13+)may be substituted for silica (Si4+)in tetrahedral layers, and magnesium (Mg2+), iron (Fez+),or less commonly chromium ( C P ) , zinc (Zn2+),or lithium (Li+) may be substituted for AP+ in octahedral layers during clay mineral formation (van Olphen, 1963). Substitutions resulting in built-in positive charge have also been reported (Gillman and Uehara, 1980; Uehara and Gillman, 1980) but are questionable (Wada and Okamura, 1983). Substitutions of A13+ or Fe3+ for tetrahedral Si4+are responsible for permanent charge in zeolites. Phyllosilicates and zeolites are the mineral groups most responsible for permanent charge in soils (White and Zelazny, 1986). Laverdikre and Weaver (1 977) reported that additions of Na-montmorillonite to suspensions of spodic soils caused a decrease in the positive surface charge almost equivalent to the negative permanent charge of the clay added. Silicaaluminum co-gelsalso have site substitutionsand vacanciesthat can result in permanent charge (Parks, 1967). Espinoza et al. ( 1975) detected negative permanent charge in two volcanic soils that had very little crystalline clay. Perrott (1977) described evidence of possible permanent charge in amorphous aluminosilicates. However, the behavior of hydrated gels, glass, and x-ray amorphous aluminosilicates (allophane and imogolite) most often resembles variably charged minerals (Parfitt, 1980; Uehara and Gillman, 1981; White and Zelazny, 1986).
B. VARIABLE CHARGE Variable charge refers to surface charge that changes with variations in the PDI. The PDI for oxide and hydroxide minerals with low solubilities are protons (H+)and hydroxyls (OH-). Other speciesmay interact with PDI and affect their activity and thus the surface potential (Hohl et al., 1980). For salt-type minerals (for example, fluorite), ions that are an integral part ofthe mineral are the PDI, and surface charge is independent of pH (Parks, 1967). For calcite, Ca2+,HCOF, H+, and OH- are the predominant PDI (Somasun-
206
ANNE LEWIS-RUSS
daran and Agar, 1967) and surface charge depends on pH and Caz+ion concentration. Mineral surfacesinteract with PDI because of charge deficienciesresulting from crystal faces with unsatisfied electrical charges and from broken bonds. When oxide and hydroxide minerals are placed in an aqueous environment, unsatisfied charges are compensated by coordination with water molecules that dissociate into H+ and OH- (Parks, 1967).The resulting OH($ groups function as protonic acids, gaining or losing H+ or OH- depending on the pH of the solution and the dissociation constant (KJ of the solid-OH group. In acid solutions, for example, the surface becomes dominated by H+ and is net positively charged: solid-OH
+ H+ = solid-OH:
(2)
In basic solutions, the surface reacts with OH- and becomes net negatively charged: Wlid-OH
+ OH- = solid-0- + HZO
(3) Although minerals are consideredeither permanently or variably charged, these classificationsare idealized endpoints. Isomorphous substitutionscan occur in oxides and hydroxides, thus imparting some permanent charge to these “variably charged” minerals. Potential-determining ions such as protons and hydroxyls can compensate for permanent charge as well as variable charge, thereby causing the permanent charge to vary as solution pH varies (Schulthessand Huang, 1990). Single minerals may exhibit both permanent and variable charge. For example, the broken edges of permanently charged clay minerals are sites of variable charge. Uehara and Gillman ( 1981)classified common soil silicates as those dominated by permanent charge or by variable charge (Table 11). The variably charged chloritic class consists of montmorillonite and vermiculite that formed in acidic conditions and incorporated dissolved aluminum in the interlayers. These interlayers compensate for any permanent-charge substitutions so the charge character of these clays is variable (Uehara and Gillman, 1981).
C. SPECIFICALLY ADSORBED AND INDIFFERENTIONS All ions in solution influence surface potential or surface charge because increases in ionic strength cause increases in charge density. Some ions also have specific interaction with the charged surface. SpeciJicullyadsorbed ions have an affinity for a surface in excess of the electrostatic attraction (Lyklema, 1984)due to van der Waals forces or chemicalbonds (Breeuwsmaand
MEASUREMENT OF SURFACE CHARGE
207
Table I1 Dominating Charge Characteristics of Some Silicate Mineral Classes“ Permanent charge chloritic
Variable charge Chloritic (Aluminum interlayered)
Glauconitic Halloysitic Illitic Kaolinitic Micaceous Serpentinitic Montmorillonitic Vermiculitic Used with permission, from Uehara and Gill-
man (1981, Fig. 2.7).
Lyklema, 1973). These ions may be attracted to a surface regardless of net particle charge and may even cause charge reversal because their adsorption seemingly does not depend on Coulombic forces (Fuerstenau et al., 1981). For example, Fuerstenau et al. (1981) reported that barium (Ba2+)was adsorbed by rutile at pH values above the zero point of charge (ZPC). Potential-determining ions and specifically adsorbed ions both affect the surface charge, and the terms have been used interchangeably. More commonly, “PDI” is used to designate species that are ionic constituents of a mineral lattice and the ions in equilibrium with them (Parks and de Bruyn, 1962; Madrid et al., 1984), and “specifically adsorbed ions” refer to other adsorbing species in solution. For example, in a suspension of hematite (Fe20,) in a calcium chloride (CaCl,) solution, Fe3+,W-, H+, and OHwould be PDIs, Ca2+would be a specifically adsorbing ion, and C1- would be an indifferent ion. Indiferent ions can only adsorb on a charged surface and do so weakly. This is because the indifferent anions (chloride, C1-; perchlorate, C10,; nitrate, NO,) are strong acids that have weak affinities for H+, and the indifferent cations (sodium, Na+; potassium, K+) are extremely weak acids that fail to dissociate H+ from their surrounding sheath of water molecules until very high pH values are reached (Bowden et al., 1980b). Both indifferent cations and anions have a constant hydration state for the pH range of 3 - 10 (Bowden et al., 1980b).Indifferent electrolytesconsist of a cation and an anion that have either nonspecific adsorption or equivalent adsorption to the surface of interest. The simplest indifferentelectrolytesconsist of a single
208
ANNE LEWIS-RUSS
1 : 1 electrolyte, composed of a cation and an anion of similar hydrated radius and mobility, each ion capable of forming only outer-sphere surface complexes with the solid. Sposito (198 1) has called specifically adsorbing ions inner-sphere complexes in contrast to indifferentions that form outer-spherecomplexes with a surface. The two types of complexes have been differentiated based on the heats of adsorption. For example, Wayman (1967) used 10 kcal/mole as a cut-off value to distinguish between “physical” (outer sphere) and “chemical” (inner sphere) adsorption. More recently, instrumental techniques, such as electron nuclear double resonance spectroscopy and cylindrical internal reflection Fourier transform infrared spectroscopy, have been used to differentiateinner and outer sphere adsorption (Stumm and Wollast, 1990).
D. PROBLEMS WITH ION CLASSIFICATIONS Although differentiation of ions into PDIs and specifically adsorbing ions is meaningful for electrically conductive solids, such as silver iodide (AgI), the separation is rather artificial for soil systems. Natural minerals are rarely pure phases and dissolution can add lattice ions to the solution that are not normally mineral components. Protons and hydroxyls are generally accepted as PDIs for oxides, although they are not crystal constituents of the pure minerals. There are also several problems with the classification of ions as specifically adsorbed or indifferent. According to several authors (see, e.g., Davis et al., 1978; James and Parks, 1982) no ion is indifferent.Electrolyte ions form complexes with surface hydroxyls or surface sites and the strength of these complexes can be measured by potentiometric titrations done at various concentrations of background electrolytes. Both specifically adsorbed and indifferent ions can compensate for charge imbalances caused either by permanent or variable charge and may compete with H+ and OH- (Schulthess and Sparks, 1989). The attraction of ions changes with variations in solution conditions. Ions that are specifically adsorbed at some pH conditions may become indifferent at others (Healy et al., 1968). At all but extreme pH values relative to the zero point of charge (ZPC), positively and negatively charged sites can coexist on a particle. At the ZPC the net surface charge is zero, but both positive and negative sites are usually present. This coexistence is particularly important in adsorption studies because it means that anions can adsorb at pH values above the ZPC by occupying the subset of positively charged sites (Barrow and Ellis, 1986). Therefore, “specifically adsorbed” may be operationally difficult to define because the ion may be adsorbing on a minority of oppositely charged sites.
MEASUREMENT OF SURFACE CHARGE
209
Such adsorption is probably the mechanism of much trace metal adsorption below the ZPC, as in the adsorption of uranyl by amorphous femc oxyhydroxide, which is nearly complete by pH 5 , well below the measured ZPC value for amorphous femc oxyhydroxide of 7.9 (Hsi and Langmuir, 1985). As Langmuir (1 979) reported, inner- and outer-sphere complexes are ideal definitions of what, in fact, is a continuum of behavior.
IV. MEASUREMENT OF SURFACE CHARGE Because so many properties of particle suspensionsare affected by surface charge, its measurement should be straightforward. Parks and de Bruyn ( 1962) reported that the ZPC is the point of minimum solubility for a solid phase in equilibrium with the solution. Dissolution is promoted by adsorption of H+ or OH-. The slowest rates of dissolution occur in the region of the ZPC, where the predominant surface sites generally are neutral. Coagulation and sedimentation rates are maximum at the ZPC (Parks, 1967). A suspension is stable because of the repulsive effect of like-charged particles. At the ZPC, the charge is neutralized, enabling closer approach of particles and, due to van der Waals attraction, coagulation, and settling. Breeuwsma and Lyklema (1973) used the charge characteristics caused by the adsorption of phosphate ions to explain the stability of a hematite sol in the presence of phosphate. Hohl et al. (1980), using equilibrium constants for phosphate - goethite surface reactions, produced model plots of areas of expected coagulation that were similar to the experimental results of Breeuwsma and Lyklema (1973).
A. POTENTIOMETRIC TITRATION The ZPC is an expression of the electroneutrality of a surface, the point at which positive and negative charges are equal. This point can be determined by potentiometrictitrations because at every point in a titration, the sum of positive charges is equal to the sum of negative charges for the solution as a whole. Positive charges are due to the positive surface sites, and to H+ and other cations in the solution. Negative charges are due to the negative surface sites, and to OH- and other anions in the solution. Titrations are done by adding known volumes of acid or base to a rock or mineral suspension and recording the pH change for each addition. The titrations are repeated in solutions of different ionic strengths of an indifferent electrolyte. If the electrolyte is assumed to have no net effect on surface charge, the
210
ANNE LEWIS-RUSS
ZPC can be determined from the change of pH, which is a measure of H+and OH- remaining in solution. As ionic strength increases, the density of ions close to the surface increases. This density increase allows more surface charge to develop, making the net surface charge versus pH curves steeper (Bowden et ul., 1980b). The shielding of the surface by the ions has the opposite effect on electrophoretic mobility (discussed in Section IV,B). As ionic strength increases, the electrophoretic particle mobility decreases. These separate effects are shown in Fig. 1 for rutile. At the ZPC, the overall surface charge is neutral and, therefore, is independent of the ionic strength of an indifferent electrolyte. The pH at which titration curves for several ionic strengthsof solution cross is the point ofzero salt effect (PZSE). At pH values above the PZSE, the surface charge is negative (excess OH-), and below the PZSE the surface charge is positive. If the electrolyte does have a net effect on surface charge, the PZSE will not occur at the same pH as the ZPC. Therefore, PZSE is an operational term, referring to a measurement technique, and ZPC is a conceptual term. Although potentiometric titration is the most common technique of determining ZPC, especially for oxides and hydroxides, the method is not without problems in execution and in interpretation (Zeltner, 1986).When calculating the surface charge, the volume of acid or base consumed changing the pH of the electrolytesolution needs to be subtracted. This volume to be subtracted can easily be determined by titrating a "blank" of electrolyte solution alone if the particle of interest has minimal solubility.However, the situation becomes more complex as the solubility of the particle increases. For complex geologic materials, the overall ZPC may be at a pH where the solubilitiesof some componentsare enhanced. Schulthessand Sparks( 1986) reported that, in such a situation, the solubility of the particle and the speciation of the dissolved species need to be known, but may not be easy to obtain. The problem of soluble test materials may be resolved by subtracting the proton consumption of the appropriate supernatant. Murray (1974) titrated the centrifuged supernatant from a MnOzsuspension when determiningthe ZPC of hydrous MnO, and used this as the blank. Huang (1981) used a similar technique with y-Al,03, centrifugingto obtain the supernatant after stimng the y-Al,03 in the electrolyte solution for 30 minutes. Although this procedure may be superior to using a simple electrolyte for a blank, it does not account for the increased (or decreased) solubilitiesat pH values different from those of the original electrolyte-particle suspension. For changing solubilities, a butch titrution technique, using supernatant obtained from suspensionspreadjusted to a range of pH values, can be used. The batch, or serial-titrution technique, was developed by van Raij and Peech (1972) to determine the ZPC of some tropical soils. Separate soil
MEASUREMENT O F SURFACE CHARGE
21 1
10
"E 0
34 4
-
0
L
iii
0' Q)
-10
0
a
d
-20
4
5
6
7
8
9
10
11
PH Fig. 1. (a) Effect of varying ionic strength on surface charge density of rutile as measured by KNO,, 0.1 M, @), KNO,, 0.01 M, (O),KNO,, 0.001 M . (Data potentiometric titration. from James ef a/. 198 1.) (b) Effect of varyingionic strength on electrophoretic mobilityof rutile. (m), NaNO,, 0.005 M, (O), NaNO,, O.OOO1 M, (O), NaNO,, O.oooO1 M. (Data from Fuerstenau et nl., 198 1 .)
m,
212
ANNE LEWIS-RUSS
samples were used to determine each point of the titration curve. Each soil - electrolyte-acid (or base) mixture was equilibrated in closed containers for 3 days before the pH was measured. Madrid et al. ( I 983) used a similar technique for iron oxides and some soils, flushing the suspensions with and measuring pH nitrogen (N,) gas that was free of carbon dioxide (CO,), values after the suspensionswere shaken for 1 hour. For both these applications, the blank consisted of the simple electrolytethat had not been equilibrated with the soils. Schulthess and Sparks (1 986) compared several types of blanks in their batch titration of pA1203.Suspensionswere shaken overnight before measuring pH. Titrations of the electrolyte solution alone were compared to titrations of the supernatant obtained by equilibrating the suspension overnight and then centrifugingand filtering. All points plotted on the calculated ideal titration curve, indicating minimal solubility of y-Al2O3 at the zero point of titration (ZPT) of the suspension. However, titration curves of supernatants from initially acidic or basic batches diverged, indicating that dissolution vaned with pH (Schulthess and Sparks, 1986). To account for changing solubilitiesat different pH values, Schulthessand Sparks (1986) developed a backtitrution technique. After the pH was measured in y-A1203- electrolyte- acid (or base) suspension that had been shaken overnight, the supernatants from each batch were separated and titrated to pH 7 using acid or base. The volume of acid or base needed for the backtitration was subtracted from the original volume of acid or base added to the suspension to obtain the quantity of H+ adsorbed or desorbed. When the titration points for four different ionic strengths were plotted, the curves crossed at one point (PZSE),but otherwise the curves were more parallel and less smooth than typical titration curves. In addition to blanking, there are other problems with potentiometric titration. For some minerals (for example, quartz), low pH values are needed to define the PZSE. Outside the pH range of 4- 10, sensitivity decreases as concentrationsof H+ or OH- increase exponentially. Blank corrections for titrations below pH 4 or above pH 10 require the subtraction of much larger quantities of H+ (or OH-). The loss of sensitivity is not compensated even though more surface sites may be protonated (or hydroxylated) (Bolt, 1957; Zeltner, 1986),because generally, surfaceprotonation increases only slightly as pH decreases (Stumm and Wollast, 1990). Another problem inherent in potentiometric titration is that colloids may tend to aggregate during the titration. This effect is enhanced close to the ZPC, and as ionic strength is increased. The aggregation of colloids may decrease the readily accessible surface area, resulting in diffusion-controlledadsorption kinetics (Zeltner, 1986). More work needs to be done in this area to define the effect of
MEASUREMENT O F SURFACE CHARGE
213
aggregation on surface charge measurements(W. A. Zeltner, personal communication 1990).
B. ELECTROKINETICTECHNIQUES Electrokinetic techniques provide valuable information about surface charge by direct measurement of charge-induced particle or liquid movement. This section briefly introduces some electrokinetic techniques and provides a few examplesof applicationsto surface charge measurement. For a more comprehensive discussion, “Zeta Potential in Colloid SciencePrinciples and Applications” (Hunter, 1981) is recommended. Electroosmosis(Smit and Holten, 1980), streaming potential (Somasundaran and Agar, 1967; Kuo and Yen, 1986; Sharma et al., 1987; Kuo et al., 1988),and electrophoretic mobility (Healy et al., 1968;Murray, 1974;Houchin and Warren, 1984; Hansmann and Anderson, 1985; Bryant and Williams, 1987;Carroll-Webband Walther, 1988;Schulthessand Sparks, 1988) can be used to determine the zeta potential, which can be related to the surface charge in some circumstances (Hunter, 1981; Dzombak, 1986). These electrokinetic techniques all directly measure the electric potential (zeta potential, () developed when the liquid and solid phases move in relation to each other (Hunter, 198I). In electroosmosis, the application of an electric field applied to the solution causes the liquid phase to move along a stationary,charged solid surface (van Olphen, 1963).The measurement of the liquid velocity or of the volume of liquid transported per unit of current flow provides information about the charge on the solid surface (Hunter, 1981). In streaming-potential measurements, the solution is forced to flow along a charged surface, creating an electric potential. In electrophoresis,an electric field is applied to a suspension, which causes charged particles to migrate to oppositely charged electrodes. Average particle velocity is used to calculate the charge per particle. The plane of movement, or slipping plane, is located at some unknown distance from the particle surface. Although the plane is often assumed to coincide with the outer limit of specific adsorption (or Stem layer), this approximation is only correct for small potentials and weak ionic strengths (Dzombak, 1986).As ionic strength increases, the potential near the particle surface is larger. Viscosity of the solution increases so more solution moves with the particle and mobility decreases (Fig. I b). Depending on cell configuration, the ionic strength of solutions measured with electrophoresis is limited by the increased viscosity and by coagulation and sedimentation. The interpretation of electrophoreticmobility as zeta potential (0is based
214
ANNE LEWIS-RUSS
on the evaluation of the forces affecting the particle. While appropriate relations have been derived for simple particle geometries, mobility/zetapotential relations for irregular particles are unknown. Oxide particle shapes have been approximated as spheres or cylinders (Dzombak, 1986). These approximations limit the interpretation of electrophoretic data as diffuselayer potential. However, determinationof the zero zeta potential or isoelectric point (IEP)is direct when measured in an indifferent electrolyte because, at this pH, both potential and mobility are zero. This direct determination means that, unlike potentiometric titration, mobility only needs to be measured at one ionic strength. Theoretically, the IEP would be equal to the ZPC if there is no specific adsorption, but Zeltner (1986) gives numerous examples from the literature of this not being the case. Zeltner (1 986) reported that the IEP he determined for goethite by electrophoretic mobility was 0.4 pH units higher than the ZPC determined by potentiometrictitrations. He suggested that such differences resulted from the more highly hydrated cationsbeing farther from the surface than anions. At the ZPC, any adsorbed cations would be slightly closer to the slipping plane than adsorbed anions, resulting in a slightly positive zeta potential. More balancing hydroxyls would be required to reach zero mobility, resulting in a higher IEP. Zeltner (1986) proposed that the difference between IEP and ZPC could be used as a rough estimate of the number of charged sitespresent at the ZPC, because more protruding cations would cause more separation between IEP and ZPC. When the IEP is lower than the ZPC, contamination by silica (Le., from glass storage containers) is to be suspected (Hunter, 198 1). Electrophoresis has been used to determine the IEP for colloids that have low ZPC values not amenable to potentiometric titration measurement. Healy et al. (1 966) reported excellent agreement of IEP and ZPC values of manganese dioxide solids (MnO,) obtained using microelectrophoresis and coagulation-sedimentation. Healy et al. ( 1968) used coagulation and electrophoretic mobility to measure the ZPC of quartz. Murray (1974) was able to measure the ZPC of 6MnO2 using electrophoresis and Na+ and K+ adsorption. Carroll-Webb and Walther (1988) used the asymmetry of their electrophoreticmobility curve for kaolinite as evidence supportinga two-site model for surface charge. The pH range of electrophoreticmeasurements is limited by instrumental considerations and by particle response. Platinum electrodes and instrument membranes can be damaged by extreme pH values. At extreme pH values, H, and 0,gases can form at the anode and cathode, particularly at high ionic strengths and applied voltage. Use of reversible electrodes eliminates this problem. Particle settling, which interferes with measurements, occurs at low pH values (Murray, 1974). Williams and Williams (1978)
MEASUREMENT OF SURFACE CHARGE
21s
reported that the settling caused less change in the velocity profile when using a rectangular cell. Streaming potential implicitly averagesthe zeta potential of the surface so the measurements are not as sensitive to small surface pores as are potentiometric titrations (Sharma et al., 1987). Sharma et al. (1987) compared the surface charge of untreated and acid-leached mixed oxides measured with potentiometric titration, streaming potential, and electrophoresis. Acid leaching, which probably resulted in surface pitting and etching, caused significant changes in potentiometric titration measurements, whereas streaming potential and electrophoresismeasurementswere unaffected. Because of these variations, Sharma et al. (1987) recommended that the last two techniques be used for naturally occurring samples. Somasundaran and Agar ( 1967)used a unique combination of streaming potential, solution equilibrium, and flotation measurements to bracket the ZPC of calcite. Regazzoni et al. (1983) compared potentiometric titration, electrophoretic mobility, and the mineral addition technique (discussed in Section IV,E) for measuring the ZPCs of magnetite (Fe30,) and zirconium dioxide (ZrO,). Results from potentiometric titration and electrophoretic mobility were similar, whereas the addition method produced values that were 0.4 pH units lower. Streaming potential measurement is better adapted to large particles or whole core samples; however, this technique seems to be greatly affected by particle size, surface area, and pore size (Kuo and Yen, 1986; Kuo et al., 1988). Electrophoretic mobility is more practical for fine particles when sedimentation during measurement is not a problem.
C. ION RETENTION The ion retention technique (Schofield, 1949) is based on changes in ion retention with varying surface charge. Modifications of this technique have been used extensively in soil studies to measure the combined effects of permanent and variable surface charge (see, e.g., van Raij and Peech, 1972; Gillman and Uehara, 1980;Bolan et al., 1986). The soil is equilibrated with an electrolyteat a variety of known pH values. The equilibrated electrolyteis then measured to determine the concentrations of cations and anions remaining on the soil surface. The pH value at which cation and anion retentions are equal is the point of zero net charge (PZNC). An example of results from this technique is shown in Fig. 2. A variation of the ion retention technique was used by Murray (1974) to determine ZPC. He measured the adsorption of Na+ and K+ in suspensions of 6-Mn0, at different pH values by withdrawing aliquots of suspension at
216
ANNE LEWIS-RUSS
0.2
0 \ U
;
0.1 -
6
P a
0
0.0
6
n
n
n
6.. ...............@-""---.@.
[email protected]
0
8
n
-0.1
-
-0.2
I
I
I
I
1
I
I
3
4
5
6
7
8
9
Ic b.
z
10
PH 0.2
D \
U
;
0.1
-
ai
P a
5
6!............. -..la-.....
u
0.0
n
V
.......
n 3 u-. ....O...-@
Q)
2
-0.1 -
Ib. c
z
-0.2
3
I
I
1
I
I
I
4
5
6
7
8
9
10
PH Fig. 2. Ion retention method of determining the point of zero net charge (PZNC).0,Cladsorption; (0),Na+ adsorption;(O),net charge. (Data from Madrid et al., 1984.)(a) Lepidocrocite and a t e , 50 :50 mixture in 0.01 MNaCI, PZNC not definable. (b) Lepidocrocite in 0.0I MNaCI, PZNC about equal to pH 8. (c) U t e in 0.01 MNaCI, PZNC not definable.
217
MEASUREMENT OF SURFACE CHARGE
2
4
6
8
10
PH Fig. 2. (continued)
various points during potentiometric titrations. Anion adsorption was not measured. The pH at which there was no measurable cation adsorption was interpreted as the ZPC, and was in agreement with the ZPC values determined from other techniques. Espinoza et al. (1975) used a similar technique, equating the lack of anion (NO;) adsorption on soils to the ZPC. The usefulness of the ion retention technique depends on choosing an appropriate electrolyte that will replace all surface-adsorbed ions without forming charged inner-sphere surface complexes(Sposito, 198 1, 1984). Different electrolytes will result in different values for the surface charge; but, according to Sposito (1984), if the electrolyte and experimental conditions are similar to the natural setting, the resulting surface charge will be of practical use, even though it may not be the optimal value. Another problem with this technique is that, similar to other cation and anion exchange methods, charge balance by protons and hydroxyls is not considered.
D. SALTTITRATION The salt titration technique consists of titrating a sample suspension of known pH with an indifferent salt. Gillman and Uehara (1980) suggested a batch technique in which samples are equilibrated at various pH values in solutions of weak ionic strength. After several days, the pH is recorded and
218
ANNE LEWIS-RUSS
the ionic strength increased. After a second equilibration of 3 hours, the new pH is recorded and compared with the first value. The pH that does not change is independent of salt concentration and, therefore, is the PZSE (Table I). A slight variation of this technique was used by Pyman et al. (1979b) to determine the PZSEs for oxide mixtures. For ZPC values higher than 5.5, Pyman et al. (1979b) passed C0,-free nitrogen gas through the suspension for 1 hour prior to pH adjustment and salt addition. Agreement of the PZSE values with the values from batch or potentiometric titration was within experimental error (k0.2 pH units). Zeltner ( 1986)attempted salt titrations of goethite suspensions using 5 M NaN03. Although the pH of the starting goethite suspensions bracketed the ZPC, the salt titration curves failed to converge with added salt solution. Both curves tended toward lower pH values as ionic strength was increased. Zeltner ( 1986)suggestedthat the acidic pH of the salt titrant was dominating pH changes, and suggested using an electrolyte solution that had a pH similar to the ZPC. Choosing such an electrolyte assumes preknowledge of the ZPC, which is not a simple matter for a complex material.
E. MINERAL ADDITION The mineral addition technique is similar to batch potentiometric titrations. The order of addition and the time ofcontact vary. Rather than adding acid or base to a suspension, in mineral addition a measured portion of ground mineral is added to a prepared solution. Rather than a single pH measurement, the pH is usually read several times to establish the trend of pH movement and the quasiequilibrium pH. A series of surface charge versus pH curves can be drawn, treating each mineral addition batch test as a single point. The crossover point is the PZSE, just as in potentiometric titrations. Alternately, the convergenceof batch tests with starting pH values that bracket the ZPC indicates the ZPC. An early application of this technique was reported by Bradfield (1923). Ahmed and Maksimov (1969) used a variation of the mineral addition technique to determine the PZSE values for cassiterite (SnO,) and rutile (Ti02). After the mineral was added to solutions varying from 0.001 to 1 M JSNO, ,the pH was measured at intervals of 1 or 2 minutes for 8 minutes. Ahmed and Maksimov ( 1969)reported that measurements were repeatable within k5% near the ZPCs. The apparent ZPC values increased slightly for both minerals as ionic strength of the test solution decreased. At pH values below the ZPCs (which were at pH values of about 5 for both minerals), the calculated surface charge density was independent of KNO, concentration.
MEASUREMENT O F SURFACE CHARGE
219
Tschapek et al. ( 1974)used the second variation of the mineral addition technique for Al,O,, measuring pH at intervals for 14 minutes. The apparent ZPC value determined was slightly dependent on the initial electrolyte concentration. The pH values determined in 0.01 MKCl and 1 MKCl(8.0 and 7.9,respectively)were similar to the value of 7.8 determined by potentiometric titration. The mineral addition technique used by Regazzoni et al. (1983)consisted of adding a sample of Fe,O, or ZrO, to 0.1M KNO, of known pH. The ZPC was determined by measuring changes in pH for 10 minutes. Husain (1984)used mineral addition to measure the ZPC of uranium oxide (UO,). Measurement of pH was continued until the readings were stable (about 6 minutes).The apparent ZPC values decreased slightlyas the concentration of the LiNO, electrolyte was increased. The use of the mineral addition technique has not been widely reported in the literature. One problem may be the variation of ZPC as ionic strength is changed (Tschapek ef al., 1974). All three of the studies cited here that used more than one ionic strength reported a decrease in apparent ZPC as ionic strength was increased.This effect is not commonly reported for the potentiometric titration technique even though mineral addition and potentiometric titration are similar methods. Possible causes of this are discussed in Section VI,B.
V. SEPARATING PERMANENT AND VARIABLE CHARGE Permanent charge and variable charge occur together in complex mineral systems, such as weathered rock and soils, and on some individual minerals, such as clays (Rand and Melton, 1975). When a negative, permanently charged component is large enough, it may be predominant so that no ZPC occurs in typical soil pH conditions (Madrid ef al., 1983,1984;Bolan et al., 1986;Hunter and Busacca, 1987). Separate measurements of the two components of charge may be possible when they are more balanced. Three techniques have been used. The first technique is a simple fitting method applied by Barber and Rowel1 (1972).They measured positive and negative adsorption of ammonium (NHt) and C1- on an iron-rich kaolinitic soil at different pH values and at differentconcentrations of adsorbates. Data did not fit a Freundlichtype model for development of concentration-dependentcharge:
(4) charge = x(n)lI2 where xis a constant and n is concentration (from van Olphen, 1963).Larger than predicted values for charge were accountedfor, in part, by overlap ofthe
220
ANNE LEWIS-RUSS
diffuse layers of the clay colloids. Negative permanent charge was used as a fitting variable because the change in negative charge was larger than the change in positive charge. The second technique uses the equilibrium pH attained between a solid and solution before acid or base additions, the ZPT. In this technique, the ZPT is considered to be similar or identical to the ZPC of the variable component of surface charge, and the displacement of the PZSE from the ZPT is used to define the permanent charge. This technique was applied to soils by van Raij and Peech (1972). For five of six soils van Raij and Peech ( 1972) tested, the PZSE was acidic compared to the ZPT. The concentration of H+required to achieve the PZSE was almost equal to the concentration of exchangeable A P . This equality is not an indication of permanent charge, however, unless complete exchange of A13+ by H+ occurs, and no H+ is consumed by other reactions. Parker et al. (1979) reported that ZPT minus PZSE overestimated the cation exchange capacity of the soils they tested. Their work indicated that about 30-90% of the acid or base lost from solution did not contribute to a change in surface charge; therefore, other reactions were consuming H+ or OH-. Such reactions (discussed in Section VI,B, 1) could move the PZNC and PZSE closer together or farther apart, or could result in fortuitousagreement of the two measurements (Parker et al., 1979).
Espinoza et al. ( 1975) compared potentiometrictitration, Nopetention, and the ZPT for two volcanic soils. They reasoned that negative permanently charged surfaces will not retain OH-, so titration curves above the ZPT indicate only variable charge,but titration curves below the ZPT indicate H+ adsorption by both permanently and variably charged parts of the soil. The adsorption of an indifferent anion (such as NO;) indicates the extent of the positive charge. Espinoza et al. ( 1975) reported that Nopetention decreased as pH increased, and seemed to confirm that ZPC for the variably charged fraction of the soil was similar to the ZPT. Espinoza et al. (1975) also reported that ZPC for the variably charged component could not be unambiguously defined from decreased anion retention because cation and anion retention are equal, but not necessarily zero, at the ZPC. However, the decrease in anion retention should presumably give a general indication of the ZPC value, because positive charge and indifferent anion retention a p proach their smallest values near the ZPC. Espinoza et al. (1975) assumed that the cation exchange capacity (CEC) measured at approximately the ZPT indicated the permanent charge. Again, complete exchange was assumed, possibly incorrectly. The third technique of separating permanent and variable charge assumes that the PZSE is equivalent to the ZPC of the variably charged component only. When the pH of a solution is adjusted so the variably charged compo-
MEASUREMENT O F SURFACE CHARGE
22 1
nent is zero, any residual charge can be attributed to permanent charge. This permanently charged component can be measured as the difference between cation and anion retention at the ZPC of variable charge (Gillman and Uehara, 1980;Uehara and Gillman, 1980,1981;Singh and Uehara, 1986). This measurement is done operationally by defining the PZSE by potentiometric titrations and then measuring the cation and anion exchange capacities at this pH value. Again, assumptions about the completeness of exchange are made. Also, why the PZSE would define only the ZPC for variable charge rather than for the composite ZPC is not clear. Madrid et al. (1984)used this charge-separation technique in a study of mixtures of iron oxide (lepidocrocite)and illite. The permanent charge was estimated as the difference between cation and anion retention at the PZSE. For example, for the 50 :50 mixture of lepidocrociteand illite, the PZSE was measured as 7.24.Referring to Fig. 2a, it can be seen that at a pH of 7.24there is an excess of negative surface charge, as measured by ion retention. This excess charge is assumed to be permanent charge. By contrast, the PZSE for lepidocrocite alone was 7.35.Referring to Fig. 2b,at a pH of 7.35 there is a slight excess of positive surface charge. This, apparently, indicates a slight error in the ion retention technique, because PZSE should equal PZNC for this oxide. The PZSE for illite is indeterminate (Fig. 2c) because negative surface charge predominates. Madrid et a/. (1984)reported that the values determined for permanent charge increased as the ionic strength of the test solution was increased. This may indicate variations in the exchange capacities of the ion retention techniques. Madrid et al. ( 1984)also warned of difficultieswith rigorous mathematical treatment of materials that had permanent and variable charge and questioned the model derived by Uehara and Gillman (1980). Gonzales-Batistaet al. (1982)compared the second and third techniques of separating permanent and variable charge (ZPT minus PZSE versus the difference between cation and anion retention at the PZSE) for six allophanic clays. There were no differences between ZPT and PZSE for four samples. Permanent charge was indicated for the two remaining samples. These samples were the least weathered, had the largest Si :(Si Al) ratios (0.56and 0.64 compared to 0.35-0.47for the other samples), and were probably composed of the most reactive materials. Results of the third technique indicated permanent charge for all samples. For one sample, the large difference between the cation and anion retention at the pH of the PZSE ( 125 meq/kg) was attributed to specifically sorbed sulfate (Se-that ) was not displaced during anion exchange. Differences between the two techniques for the five remaining samples varied from 15 to 35 meq/kg. Gonzales-Batista et al. ( 1982)cautioned against using either charge-separationtechnique, especially when PZSE occurs at low pH values (below 4.5). At such pH
+
222
ANNE LEWIS-RUSS
values, clay minerals may become soluble, adding PDI other than H+ and OH- to the solution. If the permanently charged component could be isolated, any residual charge could be attributed to variable charge. The problem is in charge separation. There are two problems with currently used techniques of measuring permanent charge by cation and anion retention. The first problem is determining if the ion exchange is complete. The second problem is that these measurements are done at varying pH values, and the tested material may change as the solution pH changes.
VI. MEASUREMENT PROBLEMS The surface charge is a function of the material tested and its interaction with the aqueous solution. Changes to the material during sample preparation or during establishment of particle-solution equilibrium will alter the surface charge and the ZPC measured.
A. ALTERING THE SOLIDPHASE There are generally two kinds of materials tested for surface charge: synthetic precipitates and naturally occurring solids. Aging of precipitates can cause changes in the ZPC (Parks, 1965, 1967; Zeltner, 1986), and naturally occurring solids can be altered by material preparation. Healy and Fuerstenau (1 965) noted that ZPC values tend to shift to higher pH values as the precipitate changes from amorphous to crystalline phases with aging. Murray ( 1974) attributed changes in ZPC values for SMnO, to the condensation and dehydration occurring as the freshly precipitated mineral ages. Husain (1984) reported that the ZPC of UO, samples shifted with aging from pH 5.8 (fresh samples) to pH 4.8 (aged 36 days). This pH shift may indicate partial surface oxidation or increased crystallinity (Parks, 1967). The kinetics of COzadsorption on these oxides can also be involved in the aging effects (C. P. Schulthess, 1990 personal communication). Healy et al. ( 1966) reported that the ZPCs of a series of manganese oxides increased as atomic packing in the crystal lattice increased. The increased packing results in an increased electrostatic field above the lattice due to the change in the types of bonds that are broken in the fracture process. When these bonds are of the nonionic van der Waals type, a nonpolar surface is in contact with water. Conversely, fracturing of ionic or covalent bonds creates a polar surface that interacts strongly with water molecules, increasing the
MEASUREMENT OF SURFACE CHARGE
223
ZPC. Parks ( 1965) noted similar trends of increasing ZPC with increasing metal :oxygen crystalline ratio. The acidity of the fluid in contact with an aging precipitate also can affect ZPC. In acidic solutions, H+ may adsorb onto oxide surfaces, inhibiting adsorption of more protons and thus producing a more negative charge (Sharma et al., 1987). Zeltner (1986) reported that goethite aged in acid solutionshad a higher ZPC than when aged in basic solutions. Other changes can occur during aging. Nail et al. (1976) reported that the pH changed substantially(a decrease of about 2 pH units) during aging of aluminum gels. Swallowet al. ( 1980)reported that H+ consumption by hydrous ferric oxide decreased (and the measured PZSE increased) as the precipitate aged and stabilized after 24 hours. They attributed the changes to release of H+ (or consumption of OH-) by the chemically precipitated oxide. Zeltner (1986) reported that the ZPC of goethite changed as the precipitate aged for several months. He also attributed these changes to H+ desorption. Most geologic materials need to be modified before testing for surface charge. These changes include sizing the particles by grinding, sieving, or both; cleaning the particle surface; or saturating the particles with uniform exchangeable ions. Any sample pretreatment that modifies surface characteristics would be expected to change the ZPC. A single mineral may have several different independent sites, each with its own ZPC. Modifications of surfaces could favor some sites, changing the proportion of the sites and, thus, changing the measured ZPC (White and Zelazny, 1986). The disturbed surface layer resulting from grinding differs from the underlying mineral and is more susceptible to hydration. Rapid hydration of cu-Fe,03 causes a surface layer of a-FeOOH. Parks ( 1965, 1967)noted that hydrous oxides had higher ZPC values than did corresponding dehydrated oxides. For Fe-, Al-, and Ti-oxides, ZPC values of the hydrous oxides were an average of 2 pH units higher than ZPC values of corresponding dehydrated oxides, The disturbed surface layer on a-Al,O, hydrates in water to produce a surface layer of gibbsite that has a higher ZPC than the underlying mineral (Smit and Holten, 1980).Grinding can cause a substantial loss of crystallinity, an increase in surface area, and greater release of metal ions (Torres Sanchez et al., 1988).Ultrahe particles created by grinding have enhanced rates of dissolution (Holdren and Berner, 1979; Helgeson et al., 1984). The increased dissolution can add new PDI to the suspension and result in hydrolysis reactions that competewith the surface for H+. Grinding time has been correlated with increased CEC and PZSE (Torres Sanchez et al., 1988), as shown in Fig. 3. The increased CEC can be attributed to an increase in the number of exchange sites (C. P. Schulthess, 1990personal communication). Acid washing of geological materials may dissolve femc or aluminum hydroxides and unblock sites of permanent charge. The increase in exposed
224
ANNE LEWIS-RUSS
0
100
200
300
400
500
600
700
800
Grinding Time, Seconds Fig. 3. Changes in cation exchange capacity (CEC) and point of zero salt effect (PZSE)of kaolinite with increased impact crushing. m,CEC; 0,PZSE. (Data from Torres Sanchez et al., 1988.)
negative permanently charged sites w ill lower the overall charge, resulting in a lower PZSE. Acid washing could conceivably dissolve permanently charged sites, resulting in higher values for PZNC and PZSE. Measurements of PZNC and PZSE are made over a range of pH values, resulting in a further source of ambiguity. The dissolution of both permanently charged sites and oxides blocking these sites depends on pH. The variations in stability for different pH values could cause changes in surface characteristics during potentiometric titrations and during ion retention measurements. Gillman and Bell ( 1976) reported that acid washing of soil samples at pH values less than 2 caused significant changes in PZSE. The decrease in PZSE was attributed to dissolution of aluminum-gel coatingsthat had blocked sites of negative permanent charge on clays. Parker et al. (1979) observed increased CEC (originally measured at pH 5.45) after acid washing (at pH 2.0) soils. For one soil, the CEC decreased, perhaps due to dissolution of less stable phyllosilicates. Madrid et al. (1983) reported that illite can be easily altered in acidic solutions, whereas it is more stable in basic solutions. Ion-saturationtechniques, frequently used in studiesof clay minerals, can affect the ZPC in ways similar to acid washing. The concentrated solutions used for ion-saturationpromote the replacement ofexchangeableions by the electrolyteion. Divalent and trivalent cations are exchanged for monovalent
MEASUREMENT O F SURFACE CHARGE
22s
cations. During a subsequent potentiometric titration, the monovalent cations are more easily replaced by H+ than are divalent or trivalent cations. Treated materials would consume more H+ by exchange, resulting in a measured PZSE that was lower than that for the same, but untreated materials. Treated samples could also have more negative permanent charge. This would depend on the efficiency of the cation exchange technique in replacing divalent and trivalent, as well as monovalent cations. Laverdiere and Weaver ( 1977)reported a decrease in PZSE for four of the six soils they tested when pretreated by Na+-saturation.They suggested that, especially with acidic soils, A13+ was present as an exchangeable ion, which blocked negative permanent charge during titration unless first displaced. Hendershot (1978) also reported that measured charge characteristics for NaC1-saturated samples differed from untreated samples. The NaC1-treated samples had more measured permanent charge (defined as the displacement of the PZSE below the ZIT) and lower PZSEs compared to untreated samples that had PZSE values less than 5. The variation between treated and untreated samples decreased as the measured permanent charge decreased and the PZSE increased. Hendershot (1978) attributed these changes to replacement of A13+by Na+, which unblocked sites of cation exchange.
B. INTERACTIONSWTH THE ELECTROLYTESOLUTION Once the material is in the test solution, aqueous-phaseeffects can alter the measured ZPC. In addition to the H+ adsorption during aging discussed in Section VI,A, the pH, ionic composition, and ionic strength ofthe electrolyte can cause surface charge variations that have complex relations to H+ and OH- adsorption. Sposito ( 1984)warned that a measurement of ZPC can be accurately obtained only if there are two final states for H+ or OH- in the solution/solid system: as a free ion in solution or as part of a pH-dependent, surface functional group. Competing reactions that consume H+ or OHmay be impossible to suppress, especially when titrating soils that have exchangeablecations or readily soluble, hydroxy polymer coatings, or both. The ZPC is almost always determined by placing the test material in an aqueous solution that contains a background electrolyte to ensure ease in measuring pH and to mask effects on surface charge caused by ionic strength changes during titration. Ideally, the electrolyte is indifferent. When the electrolyte is not indifferent, unbalanced adsorption of one ion will shift the ZPC. Even when using indifferent electrolytes, ion exchange reactions may need to be considered, especially at high ionic strengths (Schulthess and Sparks, 1987).
226
ANNE LEWIS-RUSS
1. pH Effects and Competing Reactions
Measurements of particle surface charge, such as potentiometrictitration, assume that H+no longer present in a solution (as determined by pH values) is adsorbed on the surface by negative sites. Other reactions with H+ that can occur in the same time frame as H+ adsorption are possible; for example, complexation and ion exchange (Langmuir and Mahoney, 1985). As particle- solution contact time increases, as in overnight batch titrations, oxidation - reduction and mineral dissolution-precipitation can result. In addition, proton -carbonate reactions become likely at basic pH values, when there is access to atmospheric COz. All these reactions can consume H+ and, if unaccounted for, will result in a false value for surface charge and ZPC. Competitive reactions have been used to account for surface charge characteristics deviating from “ideal behavior” (Espinoza et d.,1975; Hendershot, 1978; Madrid et al., 1984). Parker et al. ( 1979) suggestedthree competitive reactions that could consume H+ without affecting surface charge: ion exchange of H+ for surface-adsorbedcations, partial neutralization of adsorbed hydroxy A13+ (or Fe3+), and dissolution of aluminum (or ferric) hydroxides. Concern about these competitive reactions has resulted in two types of responses. Schulthess and Sparks (1986) developed a technique for dealing with such reactions implicitly, whereas Zeltner (1 986), Schulthess and Sparks (1 987) and Carroll-Webband Walther (1988) dealt with them explicitly. The technique of Schulthessand Sparks(1986), described in Section IVY A, uses backtitration, that enabled them to subtract the consumption of H+ by species present in solution due to dissolution of the solid. In relatively simple systems where dissolution is the main competing reaction, such as y-Al,O,-indifferent electrolyte, or in more complex systems where dissolution-precipitation occurs but the products are not known, this technique is an improvement over traditional potentiometric titration techniques. However, backtitration will not account for other reactions that consume protons. For example, Schulthessand Sparks (1986) discussed the possibility of competition of H+ and OH- with electrolyteions, especially at higher ionic strengths. In a later publication, Schulthessand Sparks (1 987) developed a more complex, explicit model that accounted for surface reactions with aqueous COz and competitive reactions with electrolyte ions, as well as dissolution. Schulthess and Sparks (1987) then used this model to interpret the irregularities of the titration curves that had resulted from backtitration of y-Alz03. Carroll-Webb and Walther ( 1988) explicitly accounted for dissolution of alumina and for inherent carbonate contamination in the NaOH used dur-
MEASUREMENT O F SURFACE CHARGE
227
ing potentiometric titration of kaolinite and corundum. Suspensions were equilibrated overnight in a nitrogen atmosphereand pH readings were taken within 5 minutes of base additions.At the end ofeach titration (when pH was higher than 9), an aliquot of suspension was filtered and analyzed for dissolved aluminum. Using charge balance and dissociation constants for the aqueous aluminum and carbonate systems, Carroll-Webb and Walther (1988) subtracted the H+ and OH- consumed by these reactions from total H+ and OH- removed from solution during the titrations, assuming that dissolved aluminum was constant throughout a particular titration. Rather than one intersection point (PZSE) for kaolinite for titrations of three ionic strengths, Carroll-Webband Walther (1988)reported three areas of intersection, possibly reflecting the mineralogical structure of kaolinite. Although this explicit technique is an attempt at quantitatively accounting for competitive reactions that consume H+ or OH-, there are still problems. Dissolution of kaolinite is pH dependent (Carroll-Webband Walther, 1988) and incongruent, so silica and aluminum in solution would vary at different points in the titration. Reprecipitation of aluminum species as amorphous aluminum hydroxide, hydrolysis of soluble silica and aluminum, and possible adsorption of negative hydrolyzed aluminum ions by the surface can consume or release H+ (Riese, 1982).Only hydrolysis of soluble aluminum was accounted for by Carroll-Webb and Walther (1988). The inherent carbonate contamination in NaOH considered by CarrollWebb and Walther (1988) was based on the work by Zeltner (1 986) and Zeltner and Anderson (1 988). Zeltner reported that carbonate contamination decreased the ZPC determined by potentiometric titration, but had no effect on the zero zeta potential. He attributed these results to C02becoming an inherent part of the ferric hydroxide structure as described by Russell et al. (1975). Rather than adsorbing on the particle surface, the CO, fills a structural “trough” and redistributes the charge within the crystal, adding to the intrinsic negative charge and thus decreasing the ZPC. Scholtz et al. (1985) also showed that increased carbonate adsorption by amorphous aluminum hydroxide lowered the ZPC, although they postulated that the carbonate was surface adsorbed. 2. Specific Adsorption Effects
The interpretation of ZPC becomes complicated when titrating in the presence of specifically adsorbed ions. These ions competewith H+and OHfor surface sites and contribute to the surface charge. Researchers have differed in interpreting the influence of specific adsorbing ions on the PZSE (or common intersection point, CIP). Arnold (1978) reported that specifically adsorbed cations would cause a decrease in the ZPC and an increase in
228
ANNE LEWIS-RUSS
the IEP. Parfitt (1 980) attributed the same effect to anion adsorption; that is, specificallyadsorbed anions would result in a decrease in the CIP. According to Parker et al. (1979), the effect of an ionic species in solution on the apparent ZPC differs depending on whether the ion behaves as a readily diffusible counterion or as a PDI. If a cation is strongly adsorbed (specific adsorption), it adds its charge to the surface and requires more OH- for ZPC to be achieved, so the apparent ZPC is higher. Singh and Uehara (1986) referred to this type of adsorption as high affinity specific adsorption or chemisorption. If a cation is exchangeable under the experimental conditions, more H+ is required to achieve the pristine point of zero charge (PPZC) because some H+ is consumed in the exchange reaction and the apparent ZPC is lower. Singh and Uehara (1986) referred to this type of adsorption as low affinity specific adsorption or physical adsorption. The opposite effects occur in the presence of specifically adsorbed or exchangeable anions. In either instance, a CIP still is present, but is displaced from the ZPC (Lyklema, 1984). The effect of specific adsorption is enhanced by divalent or trivalent ions (Lyklema, 1984) and by changes in pH. Breeuwsma and Lyklema (1973) presented titration curves for hematite in Ca(NO,), and KCl that a p proached each other as pH decreased. As pH decreased, the surface became more positive and the adsorption preference of Ca2+over K+ no longer was meaningful. A one-dimensional schematic of the changes in PZSE of a hypothetical material (PPZC = 5) in varying conditions of specifically adsorbed and exchangeable ions is shown in Fig. 4. Pyman et al. (1979a) demonstrated that the opposite effect results when the specificallyadsorbed anion is the only anion present in the electrolyte. In this instance, the crossover point is at a higher pH than the PPZC, and does not coincide with the ZPC or PZSE. This is an artifact of the increasing adsorption of the specifically adsorbed ion as ionic strength increases. Greater anion adsorption results in more H+ being consumed to neutralize the negative charge added to the surface. Because this effect varies with ionic strength, at some point the negatively charged ends of the titration curves will cross; therefore, as the surface charge becomes more negative, even though the ZPC is at an acidic pH, a crossover occurs at a basic pH. Specific adsorption also will affect electrophoretic mobility. McKenzie ( 1983) reported that the mobility of goethite in the presence of molybdenum was dominated by adsorption of the anion. Mobility was most negative at pH 4 where adsorption was greatest. The negative charge of the molybdateanion added directly to the surface charge. In a study of sulfate, selenite, and phosphate adsorption on goethite, Hansmann and Anderson (1985) reported that, although electrophoreticmobility seemed to be a complex function of pH and anion concentration, the free energy of adsorption was the
MEASUREMENT OF SURFACE CHARGE
- --
+
PPZC
3
229
7 PH
5
Add specifically adsorbed cation
Add specifically adsorbed anion
-_ -
--+-+ 3
7 PH
Add Exchangeable cation
Add Exchangeable anion
--
+
-- --
3 7 PH Fig. 4. Schematic of zero point of charge (ZPC)changes in varying conditionsas measured by potentiometric titration. (*) indicates hypothetical pH of new ZPC due to presence of indicated ion. Pristine point of zero charge (PPZC) = 5.
predominant factor. The strongly adsorbed phosphate anion caused a decrease in mobility and a narrowing of the mobility range. According to Sposito ( 198I), PZSE is equivalentto ZPC in the presence of inner-sphere complexes only ifthe charge due to specificadsorption does not change as ionic strength changes. This condition can be met by using a swamping, indifferent, background electrolyte to change ionic strength. Sposito (1 98 1) cautioned that the PZSE thus measured will be shifted from the ZPC -shifted upward in the presence of a specificallyadsorbed cation or shifted downward in the presence of a specifically adsorbed anion. 3. Ion Exchange Effects
Ion exchange effects can be intentional, as in the ion-saturation technique (Section VI,A), or a result of a measurement technique. Various ionic strengths can be used in all the surface charge techniques discussed, are essential for potentiometric titration and salt titration, and are generally used for mineral addition. As ionic strength is increased,the tendency for ion exchange also is increased. This exchange can be ion exchange at permanently charged sites or low affinity specific adsorption at variably charged
230
ANNE LEWIS-RUSS
sites. Ion exchange can also cause variations in measured surface charge due to irreversible exposure of permanently charged sites. Gillman and Bell (1976) reported that aluminum concentration in solution increased as the concentration of NaCl electrolyte was increased during potentiometric titrations of tropical soils. The 1 MNaCl titration curve was displaced to a more acidic value for many of the soils that Gillman and Bell (1976) studied. Madrid et al. (1984) also reported changes in ZPC with changes in ionic strength during studies of a t e . As the ionic strength of the NaCl test solutions was increased, the permanent charge measured also increased. The removal of an aluminum oxide coating could account for the larger permanent charge. The coatingcould block exchange sites, preventing ion exchange; but at a larger ionic strength, the replacement of AIMby Na+ becomes more probable and increased permanent charge would be measured. Even when the electrolyte solution is composed of “indifferent” ions, variations in the ZPC values of oxides as ionic strength varies have been recorded. Some examples were cited in Section IV,E. Only one example of this effect was found for potentiometric titration with an indifferent electrolyte (Breeuwsmaand Lyklema, 1971). One difference between potentiometric titration and mineral addition is the order of addition. In potentiometric titration, the mineral sample is preequilibratedwith the electrolyte solution before acid or base is added, so exchange reactions with ions would be more complete and more uniform. In mineral addition, adding the mineral sample to the pH-adjusted solution may result in more competition for sites between cations and H+, especially as ionic strength increases. 4. Equilibration
Perhaps the best overall measure of the effectsof the potentiometric titration technique is the change with time of the PZSE. The purpose of potentiometric titration is to measure adsorption of H+ and OH- on the surface. Because this adsorption is very fast (Sasaki et al., 1983), changes of results with time may indicate competing reactions, H+difision (Casey et at., 1988),or other changesto the surface, to the solution, or both. Some of these reactions (for sample, dissolution or precipitation)can be very slow and may last for days or weeks (Ahmed and Maksimov, 1969).Conversely,stability of PZSE with time may indicate a system at equilibrium. Breeuwsma and Lyklema (1971) compared titrations of chemically precipitated hematite in which acid/base additions were made at 2- to 5-minute intervals or at 20-minute intervals. Similar titration curves were obtained, but the titrations were not completely reversible. The maximum hysteresis occurred at about pH 7. Parker et al. (1979) compared batch titrations of
MEASUREMENT OF SURFACE CHARGE
23 1
Na-saturated soils that were equilibrated for 1 hour and for 2 days. For all soils, PZSE was higher after longer equilibration by 30 - 70% (average increase of 47%). This higher PZSE indicated a change in the soil-solution system with time. The Na-saturated soils may become Na+-A13+systemsin which A13+behaves as a PDI. More H+/Na+exchange resulted, as compared to H+/A13+(or H+/Ca2+),in the shorter titration time tested, because of the more specific adsorption of A13+ (or Ca2+). Hendershot ( 1978) compared titration techniques for determiningsurface charge of soils with permanently charged components. For NaC1-saturated samples that were dominated by permanent charge, the PZSE (when definable) was higher for soils equilibrated for 3 days than those that were equilibrated for 2 minutes. Soils dominated by variable charge had smaller differences or a decrease in PZSE with longer equilibration time. For all soils, permanent charge (measured as displacement of the crossover point below the ZPT) increased with longer titration time. Hendershot (1 978) credited the time effects to incorporation of H+ and OH- ions into the solid phase. Proton adsorption would account for an increase in measured PZSE with time. The decrease in PZSE with time for soils dominated by variable charge may indicate increased consumption of H+ by competing reactions, such as dissolution and hydrolysis of AP+ and Fe3+.The increase in measured permanent charge may be due to unblocking of exchange sites.
VII. APPLICATIONS AND PREDICTIONS FOR COMPOSITE MATERIALS When surface properties of a mineral or soil are described,ZPC, PPZC, or IEP is assigned a single value; however, unless the material is pure and has a single type of site, the ZPC, at best, represents a mean value for a distribution of charged sites. Even pure minerals can have heterogeneous surface sites due to the presence of different crystal faces and surface imperfections. Certain sample pretreatments, such as grinding, may increase this variability. In a modeling study, van Riemsdijk et al. (1986) reported that the shape of the titration curves was insensitiveto the degree of site heterogeneity, but that the PPZC could be significantly affected. In complex materials (or composite minerals),the ZPC is a function of the various sites present and their abilityto dominate the surface.At the ZPC of a composite oxide, the surface potential for each component is not necessarily zero as it is for a simple oxide (Kuo and Yen, 1988). Madrid et al. ( 1984) examined charge properties of mechanical mixtures of iron oxide (lepidocrocite) and illite. For pure illite and 10% :90% lepidocrocite :illite, there
232
ANNE LEWIS-RUSS
were several crossover points. For 50%or more lepidocrocite, PZSE values were very similar. Either lepidocrocite dominated the mixtures or surface coatings of an oxide on the illite (possiblyaluminum oxide)were suppressing the surface activity of illite. Carroll-Webb and Walther ( 1988) reported three areas of crossover (PZSE) for potentiometric titrations of kaolinite. These areas of crossover were at pH 4.0-4.5, pH 5-6, and pH 7.8-8.5. Carroll-Webb and Walther (1988) suggested that the low and high crossovers represented the ZPC for the SiO, and A1203parts of the minerals, respectively; the middle crossover resulted from the combined behavior of the two layers that comprise kaolinite. This interpretation is substantiated by the dissolution data. Kaolinite dissolution is pH-dependent and incongruent. At pH lower than 6.0, total silica in solution ([Si]) is less than total aluminum in solution ([All), and at pH 6.0-9.3, [Si] is greater than [All (Carroll-Webb and Walther, 1988). Also, rates for short-term dissolution (less than 200 hours) are 0.5 orders of magnitude greater than for long-term steady-state rates (Carroll-Webb and Walther, 1988). From experimental details provided by Carroll-Webband Walther (1988),short-term rates of dissolution would apply to their potentiometric titration of kaolinite. At lower pH values, SiO, surfaces would be expected to dominate the solid phase because proportionately more aluminum is in solution. The opposite would be true for pH values higher than 6.0. The middle crossover is close to the pH value between the two dissolution incongruities, at which dissolution rates would be at a minimum. For minerals composed of more than one oxide or for geologic materials composed of more than one mineral, a first approximation of the overall ZPC is the weighted sum of the PPZCs of the components, assuming each type of surface site behaves independently (Parks, 1967). As Parks (1967) reported, this first approximation is a simplification because of incongruent dissolution of minerals, variations in chemical composition of minerals, particularly hydrous precipitates, and the tendency of some minerals to break along well-developed cleavages, thus favoring particular surfaces. Even so, comparisons of predicted values with measured ZPC indicated reasonable agreement (as shown in Parks, 1967, Fig. 13). Parks’s ( 1967)theory was developed for individual minerals composed of more than one oxide group or for coprecipitates. His model has been tested by analyzing coprecipitated mixtures, mechanical mixtures, and soils. Perrott ( 1977) tested coprecipitated mixtures of amorphous aluminosilicates. His values for PZNC were similar to values expected from compositional dependence. Two allophanic soils also tested were within one-half a pH unit of expected ZPC values based on their A1 :(Si Al) composition. Perrott ( 1977) accounted for the change from tetrahedral to octahedral site coordination for aluminum with increasing Al :(Si Al) ratio for his test
+
+
MEASUREMENT OF SURFACE CHARGE
233
materials, but Pyman et al. (1979b), in their analysis of Perrott’s data, determined that these mineralogical considerations were unnecessary. Schwarz et al. (1984) tested ZPC values for a series of seven silicaaluminum oxide catalytic supports. Their results indicated linear variation by weight fraction of the pure components. The largest deviation they reported from their simple model was about 0.2 pH units. Their materialpreparation procedures were quite rigorous, consisting in part of gel formation, autoclaving, and calciningsamples at 500°C. According to Pyman and Posner (1978), calcined samples are rarely microporous and have larger nitrogen surface areas than water surface areas. Such samples would have uniformly accessible surfaces. In addition, sample preparation probably resulted in aluminum that had a uniform site coordination, because the site coordination is dependent on conditions of precipitation and subsequent sample history (Parks, 1967). Tschapek et af.(1974)tested mechanical mixtures of SiO, and Al,03.The ZPCs of the mixtures seemed to be dominated by the A1203 component rather than conforming to Parks’s (1967) model. However, when Pyman et al. (1979b) replotted the data using mole percent rather than the original weight percent, actual points were similar to values predicted using the ZPC values for the end members, approximated as SiO,.H,O and Al(OH), H20.Pyman et af.(1979b) also were able to predict ZPC values for their own mechanical mixtures of silica- aluminum and silica - ferric hydroxides based on the ZPC of end members. The ZPC values of coprecipitated mixtures had a different trend, and had lower ZPC values than those of the mechanical mixtures. A surface coating of adsorbed silica on aluminum or ferric hydroxide would cause such a decrease of ZPC values for the coprecipitated mixtures. Jepson et af. (1976) reported that zero zeta potentials (measured by electrophoretic mobility) of silica-coated gibbsite decreased as Si02adsorption increased, achieving a constant value of about pH 3. Schwertmann and Fechter ( 1982)reported that ZPC values of natural ferrihydritesdecreased as oxalate-extractable silica increased. Mild pretreatment of the ferrihydrites With NaOH to extract part of the silica resulted in higher ZPC values. When soluble silica was added after precipitation of ferric hydroxide, ZPC values were lower than if the silica was added during ferric hydroxide precipitation. Kuo and Yen (1 988) reexamined the data of Tschapek et al. (1974) for mechanical mixtures of SiO, and Al,03 and were able to model it using the formula:
-
7 r =LooA
+ ( 1 -f*)~o,B
(5)
wheref, is the fraction of the total surface area contributed by component oxide A, and oojis charge density on the surface as charge per unit area of
2 34
ANNE LEWIS-RUSS
4
8
8
PH Fig. 5. Variation of sensitivity of surface charge density to pH changes for two oxides in 0.1 M KNO, ZrO, in 0.1 M KNO, ;(El), FeO, in 0.1 M KNO, Potentiometric titration data from Regazzoni el al. (1983).
.(m,
.
oxide i. The formula also is valid i f L is the mass- weight fraction of oxide A and charge density is in terms of charge per unit mass. At the ZPC, a,, the total surface charge density, equals zero. Kuo and Yen (1988) credited the discrepancy between Parks’s (1967) model and the results of Tschapek et al. ( 1974)to the differencein sensitivity of the surface charge densitiesof oxides to changes in pH; that is, the surface charge of some oxides changes much more than others for an equal addition of acid or base (indicated by steeper slope of the surface charge/pH curve, as illustrated in Fig. 5). Such variations of surfacecharge response to pH change cause nonlinear relations between the weighted sum of ZPC values and the percent composition of oxide components (Pyman et al., 1979b). As Kuo and Yen (1988) reported, the most reasonable relation between surface charge and oxide composition would be based on the surface areas of the components, because surface charge density is a surface function. Kuo and Yen (1988) did not see this relation in silica-alumina mixtures, but explained this discrepancy from theory as compensation of the greater sensitivity (and smaller surface area) of alumina, as compared with silica. The prediction of the surface charge of complex materialshas yet to attain quantitative stature. An implication of multiple crossover work (CarrollWebb and Walther, 1988) is that determination of the ZPC for complex materials, such as soil or crushed rock, may be impossible, as the number of cation centers, each with unique charge behavior, increases (J. V. Walther,
235
MEASUREMENT OF SURFACE CHARGE
1989 personal communication). Values for the ZPC of soils have been published (see, e.g., van Raij and Peech, 1972; Espinoza et al., 1975; Laverdi2re and Weaver, 1977),but problems with material dissolution and competitive proton reactions may not have been accounted for. Several investigatorshave reported trends regarding the characteristics of complex materials and their effect on surface charge. According to van Raij and Peech (1972), the presence of iron and aluminum oxides will tend to increase the ZPC, whereas the presence of clay minerals that have negative permanent charge and the presence of organic matter will tend to shift the ZPC to lower values. Studies cited earlier in this section (Jepson et al., 1976; Schwertmann and Fechter, 1982)described the behavior of silica and silica coatingsin loweringthe ZPC. Data of Penott ( 1977)and Gonzales-Batistaet al. ( 1982)indicated that PZNC values increased as Alw: ( S P A13+)ratios increased. The use ofpredictionsof charge behavior based on reported trends, such as those in the preceding paragraph, may be more useful than attempts at experimental testing, especiallyfor systems dominated by silica. Uehara and Gillman ( 1981) have proposed a homogeneous amorphous coating on soil particles, which would dominate any ZPC determination. The low ZPC of silica can only be obtained at the risk of dissolution of test materials. Careful determination of the predominant surface materials may be more meaningful than experimental ZPC testing for mixed materials. A more methodical approach was suggested by Kuo and Yen ( 1988). They described a graphic procedure for predicting ZPC if the PPZC, the chargedensity sensitivities to pH for mineral components, and the proportional mineral composition of the surface are known. Compilations of PPZC values are available [see, e.g., Parks (1965, 1967) for metal oxides and hydroxides, and some silicates; Fuerstenau ( 1970), for quartz and hematite; Kinniburgh and Jackson ( 1981) for hydrous metal oxides]. Charge-density sensitivities can be determined using surface-charge/pH graphs for puremineral phases. Accurate determination of the proportional mineral composition of the surface may be more difficult (Calvert et al., 1990).Petrographic examination and point-countingor scanning electron microscopy with statistical analysis may allow surface mineral identification. Surface chemical composition can be determined using X-ray photoelectron spectroscopy (Siegbahn, 1982; Kelley, 1987). The major component of surface coatings can be determined using sequential dissolution techniques(Chao, 1984).For some sedimentarymaterials, microwave digestion or ultrasonic treatment in deionized water will remove surface coatings (H. K. Long, 1990 personal communication). Solution analysis can then be used to determine chemical constituents. In any determination of ZPC, the primary concern needs to be the appli-
+
236
ANNE LEWIS-RUSS
cation of the results, For industrial applications (e.g., flotation or pigment processes), tests of surface charge can be designed that are similar to working conditions. For soil applications (e.g., fertility determination or trace metal movement), test conditions begin to diverge from field circumstances. For whole-rock applications (e-g., aquifer transport or ore body formation),test conditions diverge greatly from reality. These divergences result from increasingly rigorous material preparation and necessarily inflated water :rock ratios in test conditions. These factors cause systems to be displaced from equilibrium. Surface-chargeexperimentationbecomes exceedinglycomplex as equilibration processes compete with the test situation.
VIII. SUMMARY The behavior of colloids depends on their surface charge. The surface charge results from structural factors inherent in the mineral (permanent charge) or from adsorption of potential determining ions by the mineral (variable charge). Several techniques are available for measuring surface charge. The choice of method will depend, in part, on the suitability of the technique for the material tested. Potentiometric titration has been widely used to determine surface charge, but due to the pH changes involved, this method is most suited to low solubility oxides that have ZPCs in the pH range of 4- 10. The use of a backtitration technique may allow compensation for competitive reactions as the solubility of materials increases. Electrokinetic measurements also are restricted by pH, due to the limitations of the equipment and the responses of the materials tested. Although the interpretation of surface charge may be uncertain, measurement of the isoelectric point is direct and can be determined from measurement of a suspension of a single ionic strength. Streaming potential is less sensitive than potentiometric titrations to small surface features, making this technique less affected by some sample preparation procedures. The ion retention technique uses measurement of cation and anion retention at different pH values to determine the ZPC. The technique, which requires chemical analyses for the speciesof interest, is generallyused for soil and clay measurements. The range of pH values used is limited by the solubility of the tested material. This limitation also applies to the salt titration and mineral addition techniques. Neither of these latter two techniques is widely used, although they require a minimum of equipment. If the value
MEASUREMENT OF SURFACE CHARGE
237
of the ZPC, rather than the surface charge density, is needed, either of these methods could be used as a simple first approximation. Many geological materials, such as soils or altered rocks, contain clays and oxides. The surface charge of these composite materials is comprised of both permanent and variable charge. There have been several attempts to measure these components of charge separately, but no definitive technique has been developed. Measuring the surface charge of geological materials is complex because all measuring techniques involve sample preparation and equilibration of the sample suspension at a range of pH values. Any sample preparation that alters the surface will affect the surface charge. Grinding rocks can result in a disturbed layer with different properties than the bulk rock. Acid-washing and ion-saturation can cause dissolution or ion exchange that alters the surface charge measured. Nor are chemical precipitates free of problems. The ZPC of precipitates may increase or decrease as aging progresses because of changes in surface hydration and crystallinity. Precipitates are also affected by the pH of the aging solution. Geologic materials will react differently in solution as pH, ionic strength, and ionic composition are changed. In addition to variations in dissolution rates with changes in pH, other reactions occur that compete with the surface for protons and hydroxide. Some examples of competing reactions are: dissolution, precipitation,formation of dissolved metal complexes, proton carbonate reactions, and proton/cation exchange. Contributions of some of these reactions can be accounted for by using a backtitration technique for potentiometrictitration, and some can be explicitly accounted for if solution parameters are well known. As the concentration of any ion in solution increases, the tendency for exchange and adsorption also increases, causing competition for surface sites. Exchangeable ions and specifically adsorbed ions can cause shifts in the measured ZPC. For complex geological materials, the overall ZPC is a composite value that depends on the various surface sites present and their ability to dominate surface reactions. A first approximation of a composite ZPC is the weighted sum of the ZPCs of the components. This simplified approach can be refined by considering the surface area rather than the weight fraction, and accounting for the varying surface charge/pH sensitivitiesof each component. The sensitivity of composite geological materials to the surface charge measuring procedures makes a predictive scheme for ZPC desirable. An important element in such a scheme is accurate identification of the minerals and mineral coatings that dominate the exposed surfaces. Prediction could be especially valuable because most testing conditions for ZPC diverge greatly from the field conditions.
238
ANNE LEWIS-RUSS
ACKNOWLEDGMENTS The comments and advice of the following reviewers are gratefully acknowledgd Cristian Schulthess, Kathleen Smith, Maynard Slaughter, Ken Stollenwerk, and James Ranville.
REFERENCES Ahmed, S. M., and Maksimov, D. 1969. Studies of the double layer on cassiterite and rutile. J. Colloid Intet$ace Sci.29,97- 104. Arnold, P. W. 1978. Surface-electrolyteinteractions. In “The Chemistry of Soil Constituents Greenland” (D. J. and M.H. B. Hayes, eds.),pp. 355-448. Wiley (Interscience),Clichester, England. Barber, R. G., and Rowell, D. L. 1972. Charge distribution and the cation exchange capacity of an iron-rich kaolinitic soil. J. Soil Sci. 23, 135- 146. Barrow, N. J. 1985. Reactions of anions and cations with variablecharge soils. Adv. Agron. 38, 183-230.
Barrow, N. J., and Ellis,A. S. 1986. Testing a mechanistic model-Part V. The points of zero salt effect for phosphate retention, for zinc retention and for acid/alkali titration of a soil.J. Soil Sci. 37, 303 - 3 10. Bolan, N. S.,Syers, J. K., and Tillman, R. W. 1986. Ionic strength effects on surface charge and adsorption of phosphate and sulphate by soils. J. Soil Sci. 37,379-388. Bolt, G. H. 1957. Determination of the charge density of silica sols. J. Phys. Chem. 61,11661169.
Bowden, J. W., Nagarajah, S., Barrow, N. J., Posner, A. M., and Quirk, J. P. 1980a. Describing the adsorption of phosphate, citrate and selenite on a variablecharge mineral surface.Aust. J. Soil Res. 18,49-60. Bowden, J. W., Posner, A. M., and Quirk, J. P. 1980b. Adsorption and charging phenomena in variable charge soils. In “Soils with Variable Charge” (B. K. G. Theng, ed.), pp. 147- 166. New Zealand Soc. Soil Sci., Lower Hutt. Bradfield,R. 1923. The nature of the acidity ofthe colloidalclay of acid soils.J.Am. Chem. SOC. 45,2669-2678.
Breeuwsma, A., and Lyklema, J. 1971. Interfacial electrochemistry of haematite (a-Fe,O,). Discuss. Faraday Soc.52,324-333. Breeuwsma,A., and Lyklema,J. 1973. Physicaland chemicaladsorption of ions in the electrical double layer of hematite (a-Fe203).J. Colloid Jntqhce Sci. 43,437 -448. Bryant, R., and Williams, D. J. A. 1987. The electrochemistry of colloidal particles from a proglacial lake. Chem. Geol. 62,29 1 - 305. Calved, C. S., Pallcowsky, D. A., and Pevear, D. R. 1990. A combined X-ray powder diffraction and chemical method for the quantitative mineral analysisof geologic samples. In “Quantitative Mineral Analysis of Clays” (C.S. Pevear and F. A. Mumpton, eds.),CMS (Clay Mineral Society) Workshop Lectures,Vol. 1, pp. 153- 166. Clay Miner. Soc., Evergreen, Colorado. Carroll-Webb, S. A., and Walther, 3. V. 1988. A surface complex reaction model for the pHdependence of corundum and kaolinitedissolution rates. Geochim. Cosmochim. Acta 52,2609-2623.
Casey, W. H.,Westrich, H. R., and Arnold, G. W. 1988. Surface chemistry of laboradorite
MEASUREMENT O F SURFACE CHARGE
239
feldspar reacted with aqueous solutions at pH = 2,3,and 12.Geochim. Cosmochim. Acta 52,2795-2807. Chao, T. T. 1984.Use of partial dissolution techniques in geochemical exploration. J. Geochem. Explor. 20, 101 -- 135. Charlet, L., and Sposito, G. 1987.Monovalent ion adsorption by an oxisol. SoilSci. SOC.Am. J. 51, 1155-1160. Davis, J. A., James, R. O., and Leckie, J. 0. 1978.Surface ionization and complexation at the oxide/water interface. J. Colloid Interface Sci. 63,480-499. Dzombak, D. A. 1986.Toward a uniform model for the sorption of inorganic ions on hydrous oxides. Ph.D. Thesis, Massachusetts Inst. Technol., Cambridge. Espinoza, W. G., Gast, R. G.,and Adams, R. S., Jr. 1975.Charge characteristics and nitrate retention by two Andepts from south-central Chile. Soil Sci. Soc.Am. Proc. 39,842-846. Fuerstenau, D. W. 1970.Interfacial processes in mineral/water systems. Pure AppI. Chem. 24, 135- 164. Fuerstenau, D. W., Manmohan, D., and Raghavan, S. I981. The adsorption of alkaline-earth metal ions at the rutile/aqueous solution interface. In “Adsorption from Aqueous Solutions” (P. H. Tewari, ed.), pp. 93- 1 17.Plenum, New York. Gillman, G.P., and Bell, L. C. 1976.Surface charge characteristicsof six weathered soils from tropical North Queensland. Aust. J. Soil Res. 14,351- 360. Gillman, G. P., and Uehara, G. 1980.Charge characteristicsof soils with variable and permanent charge minerals-Part 11. Experimental. Soil Sci. Soc. Am. J. 44,252-255. Gonzales-Batista, A., Hernandez-Moreno, J. M.,Fernandez-Caldas, E., and Herbdon, A. J. 1982.Influence of silica content on the surface charge characteristicsof allophanic clays. Clays Clay Miner. 30,103- 110. Hansmann, D. D., and Anderson, M. A. 1985. Using electrophoresis in modeling sulfate, selenite, and phosphate adsorption onto goethite. Environ. Sci. Technol. 19, 544-55 1. Healy, T. W., and Fuerstenau, D. W. 1965.The oxide-water interface-Interrelation of the zero point of charge and the heat of immersion. J. Colloid Inteflace Sci. 20, 376-386. Healy, T. W., Hemng, A. P., and Fuerstenau, D. W. 1966.The effect ofcrystal structure on the surface properties of a series of manganese dioxides. J. Colloid InterfaceSci. 21,435-444. Healy, T. W., James, R. O., and Cooper, R. 1968.The adsorption of aqueous Co(I1) at the silica-water interface. In “Adsorption from Aqueous Solution” (W. J. Weber, Jr. and E. Matijevib, eds.),Advances in Chemistry Series, No. 79,pp. 62-73. Am. Chem. Soc., Washington, D.C. Helgeson, H. C., Murphy, W. M., and Aagaard, P. 1984. Thermodynamic and kinetic constraints on reaction rates among minerals and aqueous solutions. Part 11. Rate constants, effective surface area, and the hydrolysis of feldspar. Geochim. Cosmochim. Acta 48, 2405-2432. Hendershot, W. H. 1978.Measurement technique effects of the value of zero point of charge and its displacement from zero point of titration. Can. J. Soil Sci. 58,439-442. Hohl, H., Sigg, L., and Stumm, W. 1980. Characterization of surface chemical properties of oxides in natural waters, the role of specific adsorption in determining the surface charge. In “Particulates in Water, Characterization, Fate, Effects, and Removal” (M. C. Kavanaugh and J. 0.Leckie, eds.),Advances in Chemistry Series,No. 189,pp. 1 - 3 1. Am. Chem. Soc.,Washington, D.C. Holdren, G.R., Jr., and Berner, R. A. 1979. Mechanism of feldspar weathering-Part I. Experimental studies. Geochim. Cosmochim. Acta 43, 1 16 I - 1 171. Houchin, M. R., and Warren, L. J. 1984.Surfacetitrations and electrokineticmeasurements on stannic oxide suspensions. J. Colloid Intedace Sci. 100,278-286.
240
ANNE LEWIS-RUSS
Hsi, C. K. D., and Langmuir,D. 1985. Adsorption of uranyl (VI)onto ferric oxyhydroxides: application of the surface complexation site-bindingmodel. Geochim. Cosmochim. Acta 49, 1931-1941. Huang, C. P. 1981. The surface acidity of hydrous solids. In “Adsorption of Inorganics at Solid-Liquid Interfaces” (M. A. Anderson and A. J. Rubin, eds.), pp. 183-217. Ann Arbor Sci., Butterworth Group, Ann Arbor, Michigan. Hunter, C. R., and Busacca, A. J. 1987. Pedogenesis and surface charge of some Andic soils in Washington, U.S.A. Geoderma 39,249-265. Hunter, R.J. 1981. “Zeta Potential in Colloid Science--Principles and Applications.” Academic Press, London. Husain, A. 1984. Charge development at the uranium oxide-solution interface. J. Colloid Interface Sci. 102,389-399. James, R. 0. 1981. Surface ionization and complexation at the colloid/aqueous electrolyte interface. In “Adsorption of Inorganics at Solid-Liquid Interfaces’’ (M. A. Anderson and A. J. Rubin, eds.),pp. 2 19- 26 1. Ann Arbor Sci., Butterworth Group, Ann Arbor, Michigan. James, R. O., and Parks, G. A. 1982. Characterization of aqueous colloids by their electrical double-layer and intrinsic surface chemistry properties. In “Surface and Colloid Science” (E. Matijevii, ed.), Vol. 12, pp. 119-216. Plenum, New York. James, R. O., Stiglich, P. J., and Healy, T. W. 1981. The TiOJaqueous electrolyte systemApplications of colloid models and model colloids. In “Adsorption from Aqueous Solutions” (P. H. Tewari, ed.),pp. 19-40. Plenum, New York. Jepson, W. B., Jeffs, D. G., and Ferris, A. P. 1976. The adsorption of silica on gibbsite and its relevance to the kaolinite surface. J. Colloid Interjace Sci. 55,454-46 1. Kelley, M. J. 1987. Surface-sensitiveanalytical techniques, Part 2. CHEMTECH 17,98- 105. Kinniburgh, D. G., and Jackson, M. L. 1981. Cation adsorption by hydrous metal oxides and clays. In “Adsorption of Inorganicsat Solid- Liquid Interfaces” (M. A. Anderson and A. J. Rubin, eds.), pp. 91 - 160. Ann Arbor Sci., Butterworth Group, Ann Arbor, Michigan. Kuo,J. F. 1987. Interface electrochemical properties of composite oxides in aqueous solution. Ph.D. Thesis, Univ. of Southern California, Los Angeles. Kuo, J. F., and Yen, T. F. 1986. Streaming potential measurements on mixed oxide systems, Proc. IECEC, San Diego, Calg, 2Jst 3,25 1-256. Am. Chem. Soc., Washington, D.C. Kuo, J. F., and Yen, T. F. 1988. Some aspects in predicting the point of zero charge of a composite oxide system. J. Colloid Interjace Sci. 121,220-225. Kuo,J. F., Sharma, M. M., and Yen, T. F. 1988. Electrokineticbehavior ofporouscomposite oxide matrix. J. Colloid Interface Sci. 126,537-546. Langmuir, D. 1979. Techniques of estimating thermodynamic properties for some aqueous complexes of geochemical interest. In “Chemical Modeling in Aqueous Systems, Speciation, Sorption, Solubilityand Kinetics” (E.A. Jenne, ed.), ACS SymposiumSeries,No. 93, pp. 353-387. Am. Chem. Soc., Washington, D.C. Langmuir, D., and Mahoney, J. 1985. Chemical equilibrium and kinetics of geochemical processes in ground water studies. Pract. Appl. Ground Water Geochem.. Proc. Can./Am. Conj Hydrogeol., Jst (B. Hitchon and E. I. Wallick, eds.), pp. 69-95. Nat. Water Well Assoc., Worthington, Ohio. Laverdikre,M. R., and Weaver, R. M. 1977. Charge characteristicsof spodic horizons. SoilSci. Soc. Am. J. 41,505-510. Lyklema, J. 1984. Points of zero charge in the presence of specific adsorption. J. Colloid Interface Sci. 99, 109- 1 17. Madrid, L., Dim, E., Cabrera, F., and de Arambani, P. 1983. Use of a three-plane model to describe charge properties of some iron oxides and soil clays. J. Soil Sci. 34,57-67.
MEASUREMENT OF SURFACE CHARGE
24 1
Madrid, L., Dim, E., and Cabrera, F. 1984. Charge properties of mixtures of minerals with variable and constant surface charge. J. Soil Sci. 35,373-380. McKenzie, R. M. 1983. The adsorption ofmolybdenum on oxide surfaces.Aust. J.SoilRes. 21, 505-513. Morel, F. M. M., Westall, J. C., and Yeasted, J. G. 1981. Adsorption models: A mathematical analysis in the framework of general equilibrium calculations.In “Adsorption of Inorganics at Solid- Liquid Interfaces” (M. A. Anderson and A. J. Rubin, eds.),pp. 263 -294. Ann Arbor Sci., Buttenvorth Group, Ann Arbor, Michigan. Murray, J. W. 1974. The surface chemistry of hydrous manganese dioxide. J. Colloid Interface Sci. 46,357-371. Nail, S. L., White, J. L., and Hem, S. L. 1976. Structure of aluminum hydroxide gel, Part 11: Aging mechanism. J. Pharm. Sci. 65, 1 192- 1 195. Parf~tt,R. L. 1980.Chemical properties of variable charge soils. In “Soils with Variable Charge” (B. K. G. Theng, ed.),pp. 167- 194. New Zealand Soc.Soil Sci., Lower Hutt. Parker, J. C., Zelazny, L. W., Sampath, S.,and Harris, W. G. 1979. A critical evaluation ofthe theory to soil systems. Soil Sci. SOC.Am. J. 43, extension of zero point of charge (ZPC) 668-674. Parks, G. A. 1965.The isoelectricpoints of solid oxides, solid hydroxides, and aqueous hydroxo complex systems. Chem. Rev. 65, 177- 198. Parks, G. A. 1967. Aqueous surface chemistry of oxides and complex oxide mineralsIsoelectric point and zero point of charge. In “Equilibrium Concepts in Natural Water Systems” (W. Stumm, ed.), Advances in Chemistry Series, No. 67, pp. 121 - 160. Am. Chem. Soc.,Washington, D.C. Parks, G. A., and de Bruyn, P. L. 1962. The zero point of charge of oxides. J. Phys. Chem. 66, 967-973. Perrott, K. W. 1977. Surface charge characteristicsof amorphous aluminosilicates.Clays Clay Miner. 25,417-421. Pyman, M. A. F., and Posner, A. M. 1978. The surface areas of amorphous mixed oxides and their relation to potentiometric titration. J. Colloid Interface Sci. 66,85 -94. Pyman, M. A. F., Bowden, J. W., and Posner, A. M. 1979a.The movement oftitration curvesin the presence of specific adsorption. Aust. J. Soil Res. 17, 191- 195. Pyman, M. A. F., Bowden, J. W., and Posner, A. M. 1979b. The point of zero charge of amorphous coprecipitates of silica with hydrous aluminum of ferric hydroxide. Clay Miner. 14, 87-92. Rand, B., and Melton, I. E. 1975. Isoelectric point of the edge surface of kaolinite. Nature (London) 257,214-216. Regazzoni, A. E., Blesa, M. A., and Maroto, A. J. G. 1983. Interfacial properties of zirconium dioxide and magnetite in water. J. Colloid Interface Sci. 91,560-570. Riese, A. C. 1982. Adsorption of radium and thorium onto quartz and kaolinite: A comparison of solution/surface equilibria models. Ph.D. Thesis, Colorado Sch. Mines, Golden. Robinson, B. P. 1962. Ion-exchange minerals and disposal of radioactive wastes-A survey of literature. Geol. Sun. Water-SupplyPap. (US.)No. 1616. Russell, J. D., Paterson, E., Fraser, A. R., and Farmer, V. C. 1975.Adsorption ofcarbon dioxide on goethite (a-FeOOH) surfaces, and its implications for anion adsorption. J. C. S. Faraday Z71, 1623- 1630. Sasaki, M., Morlya, M., and Yasunaga, T. 1983. A kinetic study of ion-pair formation in the surface of a-FeOOH in aqueous suspensions using the electric field pulse technique. J. Phys. Chem. 87, 1449- 1453. Schofield, R. K. 1949. Effect of pH on electric charges carried by clay particles. J. Soil Sci. 1, 1-8.
242
ANNE LEWIS-RUSS
Scholtz, E. C., Feldkamp, J. R., White, J. L., and Hem, S. L. 1985. Point of zero charge of amorphous aluminum hydroxide as a function of adsorbed carbonate. J. Pharm. Sci. 74, 418 -48 I . Schulthess,C. P., and Huang, C. P. 1990. Adsorption of heavy metals by silicon and aluminum oxide surfaces on clay minerals. Soil Sci. Soc. Am. J. 54,619-688. Schulthess, C. P., and Sparks, D. L. 1986. Backtitration technique for proton isotherm modeling of oxide surfaces. Soil Sci. SOC.Am. J. 50, 1406- 1411. Schulthess, C. P., and Sparks,D. L. 1987. Two-site model for aluminum oxide with mass balanced competitive pH &/salt dependent reactions. Soil Sci. Soc. Am. J. 51,1136-
+
1144.
Schulthess, C. P., and Sparks, D. L. 1988. A critical assessment of surface adsorption models. Soil Sci. Soc. Am. J. 52,92-97. Schulthess, C. P., and Sparks, D. L. 1989. Competitive ion exchange behavior on oxides. Soil Sci. Sac. Am. J. 53,366-373. Schwarz, J. A., Driscoll, C. T.,and Bhanot, A. K. 1984. The zero point of charge of silica-alumina oxide suspensions. J. Colloid Interface Sci. 97,55 -6 1. Schwertmann, U., and Fechter, H. 1982. The point of zero charge of natural and synthetic ferrihydritesand its relation to adsorbed silicate. Clay Miner. 17,471 -476. Shanna, M.M.,Kuo, J. F., and Yen, T. F. 1987. Further investigation of the surface charge properties of oxide surfaces in oil-bearing sands and sandstones. J. Colloid Interface Sci. 115,9- 16.
Siegbahn, K. 1982. Electron spectroscopyfor atoms, molecules, and condensed matter. Science 217, 111-121.
Singh, U., and Uehara, G. 1986. Electrochemistry of the doublelayer, principles and applications to soils. In “Soil Physical Chemistry” (D. L. Sparks, ed.), p. 1-38. CRC Press, Boca Raton, Florida. Smit, W., and Holten, C. L. M.198O.Zeta-potentialand radiotracer adsorption measurements on EFG a-Al,03single crystals in NaBr solutions. J. Colloid Inte$ace Sci. 78, 1- 14. Somasundaran, P., and Agar, G. E. 1967. The zero point ofcharge ofcalcite.J. Colloid Interface Sci. 24,433-440. Sposito, G. 198 1. The operational definition of the zero point of charge in soils. Soil Sci. Soc. Am. J. 45,292 - 291. Sposito, G. 1984. “The Surface Chemistry of Soils.” Oxford Univ. Press,New York. Stumm, W., and Wollast, R. 1990. Coordination chemistry of weathering: kinetics of the surface-controlled dissolution of oxide minerals. Rev. Geophys. 53,53-69. Swallow, K. C., Hume, D. N., and Morel,F. M.M. 1980. Sorption of copper and lead by hydrous femc oxide. Environ. Sci. Technol. 14, 1326- 1331. Torres Sanchez, R. M., Agletti, E. F., and Port0 Lopez, J. M. 1988. PZC modification on mechanochemically treated kaolinite. Muter. Chem. Phys. 20,27-38. Tschapek, M.,Tcheichvili, L., and Wasowski, C. 1974. The point of zero charge (pzc) of kaolinite and SiO, Al,O, mixtures. Clay Miner. 10,219-229. Uehara, G., and Gillman, G. P. 1980. Charge characteristicsof soils with variable and permanent charge minerals-Part I. Theory. SoilSci. Soc. Am. J. 44,250-252. Uehara, G., and Gillman, G. P. 1981. “The Mineralogy, Chemistry, and Physics of Tropical Soils with Variable Charge Clays.” Westview Press, Boulder, Colorado. van Olphen, H. 1963. “An Introduction to Clay Colloid Chemistry.” Wiley (Interscience),New York. van Raij, B., and Peech, M. 1972. Electrochemical properties of some oxisols and alfisols of the tropics. Soil Sci. Soc. Am. Proc. 36,587-593. van Riemsdijk, W. H., Bolt, G. H., Koopal, L. K., and Blaakmeer, J. 1986. Electrolyte adsorp-
+
MEASUREMENT OF SURFACE CHARGE
24 3
tion on heterogeneous surfaces- Adsorption models. J. Colloid InterfaceSci. 109,2 19 228.
Wada, K., and Okamura, Y. 1983. Net charge characteristicsof Dystrandept Band theoretical prediction. Soil Sci. Soc. Am. J. 47,902 - 905. Wayman, C. H. 1967. Adsorption on clay mineral surfaces. Princ. Appl. Water Chern..Proc. RudolfsRes. Con$, 4th, Rutgers Univ.( S . D. Faust and J. V. Hunter, eds.), pp. 121- 164. Wiley, New York. Westall, J. C., and Hohl, H. 1980. A comparison of electrostaticmodels for the oxide/solution interface. Adv. Colloid InterfaceSci. 12,265 - 294. White, G . N., and Zelazny, L. W. 1986. Charge properties of soil colloids. In “Soil Physical Chemistry” (D. L. Sparks, ed.),pp. 39-81. CRC Press,Boca Raton, Florida. Williams, D. J. A., and Williams, K. P. 1978. Electrophoresis and zeta point of kaolinite. J. Colloid Interface Sci. 65,19 - 81. Zeltner, W. A. 1986. Charge development at the goethite/water interface: Effects of aggregation and carbonate adsorption.Ph.D. Thesis, Univ. of Wisconsin, Madison. Zeltner, W. A., and Anderson, M. A. 1988. Surface charge development at the goethite/aqueous solution interface: effects of C02 adsorption. Langmuir 4,469-474.
This Page Intentionally Left Blank
GENETIC IMPROVEMENT OF MAIZE YIELDS W. A. Russell Department of Agronomy Iowa State University Ames, Iowa SO01 1
I. 11. 111. IV. V. VI. VII. VIII.
Introduction Genetic Gains in Grain Yield Genetic Gains: Stress versus Nonstress Environments Response to Increase in Plant Density and Nitrogen Fertility Changes in Other Plant Traits Genetic Gains via Recurrent Selection in Populations Improvement of Inbred Lines Future Trends References
I. INTRODUCTION Grain yield increase for maize, Zeu mays L., in the United States occurred in the last 60 years, which has been the period of hybrid maize production. Nonetheless, it was during the first three decades of the Twentieth Century that the basis for hybrid maize was developed. East (1908), Shull(l909), and others began maize breeding for hybrid development in the early 1900s in the United States, although a primitive type of breeding had been conducted for thousands of years by the Native Americans. Pre-Twentieth Century breeding was important because it provided the open-pollinated varieties (O.P.) from which the parental inbred lines of the first double-cross hybrids were developed. These O.P., such as “Reid Yellow Dent” and “Lancaster Sure Crop,” were developed in the 1800s from crosses of the northeastern flints and the southern dents. Although maize gemplasm from other countries has been used in the development of some U.S. Corn Belt parental inbred lines, exotic germplasm has had a relatively insignificantrole in Corn Belt breeding programs (Brown, 1975; Goodman, 1985, 1990). Breeding procedures were used to improve and develop new strains of the 245 Amurncdr in A p n ~ w ~ Vy d , 46 Copyright 0 1991 by Academic Press, Inc. AU rights of reproduction in my form reserved.
246
W. A. RUSSELL
open-pollinatedcultivars in the late 1800s and early 1900s before the development of inbred lines for hybrid seed production was begun. These breeding procedures, which included varietal hybridization, mass selection, and ear-to-row selection, did not effect important yield improvements. In some instances, varietal hybridizationdeveloped crossesthat produced better than the higher-yielding parent, but the procedure was not accepted widely for commercial use. Selectionprogramswere successfulin producing numerous strains that differed for maturity, plant type, ear and grain type, and pest resistance. Maize shows conducted in the early part of the Twentieth Century also led to selection for distinct ear and grain appearance. Close selection to type, however, may have caused some inbreeding, which may have been the primary reason that yield improvementswere not realized. Many of these O.P. were the bases for subsequent maize breeding programs. Mass selection and ear-to-row breeding were graduallydiscontinued as inbred line development for hybrid use became the accepted method. More recently, modifications to mass selection (Gardner, 196 1) and to ear-to-row selection (Lonnquist, 1964) have been used to enhance the effectiveness of these selection methods for improving yield in breeding populations. Most maize breeding research has been conducted in this century. Research with maize inbred lines and hybrids was described in detail by Shull (1908, 1909, 1910), and many of the earlier studies have been reviewed by Hayes (1 963). The pure-line method of maize breeding evolved from Shull’s work, and is still the basic maize-breedingprocedure in the United Statesand in many other parts of the world. Jones’s ( 19 18, 19 19) suggestion for the commercial use of double-crosshybrids was a major contributiontoward the adoption of hybrid maize in the United States (Hayes, 1963). Many U.S. agricultural experiment stations and USDA programs for the development and evaluation of inbred lines began during the period of 19 15 - 1925. Crabb ( 1947) wrote about early maize breeders and their programs. Maize breeding as a private commercial enterprise began in the 1920s, and by the 1930s farmers’ use of hybrid seed became an acceptable practice; hybrid maize occupied approximately 100% of the maize area in Iowa by 1943,90% of the maize area in the U.S. Corn Belt, but only 60% of the maize area for the entire United States (Fig. 1). Double-cross hybrids were the predominant type in the United States until about 1960, when the use of single crosses began to increase. Single crosses became the predominant type within a few years, with considerably fewer hectares being planted to related-line single crosses, three-way crosses, and double crosses. Maize breeding for the development of inbred lines and hybrids in other parts of the world expanded rapidly after World War 11. Maize is a flexible species amenable to selection; consequently, progress has been achieved to develop types adapted to many areas where previously it either was not
GENETIC IMPROVEMENT OF MAIZE YIELDS
247
100
90 $? 80
.-E6
3 0
5
70 60
50 40
10
Years, 1930-1960
Fig. 1. Percentage of maize planted with hybrid seed in Iowa (A), U.S. Corn Belt (0),and the United States (0)for 1930- 1960.
grown or was relatively unimportant in earlier years. For example, there has been the tremendous expansion in some European countries, made possible by selection for earlier maturity and better adapted cultivars that can be grown successfully in areas where maize production was not feasible40 years ago (Trifunovic, 1978). Some inbred line development and evaluation were done in a few European countries before 1940, and a great expansion of maize breeding came after 1945. In the first period of this expansion, materials from the United States were evaluated and some U.S. hybrids were used. Subsequently, European and U.S. lines in hybrid combinations were developed and these hybrids helped to expand the maize-growing area and improve yields. The European lines introduced greater cold tolerance and adaptation for earlier maturity, and the U.S. lines added improved yield and standability. These combinations permitted the expansion of hybrid maize into central Europe. In France, for example, the hectarage planted to maize has increased more than six times from the early 1950s to the late 1980s, and average yields have increased about 4-fold. Indeed, France is now a leading maize producer in Europe. The development of earlier-maturityhybrids has been a prime factor in this increase. In Yugoslavia, which is also a major maize-producing country in Europe, genetic improvement of hybrids and better cultural practices have given a yield increase of 0.94 q ha-' yr-I from 1966 to 1989 (Kojic, 1990). Similarly in Romania, the leading producer of maize grain in Europe, the yield increase has been 1.35 q ha-' yr-' (Kojic, 1990).
World maize production now normally exceeds 440 million metric tons and usually ranks third after wheat and rice in total production among the
W. A. RUSSELL
248
..
75 US. Maize Yield, 1930-1989
65
55 45
35 25 15
..
r2= 0.93
5 1
I
1930
1940
1950
1960
I
1970
1980
1990
Year
Fig. 2. Observed and predicted maize yields in the United States for 1930- 1989. 9 = 10.77 0.62 (X- 1930) 0.00079 (X- 1930)2;9 = 0.93.
+
+
world cereal crops (U.S. Department of Agriculture, 1930- 1989). During the period 1970- 1972 to 1987- 1988, world maize production increased about 172 million metric tons, which represents a 62% increase in the world supplies and a rate of 3.5% yI". More than 50% of the total world area planted with maize is in Latin America, Africa, and Asia, but probably less than 35% of the total world grain maize production is in these areas. Except for a few countries in these three areas, average yields per hectare are very low. Hybrids are the primary type grown in some countries, whereas other countries use O.P.,improved synthetics, variety crosses, and hybrids. As farm technology is improved, it seems likely that hybrid types will become more important in most of the tropical and subtropical areas. Average maize yields in the United States have increased continuously from mid- 1930s to 1989 (Fig. 2). Before 1930, yields were static because no yield gain was realized from breeding, and essentially no improvement occurred because of changes in cultural practices. The yield increases that started after 1935 coincided with the use of hybrids, as shown in Fig. 1. Yield increases have occurred since the mid- 1930s because of the use of hybrids, increased use of fertilizers, more effective weed control, higher plant densities, and improved management. The yield increase trend shows evidence of
GENETIC IMPROVEMENT OF MAIZE YIELDS
249
an increased rate ofgain beginning about 1960,when single crossesstarted to replace double crosses. Moreover, 1960- 1970was a period of rapid increase in the use of nitrogen fertilizer (Thompson, 1982) and in greater plant densities. Mean yields plotted in Fig. 2 show greater variability among years for average yields in the 1970- 1980s, but there is no definite evidence of a plateau effect. Thompson (1979, by using a regression equation derived from maize yields and weather data in Illinois, predicted that yields in that state would stabilize at 75 q ha-’ by 1985. This stabilization has not occurred, however, partly because of the introduction in the mid- 1970sof new hybrids with higher yield potentials. Estimates of 25 - 35%were frequently quoted for the yield improvements of the first hybrids over O.P. Frey (197 1) presented data, based on the Iowa State Corn Yield Tests, showing that the initial increase was at 7- 1 1% during 1926- 1932; but it reached 18% in 1943, when 100% of the Iowa maize was hybrid. By using a system of comparisonsthrough check hybrids, he showed that the yield advantage had increased to 56% by 1968. He showed further that greater lodging resistance was an important change occurring from 1950 to 1969 because less lodging meant less harvest loss. Using estimates from the Iowa State Corn Yield Tests of 1930- 1970, L. L. Darrah (unpublished observations) showed that the genetic gain for yield was 33%.
11. GENETIC GAINS IN GRAIN YIELD The first experimental evidence for the contribution of breeding to maize yield increases since the introduction of maize hybrids was reported by Russell ( 1974). He evaluated five groups of experiment station hybrids representing approximately 10-year “eras” from 1930 to 1970 and one O.P. at four central Iowa locations during 1971- 1973. Hybrids in eras 1930- 1960 had four double crossesper era, and era 1970had four single crosses. He used three plant densities: (1) 29,700 plants ha-’, a typical farm density for the early 1950s; (2) 44,500 plants ha-’, a typical harvest density for the early 1970s; and (3) 59,300 plants ha-’, a density for higher-level management, particularly in areas of higher average rainfall. Yield data averaged for 11 environmentsfor eras ofhybrids and the O.P. at each plant density are shown in Table I. When comparisons are made at optimal densities, yields from the 1930 era were only 5% greater than those for the O.P., which was much less than the 25 - 35% frequently quoted in earlier years. There seemed a plateau effect with the 1940 and 1950 eras; however, the 1960hybrids yielded significantlyhigher than the 1950era did,
2 so
W. A. RUSSELL Table I
Mean Yields and Linear and Quadratic Regression Coefficients for S i u Eras of Maize Cultivars Evaluated at Three Plant Densities' Plants ha-'
Era
29,700 (1)
44,500 (2)
59,300 (3)
Era mean
Regressions'
b,
ba
-2.45 -4.70 -2.40 -1.35 1.10 4.20
- 1.92 -0.90 - 1.07 -1.95 -1.70 -2.43
q ha-'
O.P. 1930 1940 1950 1960 1970
L.S.D. (0.05)b a
51.5 57.8 65.0 63.0 66.9 72.7 6.4
54.8 55.8 65.8 67.5 73. I 83.7 6.4
46.6 48.4 60.2 60.3 69.1 80.6 6.4
51.0 54.0 63.6 63.6 69.7 78.8 3.7
Data summarized for four Iowa locations in 1971 and 1972 and three Iowa locations in
1973 (Russell, 1974).
Least significant difference (L.S.D.) not applicable to comparison for open-pollinated variety (O.P.). Regression coefficientscalculated by using orthogonal polynomial coefficients(Steel and Tome, 1980); S.E.(b,) = 0.61; (b,) = 0.26.
and there were further significant increases for the 1970 era. The yield increase from the 1960to the 1970eras ( 13.1%)includesthe effect for change from double-cross to single-cross hybrids. Data in Table I1 are for the two hybrids per era that had the highest average yields over all densities and environments; the two best hybrids were used because of probable seed production errors in two groups. On the basis of era means, 1930 yielded 12.2% more than the O.P.did. This may not be a meaningful comparison, however, because O.P.included only one cultivar, which is not likely very representative of cultivars grown before hybrids. With each succeeding era, there were additional increases, the smallest increase occurring between 1940 and 1950. The total increase from 1930 to 1970 was 23.0 q ha-', or 40.2%. Frey (1971)calculated a gain of 49% in the Iowa State Corn Yield Test between hybrids tested in 1926 and 1968. Only 1960 and 1970 hybrids had positive linear regression coefficients, i.e., yields were higher at 59,300plants ha-' than at 29,700 plants ha-'. Era 1930had its highest yield at the lowest density, whereas all other eras were highest at the intermediate density. Era 1970, which had two single-cross hybrids, had the greatest yield increase between the lowest and highest plant densities, b, = 5.20 q ha-'. The most legitimate comparisons may be among yields at the optimal
GENETIC IMPROVEMENT OF MAIZE YIELDS
25 1
Table I1 Mean Yields and Linear and Quadratic Regression Coefficients for O.P. and Two Hybrids per Era that Had Highest Yields Over All Densities, Locations, and Yearsa Plants ha-' Era
29,700 (1)
44,500 (2)
Regressionsc
be
b,
-2.45 -5.40 -0.20 -2.40 0.80 5.20
- 1.92
q ha-'
-
O.P. 1930 1940 1950 1960 I970 L.S.D. (0.05)b
59,300 (3)
Era mean
51.5 61.6 64.8 68.1 69.9 72.7 9.0
54.8 59.2 68.2 73.4 77.7 85.0 9.0
46.6 50.8 64.4 63.3 71.5 83.1 9.0
51.0 57.2 65.8 68.3 73.0 80.2 5.2
- 1.00 -1.20 -2.57 -2.33 -2.36
After Russell (1974). L.S.D. not applicable to comparison for open-pollinated variety (O.P.). Regression coefficientscalculated by using orthogonal polynomial coe5cients (Steel and Tome, 1980);S.E. (be)= 0.86; (b,) = 0.37. a
density for each era. Essentially, the yields had a linear increase across era groups (Fig. 3). An increased gain for the single-crosshybrids (1970 era) was observed. This increase was typical of the U.S. Corn Belt when single crosses became the prominent hybrid type in the 1960s.The yieldgain from the O.P. to the 1970 single-cross hybrids was 55.1%. The yield increase by Iowa farmers for 1922- 1970, which represents the breeding and evaluation period (49 years) for the materials used, was 0.78 q ha-' yr-', or a total gain of 38.2 q ha-'. When all environments,the best two hybrids per group, and the optimal densities are considered, the 1970 group yielded 30.2 q ha-' more than the O.P. did. Consequently, 79% of the increased yield of Iowa farmers can be attributed to the genetic improvement of the hybrids. The effects of improved cultural practices and better hybrids are inseparable in these experiments. The hybrids have given increased yields because of their continued increase in genetic potential to take advantage of improved cultural practices. A second study at Iowa State University (Russell, 1984) obtained further information on the contribution of maize breeding. Most parental inbred lines were of later maturity than those used by Russell ( 1974),and only two lines had been used in the previous study. The O.P. were four cultivars representative of maize grown in the U.S. Corn Belt before the introduction of hybrids. Each hybrid group representing the eras of 1930- 1980 had
W. A. RUSSELL
252
r 80
I
0
9 70
E a F
60
lo O.P. 1930 1940 1950 1960 1970 Cultivar Era
Fig. 3. Observed yield at optimum density for open-pollinated variety (O.P.) and hybrids of each 1 0-year era, 1930- 1970, and predicted linear response based on the observed yields of the two best hybrids in each double-cross group; k = 70.2 2.68X, (Russell,1986).
+
experiment station parental lines available to the seed industry and widely used in breeding and hybrid seed production. Single-crosshybrids were used in all groups to avoid confounding effects of hybrid types in interpreting the data. The materials were evaluated at four locations in 198 1 - 82 with three plant densities: 31,100 plants ha-', 47,800 plants ha-', and 64,500 plants ha-'. All plots were hand-harvested; thus, there were no harvest losses. Group yields in each plant density combined over all environments indicate the average trends for each 10-year period (Table 111).Decreased yields were observed for the 1940 and 1950 eras, which concernedbreeders in those years. Probably, these yield trends occurred because of breeders' efforts to improve stalk strength, which was inadequate in the 1930 era materials. The O.P. and first three hybrid groups had negative b, values because yields were lower in the highest density than in the lowest. The b, value was positive for the 1960 era, and became much greater for the 1970 and 1980 eras. When group yields at optimal densities are considered, the 1930 crosses at Density 2 yielded 19.3 q ha-' (30.4%)more than O.P. did at Density 1. The 1980 crosses at Density 3 yielded 42.1 q ha-' (66.4%) more than O.P. did at Density 1 and 22.8 q ha-' (27.6%)more than the 1930 crossesdid at Density 2. With the exception of the 1930 and 1950 crosses, the yield increases over eras, using the yield at the optimal density for each era, have a close fit to a linear response with b, = 6.25 q ha-' era-' (Fig. 4). Maize yields in Iowa increased 53.4 q ha-' during 1922 to 1980; thus, the total genetic gain was 42.1 q ha-' (Table III), or 79%of the total yield gain.
253
GENETIC IMPROVEMENT OF MAIZE YIELDS Table 111
Mean Yields and Linear and Quadratic Regression Coe5cients for Seven Eras of Maize Cultivars Evaluated at Three Plant Densities'
Plants ha-' Era
31,100 (1)
47,800 (2)
64,500 (3)
Era mean
59.9 82.7 77.7 75.4 90.5 96.6 99.4 6.5
57.6 79.2 72.6 72.7 91.1 100.5 105.5 6.5
60.3 81.0 76.1 74.7 88.2 93.7 98.1 3.2
Regressionsb
bt
b,
-2.90 -0.95 -2.70 - 1.65 4.00 8.30 8.10
-0.20 -0.85 -0.80 -0.35 -1.13 - 1.47 -0.67
-
O.P.' 1930 1940 1950 1960 1970 1980
L.S.D. (0.05)'
63.4 81.1 78.0 76.0 83.1 83.9 89.3 6.5
'Data averaged for four Iowa locations in two years, 1981 and 1982 (Russell, 1984).
Regression coefficients calculated by using orthogonal polynomial coefficients (Steel and Tome, 1980); S.E. (b,) = 0.74; (b,) = 0.43. O.P., Open-pollinated variety; L.S.D.,
Duvick ( 1977)published results for two experimentsconducted in central Iowa, by using mostly different parental materials from those of Russell. Duvick's first experiment had 19 commercial hybrids representative for the period of 1939- 1971, which were compared in plant densities of 32,000, 44,000, and 66,000 plants ha-'. Linear regression values for yield improvement from 1939 to 1971 were 0.16 q ha-' yr-l at the lowest density, 0.67 q ha-' yr-' at the intermediate density, and 0.82 q ha-' yrlat the highest density. The increase of response from lowest to highest density was caused primarily by the decrease of yield by the older hybrids as plant density was increased. His second experiment included hybrids representative of 10-year eras, 1930s- 1970s, 10 single-cross hybrids for each era that were a diallel of five inbreds chosen from the most successful hybrids of a given decade. The hybrids were compared at the same densities as used previously. Linear regression values at the three plant densitiesfor yield improvement over the five decades were 0.40, 0.69, and 0.82 q ha--' yr-'. Yield results in both experiments showed that the improvements allowed newer hybrids to take advantage of higher plant densities. Duvick calculated a gain of 0.88 q ha-' yr- I , based on yield data from Pioneer Hi-Bred maize trials during 19351975. This gain included the effectsof genotypes, cultural conditions, and weather. He determined that genetic gains for yield in the two experiments
254
W. A. RUSSELL 110
100 I-
2
90
T
sr Q)
5 80 70 0
60
I
I
I
I
I
I
-I
O.P. 1930 1940 1950 1960 1970 1980
Cultivar Era Fig. 4. Observed yield at optimum density for open-pollinatep varieties (O.P.)and hybrids of each 10-year era, 1930- 1980, and predicted linear response;Y = 85.13 6.25X, (Russell, 1986).
+
were 0.50 and 0.53 q ha-' p-l, which are 57 and 60%of the total yield gain, respectively. Additional data were presented (Duvick, 1984) from two sets of experiments conducted in Iowa during 1978- 1980 and 1977- 1979. For Experiment 1, which included one O.P. and 47 commercial hybrids evaluated at 30,000,47,500,and 64,000 plants ha-', the linear regression values for yield on year of release were 0.60,l. 10, and 1.20 q ha-' yr-l for the three densities, respectively. Grain yields improved at an increasing rate for 64,000 plants ha-' from 1930to 1980for year of hybrid release; highest yields of the newer hybrids were at the medium and highest densities, whereas the older hybrids had their highest yields at the lowest density. Duvick's second experiment included the 50 single crosses used previously (Duvick, 1977), except that three parental lines were changed for the 1970 era, and the hybrids were compared at the same densities used in the first set of experiments. Generally, the yield results agreed with the results for the commercial hybrids in the first set. For Experiment 1, Duvick estimated that the genetic gain for the breeding period was 0.92 q ha-' yr-I (89%of the total Iowa yield gain) and forExperiment2,0.73 qha-'yr-l(7l%ofthetotalIowagain).Again, much of the gain could be attributed to the increased genetic ability of newer
GENETIC IMPROVEMENT OF MAIZE YIELDS
255
materials to take advantage of higher plant densities to produce more grain per unit area. A further study by Duvick (1991) included Reid O.P. and 41 Pioneer commercial hybrids introduced during 1934 through 1989. The trials were grown at three Iowa locations, for 2 yr and at three plant densities: 30,000, 47,000, and 64,000 plants ha-l. For yields averaged over all densities, locations, and years the yield gain calculated by regression on year of hybrid introduction was 0.73 q ha-' yr-'. Similarly,the gains were 0.49 q ha-' yr-' at the lowest plant density, 0.77 q ha-I yr-' at the medium plant density, and 0.9 1 q ha-' yr-I at the highest plant density. The newer hybrids outyielded the older ones at all densities.The estimated geneticgain, calculated by using yields from the highest yielding density of each era, was 0.76 q ha-' yr-I, or 76%of 1.OO q ha-' yr-' that was the total on-farm yield gain in Iowa. This is similar to values obtained in previous experiments. Yields for newer hybrids continued to have linear gains and showed no sign of leveling out. Tapper (1983)conducted a detailed study for yield, agronomic traits, and physiologic traits of four single crosses per 10-year era for 1930- 1970. Several of the parental lines that she used were the same as those used by Russell ( 1974).Evaluations were done at low, medium, and high plant densities. The total gain for grain yield, based on optimal plant densities, was 28.4 q ha-' (41.6%) for the 1970 crosses over the 1930 crosses. Because of improved lodging resistance for the 1970 crosses, the gain for machine-harvestable yield was even greater at 36.4 q ha-l(67.2%). For machine-harvestableyield, the linear regression values based on plant densitieswere - 1 1.1, -4.7, -0.4, 1.5, and 8.1 q ha-' for the five groups of hybrids for 1930- 1970, respectively. In a similar comparison for machine-harvestableyield, Russell ( 1974) foundayieldgainof28.6 q ha-'(48.l%). Tapper's( 1983)datafortotalgrain yield showed a pronounced plateau for the 1940- 1960 hybrids, but for the machine-harvestableyields, the data showed no such plateau; the predicted gain average over densities was 8.8 q ha-' decade-'. Further studies on the genetic yield improvement of U.S. maize cultivars were conducted by Castleberry et al. (1984). The cultivars were representative of the decades 1930 to 1980s, including O.P. and composites similar to those in use before the introduction of hybrids and commercial hybrids for the eras 1940s- 1980s. Some of their parental materials would be similar to those used in previous studies (Russell, 1974, 1984; Duvick, 1977, 1984; Tapper, 1983),but most would probably be different. As an average over 1 1 location -year environments, which included fertility treatments in two environmentsand imgation treatments in two environments,Castleberryet al. showed a genetic improvement of 0.82 q ha-' y i l . They estimated that this genetic gain was 75%of the increase for averageU.S. maize yields since 1930.
W. A. RUSSELL
256
Table IV S, and S, Means for Yield and Inbreeding Depressionfor OpenPollinated Varieties and S i x Era Populations'
Yield Population
s,
SI
Inbreeding depression
q ha-'
Reid Lancaster Era I Era 2 Era 3 Era 4 Era 5 Era 6 bt
29.1 31.4 30.5 39.1 44.2 41.1 56.3 60.1
5.2 f0.4
22.5 22.1 22.3 29.2 30.8 33.8 42.5 44.1 3.8 f 0.3
6.15*~ 8.1** 8.2** 9.9** 13.4** 13.9** 13.8** 16.6**
22.1' 21.1 26.9 25.3 30.3 29.1 24.5 21.3
*' ** Significant at p = 0.05 and 0.01, respectively. After Lamkey and Smith (1981). S, mean - S, mean. 'Inbreeding depression as a percentage of the S, mean.
Lamkey and Smith (1987) used populations representative of seven eras, pre- 1930s- 1980s, to determine genetic gain from So and S, generations. The pre- 1930s were represented by two O.P.-Reid Yellow Dent and Lancaster Sure Crop-and six decades-1930s to 1980s-by inbred lines representative of the decades crossed and intermated to produce six So populations. Also, to assess changes in rates of inbreedingdepression S, bulks of each O.P. and era populations were used. Each era population had seven to nine parental inbred lines, which included most lines used by Russell (1974, 1984). The cultivarswere evaluated in one plant density, 62,000 plants ha-', which was probably too high for earlier eras but may have been near optimum for Eras 5 and 6 (Russell, 1974, 1984). Machine-harvestable yields were obtained for So and S, generations (Table IV). Except for Era 1, the So generation showed a consistent yield increase from O.P.to Era 6, bt = 5.2 q ha-'. This increase is probably not a true genetic gain because of confounding effects due to decreased lodging from O.P.to Era 6, changes in optimal plant density over eras, and earlier maturity for Eras 1 and 2 than for Era 6. This gain is less than reported by Russell for F, hybrids, probably because he
GENETIC IMPROVEMENT OF MAIZE YIELDS
257
reported gains based on optimal plant density and Lamkey and Smith data do not reflect changes in the magnitude of heterosis over eras. The S, generation had a consistent yield increase from O.P. and Era l to Era 6, bt = 3.8 q ha-'. This increase is significantly smaller than for the So generation. The inbreedingdepression increased from 8.2 q ha-' for Era 1 to 16.6 q ha-' for Era 6; however, inbreeding depression as a percentage of the So showed no directional change and was similar for all eras. The amount of inbreeding depression is determined by allelic frequency, directional dominance, and the number of segregating loci. The yields of the So and SI populations suggest that favorable allelic frequencies were below 0.5 for the earlier eras and have been increasing and/or that the more recent era populations were segregating at more loci. Meghji et al. ( 1984) also reported greaterinbreeding depression of units of grain yield for 1970s hybrids than for 1930s and 1950s hybrids. The inbreedingdepression was greaterat a high plant density than at a low plant density. This finding suggests that a portion of the inbreeding depression occurred at loci conditioning responsiveness of these hybrids to high plant densities. The first hybrid cultivar released in South Africa was a white dent topcross grown by farmers in the 1949/50 season (Kuhn and Gevers, 1980). It was reported to yield 25 to 30%above the yield of all O.P.cultivarsgrown at that time. By using data obtained in Phase I1 cultivar trials, Kiihn and Gevers determined that the increase in yield of the best-yielding hybrids above the mean of the old hybrids was 41.2%, or a 13.7%improvement per decade from 1950 to 1980. By comparing all new hybrids with old hybrids, the average advantagefor 2 1 locations in one season and 25 locationsin a second season was 36.0%. The national gain for maize grain yields in France from 1950 to 1980 was 1.44 q ha-' y r l (Derieux et al., 1987). Derieux et al. evaluated 33 hybrids -early, midseason, and full season-grown in France from 1950 to 1980, to determine the genetic gain in grain yield. The hybrids were divided into three maturity groups and grown in three plant densities. The mean genetic gain averaged for all hybrids was 0.8 q ha-' yr-l, which is 55.6%ofthe total yield gain in France during the same period. By using the optimal density, they determined that the genetic gain was 1 q ha-' yr-' for the early group, 0.6 q ha-' hr-, for the midseason group, and 0.8 q ha-' yr-I for the full-season group. All hybrids except two yielded more at the highest density than at the lowest density, and 24 hybrids had their greatest yield at the highest density. The regression of yield on density increased from 0.08 q ha-' for the 1950s hybrids to 0.59 q ha-I for the 1980s hybrids. Maize yields in Ontario, Canada increased 0.69 q ha-' yr-' (1.5% yr-l) from 1959 to 1988 (Tollenaar, 1989). Tollenaar evaluated commercial hy-
258
W. A. RUSSELL
brids, double crosses and then single crosses, grown in Ontario from 1959 to 1988, to determine genetic improvement in grain yield. The cultivars were compared in experiments that had four plant densities: 20,000, 40,000, 80,000, and 130,000plants ha-’. He used grain yields at optimal density to determine the genetic gain of maize hybrids for the 30-year period. Genetic gain was 1.7% yr-’ for the total grain yield and 2.6% y r - I for machine-harvestable yield, Hybrids used in this study were, or are, generally, grown in a large proportion of the commercial area in Ontario; thus, it seems genetic gain is a close approximation of genetic improvement of Ontario maize hybrids. Misevic et al. (1986, 1987a,b)used long-term data from national varietal trials in Yugoslavia to estimate genetic gains for maize grain yields of released single-crosshybrids in six maturity groups for the period 1966 - 1984. The estimated gains, q ha-’ y-’, were: FA0200, 1.88; 300, 1.64; 400,0.32; 500, 1.37; 600, 1.81; and 700, 1.42. Generally, higher kernel moisture at harvest and decreased percentage of lodged plants were associated with the yield increases. Also in Yugoslavia, Kojic (1 990) reported results from a study that compared yields for four sets of cultivars: an O.P. that was grown in the 1950s before hybrids were used, one double-cross hybrid grown in the 1960s, three single crosses (1 966 - 1973, and three single crosses (since 1975). The cultivars were evaluated at three plant densities: 37,590, 51,950, and 63,490 plants ha-’. Whereas the O.P. and first two sets of hybrids had their highest average yields at the lowest plant density, the most recent single crosses yielded best at the highest density. Based on average yields at the optimum densities, the third hybrid set outyielded O.P. by 91.9%. The actual yield difference was 57.8 q ha-’, or 47.9% of the yield of the single crosses, which Kojic indicated as the contribution of breeding to yield gain. The first double crosses introduced in Brazil in the 1940s had parental lines from one germplasm, “Cateto,” which is a flint cultivar (Paterniani, 1990). Subsequently, in 1954 a semi-dent double cross was introduced that had dent lines from “Tuxpeno” and “Paulista Dent,” and flint lines from Cateto. In the state of Sao Paulo, Brazil, Paterniani (1990) estimated the productivity of maize increased from 30 q ha-’ in 1946, when Catetowas the primary cultivar, to 60 q ha-’ for hybrids becoming available in the 1980s. In the 1960s and 1970s, the introduction of new germplasm, especially from CIMMYT, played an important role in the progress. Estimates of the genetic contribution to total yield gains vary widely (Table V). All estimates, except those by Frey (197 1) and L. L. Darrah (unpublished observations) were obtained in experiments conducted purposely to obtain data for assessing genetic gains. Frey’s and Darrah’s esti-
GENETIC IMPROVEMENT OF MAIZE YIELDS
2S9
Table V Summary of 14 Estimates of Total Grain Yield Gain and the Genetic Yield Gain of Maize Hybrids'
Author
Year reported
Time span
Experiment years
Frey Darrah Russell Duvick Duvick Tapper Tapperc Castleberry et al. Duvick Duvick Russell Tollenaar Derieux Carlone & Russell Kojic Duvick
1971 1973 1974 1977 1977 1983 1983 1984 1984 1984 1984 1989 1986 1987 I990 1991
1926- 1968 1930- 1970 1930- 1970 1935-1971 1935- 1972 1930- 1970 1930-1970 1930- 1980 1930- 1980 1930- 1980 1930- 1980 1959- 1988 1950- 1980 1930- 1980 1946- 1989 1930- 1989
1926- 1968 1930- 1970 1971-1973 1972- 1973 1972- 1973 1980-1981 1980-1981 1980- I981 1978-1980 1977- 1979 1981-1982 1987-1988 1983- 1984 1987 1985- 1986 1989- 1990
Totalb gain (q ha-I yrl)
Genetic? gain (q ha-l yr-l)
Genetic?
-
-
0.99 0.78 0.88 0.88 0.78 0.78 1.10 I .03 I .03 0.90 0.69 1.44 0.89 0.85
0.33 0.63 0.50 0.53 0.40 0.73 0.82 0.92 0.73 0.7 1 1.34 0.80 0.72 0.25 0.76
56 33 79 57 60 51 94 75 89 71 79
1
.oo
gain (%)
-
56 81 29 76
After Duvick (1984, 1991) and Russell (1986). Gains calculated basis US.maize yields by Castleberry et a/.;Ontario yields by Tollenaar; France yields by Derieux; Yugoslavia yields by Kojic; and state of Iowa yields for all other estimates. Gains calculated relative to 1930-era hybrids; first estimate- total yields, second estimate- machineharvest yields.
mates were obtained by using data from the Iowa State Corn Yield Tests. The estimates ofgeneticgain, compared with those of O.P., range from 56 to 94% for the planned experiments; some may be considered actual genetic gains, whereas others probably are biased because of certain procedures. For example, the estimates from machine-harvestedplots (see footnotes, Table V) are likely biased in favor of the newer hybrids because greater stalk lodging of the older hybrids would be expected to have caused greater harvest losses. Russell ( 1974,1984) and Duvick ( 199I ) made adjustmentsto the calculated genetic gains because the experimental test average yields were higher than the on-farm state averages. With these adjustments, the values for genetic gain were 63% (Russell, 1974), 56% (Russell, 1984), and 56% (Duvick, 199 1). Regardless of what biases may be present in the estimates for genetic gain, all estimates show the tremendouscontribution ofplant breeding to the
260
W. A. RUSSELL
total maize yield gain in the United States during the 50 years that hybrid cultivars have been used.
111. GENETIC GAIN: STRESS VERSUS NONSTRESS ENVIRONMENTS Yield levels among environments for the experiments reported by Russell (1974, 1984) ranged widely, primarily because of soil moisture effects. For the first evaluation (Russell, 1974), the yield range, averaged over all densities and cultivars, was from 48.0 q ha-' in the lowest-yielding environment to 84.8 q ha-' in the highest-yieldingenvironment. Consideringonly the two highest-yielding environmentsand the optimal density for each era, the 1970 hybrids yielded 28.4 q ha-' (35.4%) more than the 1930 hybrids did. A similar comparison in the three lowest-yielding environments showed that the 1970 hybrids yielded 25.6 q ha-' (59.0%) more than the 1930 hybrids did. Tapper (1983) also found that the relative superiority of the 1970 hybrids was greater in the lowest-yielding environment than in the highestyielding environment. For the second evaluation (Russell, 1984), the yield range, averaged over all densities and cultivars, was from 64.0 q ha-' in the lowest-yielding environment to 92.7 q ha-' in the highest-yielding environment. When environments were divided into two groups on the basis of average yields, and when comparisons were made on the basis of the optimal density for each era group, the 1980 hybrids had 62.0% in the high environments and 69.9% in the low environments more yield than did O.P. A stability analysis (Eberhart and Russell, 1966) of the data presented by Russell (1984) can be used to compare the seven cultivar groups for yield response with the environment yield level (Fig. 5). There are eight indexes (4 locations X 2 years), and each index is the average over 28 cultivars and 3 plant densities. In the stability analysis, the response to the environmental index, averaged for all entries, is 1.OO, and responses for individual entries are lesser or greater than 1.OO. For each era, the yield data for the optimal density were used. The 1970 and 1980 hybrids had superior yields in all environments, which included two drought-stress locations and two highyielding environments. Both groups have b values less than 1.O, whereas the 1930 and 1960 hybrids have b values greater than 1.O. These values indicate that the relative superiority of the more modem hybrids was greater in the low-yielding (stress) environments than in the high-yielding (nonstress) environments. The 1940 and 1950 hybrids and the O.P. also have low b values, but at much lower average yield levels than do the 1970 and 1980 hybrids.
GENETIC IMPROVEMENT OF MAIZE YIELDS
26 1
110
100
v-
‘m 90 r U
Li 2 >
80 r2 =0.91
1940
70
60
50
65
85
75
95
Index
Fig. 5. Yield responses, b, for open-pollinated varieties (O.P.) and six hybrid groups of 10-year eras, 1930- 1980, to eight environment indexes (4 locations X 2 years) (Russell,1986).
Consequently, there seems to be no distinct relation between response and era of the hybrids. More likely, the responseswere specific for the genotypes. For example, 1930-era hybrids had two parental lines that contributed prolificacy, which was strongly expressed in the high-yielding environments. Duvick (1984, 1991) showed substantial differences between older and newer hybrids in their abilities to yield in environments of varying production levels. Hybrids of the 1969- 1978 subset (Duvick, 1984) had a linear response of b, = 1.21, whereas those of the 1930- 1944 subset had a linear response of b, = 0.77. Also, 1980-era hybrids had a linear response of b, = 1.51 (Duvick, 199l), whereas the 1930-era hybrids had a linear response of bt = 0.4 1. Similarly, Castleberry et al. ( 1984) obtained a linear response of b, = 1.20 for 1980 hybrids and of 0.82 for 1940 hybrids. In both studies (Duvick, 1984; Castleberry et al., 1984), newer hybrids yielded more than did older hybrids in low-yielding environments, and the yield difference increased from low- to high-yielding environments. Derieux et al. ( 1987) found that yield improvement for very early, early, and midseason hybrids resulted more from yield stability than from improvement of potential yield; i.e., average responses relative to environment yields were b, = 1.20 q ha- *
262
W. A. RUSSELL
for 1950 hybrids and b, = 0.82 q ha-' for 1980 hybrids. These results are similar to those reported by Russell ( 1974), who compared 1930 and 1970 hybrid yields in low- and high-yielding environments. There is a discrepancy between the results of Russell (1 984) with those of Duvick (1984, 199 1) and Castleberry et al. (1984) in terms of the relative performance of old and new hybrids in stress and nonstress environments. Differences in germplasm may have been the cause, but this does not seem the only reason because certain genetic relations among the hybrids of the three studies also seem likely. Differences in harvest losses may have been a factor in causing the discrepancy. Russell's study obtained total yield because all plots were hand-harvested, whereas the studies of Duvick and Castleberry et al. used machine harvest, which may have caused some yield losses in plots that had a higher incidenceof stalk lodging. Old hybrids would be expected to have more stalk lodging than new hybrids would, and this greater lodging would be accentuated in higher plant densities, which would probably produce the highest yields in the high-yielding environments. Tapper (1 983) found that harvest losses were appreciably higher for the older than for the newer hybrids. Further comparison among decades of cultivars for ability to cope with moisture stress was presented by Castleberry et al. (1984). In two locations for one year, cultivars were grown under imgation and dryland. The yield response was 1.23 q ha-' yr-' under imgation and 0.84 q ha-' yr-l under dryland (Fig. 6). Modem hybrids not only yielded more grain than did the older hybrids in low-moisture environments, but they also exhibited the genetic capacity to increase this yield difference in the high-yielding environment, i.e., the 1930 cultivars were limited in their yield potential.
IV. RESPONSE TO INCREASE IN PLANT DENSITY AND NITROGEN FERTILITY The increase in plant densities used by farmersfrom the 1930sto the 1980s has been an important factor in increasing grain yield in the U.S. Corn Belt. Cardwell ( 1982) calculated that increased plant densities were responsible for 21% of the gain in maize yields by Minnesota farmers for the period 1930- 1979. Russell (1974) found that the O.P.and hybrids of the 1930, 1940, and 1950 eras all had significant, negative linear responses to plant densities, whereas the 1960 hybrids had a positive linear response, and the 1970 hybrids had a significant,positive linear response (Table I). Even in the two highest-yielding environments, the 1930 hybrids had their highest average yield at the lowest plant density, whereas the 1970 hybrids had their
GENETIC IMPROVEMENT OF MAIZE YIELDS
100
b' _r2 _ -q ha-'yr-' 0 Full Irrigation** 0.91 A Stress" 0.97
26 3
Y
1930's 1940's 1950's 1960's 1970's 1980's
Decade of Use Fig. 6. Regression analysis of decade group mean yields of maize cultivars grown under fully irrigated and stress conditions at Hastings, Nebraska and at Yuma, Colorado in 1981 (Castleberry et al., 1984). ** Correlation significant at the 0.01 level of probability. t b values different at the 0.01 level of probability.
highest average yield at the highest plant density (Fig. 7). The yield difference between the two eras was only 4.6% at the lowest plant density, but the difference increased to 44.2% at the intermediatedensity and to 70.4% at the highest density. For 1980s hybrids, Russell ( 1984) obtained a yield increase of 11.3% from a low plant density to an intermediate plant density and a further gain of 6.8% for the highest density, but the 1940 and 1950 hybrids had their highest average yields at the lowest plant density, and yields decreased successively for the intermediate and highest plant densities (Table 111). Duvick ( 1977, 1984, 199 1) and Tapper (1 983) also showed that higher plant densities contributed to newer hybrids yielding more than the older hybrids did. In all these evaluations, the superiority of the newer hybrids for resistance to barrenness was evident. For example, in the study by Russell ( 1984), the 1940 hybrids had 0%barren plants in Density 1,9% barren plants in Density 2, and 2 1% barren plants in Density 3; the 1980 hybrids had 1 1 3 ears per 100 plants in Density 1, 1% barren plants in Density 2, and 5% barren plants in Density 3 (W. A. Russell, unpublished observations). Tollenaar ( 1989) calculated optimal plant densities for yield of each hybrid by a formula presented by Duncan (1958). Tollenaar found that the optimal plant density for total yield was greater for more recent hybrids than for older ones; however, the increase in optimal plant densities did not continue beyond the 1970-era hybrids. By using yield data reported by Russell ( 1984), Tollenaar ( 1989) determined that optimal plant density of
W. A. RUSSELL
264
29.7
44.5 Plants ha-' (~1000)
59.:
Fig. 7. Observed yields for two highest-yielding hybrids per era at three plant densities averaged for the two highest yielding environments(Russell,1974). (O), 1970; (0),1930.
U.S. Corn Belt cultivars increased from the 1920s to the 1980s at a rate of 0.97 plants/m2 10-yr era-', although there was essentiallyno increase for the 1940s and 1950s cultivars. For late-maturing hybrids in France, which are representative of the 1950s to the 1980s hybrids, the linear increase in optimal plant density for total grain yield was 0.96 plants/m2 10-yr era-' (Derieux et al., 1987). The early maturing hybrids for the same eras, however, did not show a clear trend (Tollenaar, 1989).
Increased fertility levels have played a key role in the increase of maize yields since about 1950, and the dominant plant nutrient causing this increase has been nitrogen (N). Nitrogen fertilizer use in five central Corn Belt statesincreased from zero in 1945 to about 140 kg ha-' in 1980 (Thompson, 1982). Cardwell (1982) reported that the average use of N fertilizer in Minnesota increased from 16.8 kg ha-' in 1950 to 112 kg ha-' in 1979, and this N-use increase accounted for 47% of the yield increase realized by Minnesota farmers from 1930 to 1979. With adjustments for decreased use of animal and green manures; however, the net yield increase was estimated at 19%. The level of soil fertility has not been used directly as a selection tool in the developmentand evaluation of maize inbred lines. Several studies, however, present information on the improvement of maize hybrids for N-use efficiency (i.e., weight of grain produced per unit of N supplied) and yield
GENETIC IMPROVEMENT OF MAIZE YIELDS
265
response for N levels. Beginning in the late 1950s, maize yields in the United States accelerated (Fig. 2). Causes suggested for this acceleration have included higher N levels, higher plant densities, use of single-cross instead of double-cross hybrids, and favorable weather conditions. The role that improved hybrids may have played, relative to that of N response, in the accelerated yields of the 1960s is not generally known. Duvick ( 1984) evaluated four single-cross hybrids representative of four decades- 1940 to 1970 -at low, intermediate, and high N levels. At all N levels, the newer hybrids outyielded the older hybrids, but the hybrid X N level interaction was not significant. Thus, when there is N deficiency, the newer hybrids will outyield the older hybrids, Kamprath et al. (1 982) evaluated three maize population hybrids at three N levels. The population hybrids were "Jarvis X Indian Chief," Jarvis X Indian Chief after eight cycles of full-sib family selection for intrapopulation performance, and Jarvis X Indian Chief after eight cycles of reciprocal recurrent selection for interpopulation performance. The N levels were 56, 168, and 280 kg ha-'. The improved population hybrids produced more grain at each N level than did the original population hybrid. The greater grain yield was associatedwith an increase in ear number per plant as the N rate increased. Moreover, the average N-use efficiency for the improved hybrids was greater than that for the original hybrids at all N levels. Castleberry et al. (1984) compared cultivars from six decades- 1930 to 1980-in low- and high-fertility conditions at one location for 2 years. The high-fertility area had received normal fertilizer applications for 20 years, whereas the low-fertilityarea had been in continuous maize and unfertilized since 1958. For this study, the high-fertility area received approximately 200 kg ha-' N, 90 kg ha-' P,05, and 150 kg ha-' K,O in both 1980 and 198 1 ;the low-fertilityarea received no fertilizer. The yield response relative to decades of cultivars was 0.87 q ha-' yr-' in the high-fertility condition and 0.5 1 q ha-' yr-I in the low-fertility condition (Fig. 8). The newer hybrids were superior to the older cultivars in both fertility levels, and the superiority was greater for the high-fertility area than for the low-fertility area. Studies seem to show that hybrids of the 1970 and 1980 eras have the genetic potential to take advantage of both increased plant densities and N fertility; consequently, plant densities and N fertility would be expected to have cumulative effects on yield response. Carlone and Russell (1987) evaluated four O.P. cultivars and 24 single crosses, which were representative of the years 1930- 1980, in three plant densities and four N levels, for 2 years. The maize cultivars were the same as used by Russell (1984). The plant densities were 34,445, 5 1,661, and 68,889 plants ha-'; the N levels were 0, 80, 160, and 240 kg N ha-'; thus, there were 12 treatment combinations. Considering cultivar yield response to N levels, O.P. era had the lowest
W. A. RUSSELL
266
q ha'yr-'
r2
.Low Fertility" r
0.51 A
0.97
70
r u 60
I
I
I
I
I
1930's 1940's 1950's 1960's 1970's 1980's
Decade of Use
Fig. 8. Regression analysis of decade groupmean yields of maize cultivars grown under high and low fertility at Dayton, Iowa, 1980- 1981 (Castleberry d al., 1984).** Correlation significant at the 0.01 level of probability. t b values different at the .01 level of probability.
response (b, = 4.90 q ha-'); 1930, 1960, 1970, and 1980 era hybrids had intermediate response (b, = 5.70,5.33,5.66,and 6.39 q ha-', respectively); and 1940 and 1950 era hybrids had the highest response (b, = 7.01 and 7.20 q ha-', respectively). The 1940 and 1950 eracultivars,however, had the lowest average yields of the hybrid cultivars. Although these data do not indicate that greater response to N use has been an important selection achievement in newer hybrids, B73 X Mo17 of the 1970 era and B84 X Mo 17 of the 1980 era had among the highest b, values for linear response to N levels. Both densities X cultivars and N-levels X cultivars were highly significant, but the highly significant interaction for densities X N levels X cultivars may be the most important. Evidently, the responses caused by plant densities and N levels were interdependent, and interdependence varied among cultivars. Data for two hybrids (Table VI) illustrate contrasting responses that would contribute to the densities X N levels X cultivars interaction (Carlone, 1985). The B73 X Mo17 had a highly significant linear X linear component,whereas B73 X Va26 did not. For B73 X Mo 17, there was a wide range for linear responses to densities among N levels, b, = 0.60- 14.85 q ha-', whereas B73 X Va26 had a much narrower range for linear responses to densities among N levels, b, = 4.55-8.80 q ha-'. Also, B73 X Mo 17 had a much greater range than did B73 X Va26 for linear responses to N levels among different densities. The response surface for B84 X Mo17 is shown in Fig. 9 (Russell,1986). Yield decreased from the
GENETIC IMPROVEMENT OF MAIZE YIELDS
267
Table VI Grain Yields for B73 X Mo17 and B73 X VaM at 12 Density X N-level Treatments'
N-level (kgha-I)
Plant densities (q ha-l)b 1
2
3
b,"
52.0 70.1 84.0 81.0 5.04
50.7 88.8 91.2 96.2 6.94
53.2 77.3 100.3 110.7 9.78
0.60 3.60 8.15 14.85
48.1 72.9 71.5 70.0 3.22
53.4 17.2 89.3 78.1 4.3 I
58.4 90.5 80.6 85.0 3.50
B73 X Mo17 0 80 160 240 br
B73 X Va26
0 80 160 240 b,
5.15
8.80 4.55 7.50
'After Carlone (1985). Densities 1, 2, and 3 represent 34,445, 5 1,66I , and 68,889 plants ha-I, respectively. 'S.E. for plant density, br = 2.84. S.E for N-level, b, = 0.90.
lowest to the highest densities at 0 N but increased for all other rates, and yield increased as plant densities increased. Fakorede and Mock (1982) evaluated populations from two recurrent selection programs for responses to four N levels: 0, 90, 180, and 270 kg N ha-'. In one program, two synthetics-Iowa Stiff Stalk Synthetic (BSSS) and Iowa Corn Borer Synthetic No. 1 (BSCB1)-had been used for seven cycles in reciprocal recurrent selection to improve yield of the population cross. In comparisons of population crosses BSSSCO X BSCBICO, BSSS(R)C5 X BSCBl(R)CS, and BSSS(R)C7 X BSCBl(R)C7, crosses C5 X C5 and C7 X C7 yielded significantlyhigher than did CO X CO at all N levels, and C7 X C7 yielded significantlyhigher than did C5 X C5. Thus, the improved populations utilized N more efficiently than did the unimproved population, but the responses to applied N were not different among population crosses from 0 to 270 kg N ha-'. In a second program, in which the
268
W. A. RUSSELL
Fig. 9. The response surface for B84 X Mol7 evaluated in three plant densities and four N levels for two years (Russell,1986).
selection population was BS 12 with inbred B 14A as a tester, recurrent selection was used for six cycles to improve the cross of BS 12 X B 14A. The yield advantage of BS12C6 X B14A was evident at zero N and increased With successivelevels of N so that the N-linear responses were significantlydifferent. Evidently, in one breeding program, selection did not select for yield genes responding differently over N levels, whereas in the second program, selection was effective in increasing frequencies for alleles affecting N response. Modern-day hybrids have greater N-use efficiencies than do hybrids of earlier eras (Duvick, 1984; Castleberry et al., 1984; Carlone and Russell, 1987). The improvement in N-use efficiency is a part of the total genotype developed for highergrain yields. Muruli and Paulsen ( 198 1 ) presented some evidence that selection in a low or high N-level environment will give rise to materials that are N efficient or N inefficient, respectively. Do elite singlecross hybrids differ in N-use efficiency, possibly because of being developed under different levels of soil N? Tsai et al. (1984) reported results that indicate differences among hybrids for response when grown at five N-levels between 0 and 447 kg ha-’. “Pioneer 3732” attained its maximum yield at
GENETIC IMPROVEMENT OF MAIZE YIELDS
269
140
120 I’ 0
100
r
c
80
U
5 al
5 60
40
20
0
I
67
I
I
I
134
201
268
I
44;
N Rate, kg ha”
Fig. 10. Grain yields of two singlecrosshybrids at six levels of N fertilizer(Tsai el al., 1984). (0),B73 X Mo17; (O),Pioneer 3732.
67 kg N ha-’, whereas B73 X Mo17 attained near maximum yield at 268 kg N ha-* (Fig. 10). Additional analysis showed that Pioneer 3732 and B73 X Mo 17 accumulatedN differently as N fertilizer increased from 20 1 to 447 kg N ha-’. Some factors perhaps confounding definite conclusions should be indicated, however: maturity differences, Pioneer 3732 being considerably earlier than B73 X Mo17; only one plant density, 54,000 plants ha-’, which may have been near optimum for B73 X Mo 17 but probably was too low for Pioneer 3732; and genetic yield potential, which is probably greater for B73 X Mo17 than for Pioneer 3732. Tsai ef al. (1984) cited three factors affectinggrain yields of B73 X Mo 17 and Pioneer 3732: (1) the ability of the plants to absorb exogenous N after the milk stage; (2) the duration of grain fill; and (3) the rate of N synthesis (kernel N sink) affected by N fertilizer. Bundy and Carter (1988) evaluated five commercial single crosses in one season and seven commercial single crosses in a second season at four N
270
W. A. RUSSELL
levels in southern Wisconsin. Nitrogen levels and hybrids had highly significant yield differences, but the interaction of N levels X hybrids was not significant in either test. Gardner et al. (1 990) reported results from four studies that evaluated four sets of commercial single crosses at various N levels. The largest experiment had 12 hybrids grown in three N levels at five locations in 1986 and four N levels at eight locations in 1987. Mean squares for N levels and hybrids were highly significant, but N levels X hybrids was not significant in either year. None of the other experiments found significant interaction for N levels X hybrids. Authors concluded that it is difficult to predict a hybrid response to N, and that the response is highly dependent on the growing environment. If the hybrids being compared have similar maturities, perhaps the failure to obtain significant N levels X hybrid interactions should not be a surprising result. These were elite hybrids developed and selected for commercial use after extensive testing in many locationyear environments in which soil N level would be one variable. Inbred lines usually show highly significant differences for response to N levels, which one may expect to affect responses for hybrids in which the inbred lines are used as parents. Figure 1 1 shows three significantlydifferent responses for line yields averaged for three years for one location (Russell, 1984). The optimal N level was different for each inbred line. Russell and Balko (1980) studied the effects of N levels on the prediction of yield response of hybrids on the basis of the response ofinbred parents. The response curves for single-cross 43 (Fig. 12) and the average of its parents, 5 and 6, (Fig. 13) were very similar; single-cross 53 (parents 25 and 26) had a good efficiencyresponse, but single-cross 53 required 240 kg N ha-' for its greatest predicted yield. Single-cross59 was less efficient in terms of N use, a quality suggested by its parents, 37 and 38. Although in some instances predicted response showed some similaritybetween parent lines and single-crossprogenies, there were as many examples with little similarity between responses, particularly if the zero N level is omitted. The data also suggested that breeders may be able to develop more efficient N use of inbred lines by using a more moderate N level.
V. CHANGES IN OTHER PLANT TRAITS Favorable changes for agronomic traits other than yield have been an important part of the total improvement for maize cultivars through breeding, the most notable changes being resistance to root and stalk lodging, resistance to barrenness, and plant health. Much improvement in resistance to lodging was necessary to permit machine harvest and use of higher plant
GENETIC IMPROVEMENT OF MAIZE YIELDS
I
I
I
60
120
180
27 1
240
N Rate, kg ha-l
Fig. 11. Grain yields for three inbred lines to five levels of N fertilizer (Russell,1984). (O), B73; (A), H98; (0),B79.
densities. Duvick ( 1977) found that root lodging resistance improved continuously from the 1930- 1960 hybrids, with only minor improvement for the 1970 hybrids. In a second study, Duvick (1 984) showed that the root lodging resistance improved in one set of comparisons continuously from the 1930 decade to the 1980 decade and, in a second set of comparisons, from the 1930decade to the 1960decade, with little further improvement for the 1970decade. In his most recent report, Duvick ( 1991) showed significant yield improvements accompanied by favorable changes in several important agronomic traits, and the changes were achieved without unfavorable changes for plant and ear heights, days to anthesis, and percentage of grain moisture. Russell (1984) found highly significant improvement for root lodging from O.P. to 1930-era hybrids, but significant improvement was not observed for subsequent eras to 1980. Lamkey and Smith (1987), who used much of the same germplasm as did Russell, also found that significant gain was realized from O.P. to 1930-era hybrids, with no significant gain subsequently. Duvick (1977) showed that much of the improvement for stalk lodging resistance occurred from the 1940 to the 1950 decade. A similar result was obtained for one set of hybrids in a second study (Duvick, 1984), whereas for a second set much of the gain was continuous for the 1950 to
W. A. RUSSELL
272
I
I
I
60
120
180
i
N Rate, kg ha’’ Fig. 12. Genotype X yield response for three selected single crosses to five levels of N fertilizer (Russell and Balko, 1980). (A), 43; (U),53; (0), 59.
1980 decades. In his later study, Duvick (1991)found a continuous improvement for stalk lodging resistance from the 1930sto the 1980s hybrids. Russell (1 984)and Lamkey and Smith (1 987)showed improvement for stalk lodging resistance from the 1930 hybrids to the 1970 hybrids. Duvick (1 984) showed substantial improvement for stalk lodging resistance from the 1930 to the 1950 hybrids, and he attributed the lower yields of the 1940 and 1950 hybrids compared with that of the 1930 hybrids to the selection of much greater stalk strength. Tapper (1983),who used the same germplasm in single-cross hybrids as Russell ( 1974) had used, showed that significant resistance to root and stalk lodging was achieved in the 1940-and 1950-decade crosses. She had total grain yield and machine-harvested yields on the same plots. Lodging and dropped ears caused a harvest loss of 27.1% for the 1930-decadehybrids but of only 1.2% for 1960 hybrids and of3.6% for 1970 hybrids. Derieux ef al. (1 987)showed much improvement for lodgingresistance from 1950- 1980 hybrids in three maturity groups. In earlier years of inbred line development and hybrid evaluation, plant densities were relatively low (i.e., usually about 30,000plants ha-’), which was typical of densities used by farmers in the Corn Belt. Consequently, in
GENETIC IMPROVEMENT OF MAIZE YIELDS
--I-
27 3
A
80
70 60 -
2U
r
d Q)
F
50-
40-
30 20 -
s 1
lot
O O
60
120
180
2
0
N Rate, kg ha-’
Fig. 13. Genotype X yield response for three selectedparental inbred pairs to five levels of N fertilizer (Russelland Balko, 1980). (A), 43; 53; (0),59.
a),
these early years there was relatively little stress on ear development, and barren plants were of no importance. Beginning about mid-l950s, with increased use of commercial fertilizers, farmers gradually increased plant densities and breeders did likewise because it quickly became evident that some parental inbred lines and hybrids were prone to barrenness at higher plant densities. With the greater stress on plant development in the higher plant densities, silk emergence was sometimes delayed relative to time of pollen shed, and with extreme stress silks did not emerge at all. In comprehensive breeding studies, Russell and Teich (1967) and Russell and Machado (1978) found that inbred lines developed in high plant densities had greater resistance to barrenness than did lines developed at low plant densities, and this trait was transmitted to testcrosses of the parental lines. Russell (1974, 1984)had found that the number of ears per plant, either barrenness or more than one ear per plant, was an important contributing trait for yield. Duvick (1984) also found that reduced barrenness was an important yield component, particularly at higher plant densities. In all three studies, hybrids
274
W. A. RUSSELL
of more recent eras exhibited less barrennessthan did most hybrids of earlier eras. Crosbie (1982)reported that reduced barrenness, rather than increased prolificacy, was an important factor for yield increases over decades. Greater resistance to barrenness, i.e., primarily the ability to have silk emergence concurrent with tassel shed obtained by breeding of recent era single crosses, has permitted higher plant densities and, thus, greater grain production per unit land area. Several plant and ear traits have changed with yield improvement through breeding, whereas other traits have shown no significant trends. In some instances, selection during inbred development and evaluation may have been partly responsible for trait changes, whereas for others, the change may have been primarily a correlated response with the increase for yield. Russell ( 1985)evaluated 18plant, ear, and grain traits of28 cultivarsfrom seven eras (Table VII) and correlated these traits with grain yield. These cultivars were the same cultivars evaluated in a previous study (Russell, 1984). Significant differences among O.P. and era hybridswere obtained for all traits; however, the important trends were: earlier silk emergence and decreased pollen-silk interval, longer time to black layer from silk emergence (physiologicrnaturity), improvedplant health (stay-green rating 1 .O = no dead tissue, 9.0 = no green tissue at about 54 days after average silk emergence), increased total dry weight (dry stover weight probably increased because greater leaf area and plant and ear heights had no general relation to total plant weight), increased harvest index (single-cross B73 X Mo17,1970era, was 54.3% and B84 X Mo 17,1980era, was 54.7%), and increasesfor all ear traits, weight per 300 kernels (K), and grain yield. Relative yields for era grain yields were similar to those of a previousstudy (Russell, 1984),except the average for the 1980 era was less than for the 1970 era. Comparisons of the two studies at comparable plant densities showed that the maximum era gain relative to O.P.was 65.9% in the first instance and 57.5% in the second. The plant density used in the second study, 5 1,670plants ha-', may have been too high for maximum yield of O.P.and too low for maximum yield of 1970-and 1980-era single crosses. Simple correlations of plant, ear, and grain traits with grain yield (Table VIII) were greatest for total plant weight ( I = 0.94**)and harvest index (r = 0.83**).Stay-green would affect time ofphysiological maturity and rate of grain fill, and both traits had high r-values (r = 0.71** and 0.77**,respectively). The ear and grain traits, i.e., components of grain yield, had significant but lower correlations. Ears per plant did not have a significant correlation with grain yield because differences among cultivars for ears per plant were not significant.Plant density in this experiment was probably not high enough to cause barrenness in O.P.and in earlier-era hybrids.
Table VII
Mean Values for 18 Plant, Ear, and Grain Traits for Each of the Seven Eras of Maize Evaluated in Two Years,1982 and 1983' ~
~
Days
Era
Pollen shedb
O.P. 1930 1940 1950 I960 1970 1980 L.S.D. (0.05)
20.8 24.0 21.6 21.8 23.7 22.0 21.6 0.6
Silk
Height Black layef
Black laye8
Plant
emergenceb
Pollen -silk interval
24.9 27.2 25.0 24.8 26.0 23.2 23.7 0.5
4.1 3.2 3.4 3.0 2.3 1.2 2.0 0.5
17.0 17.2 16.9 19.0 21.2 24.1 24.2 2.7
54.1 52.0 53.9 56.2 57.1 62.9 62.5 2.6
233 265 242 23 1 250 237 23 1 4.3
Ear (cm) 155 148 124 112 127 115 116 4.3
staygreen (1 -9)
Total dry weight
6.2 5.4 5.7
1549 1776 1718 1759 1925 2043 2053 112
5.5
5.1 3.9 4.3
1.o
(g plot-')
Ear Harvest index (%)
Length
Era
Rate grain m (g plant-' day-')
O.P. 1930 1940 1950 1960 1970 1980 L.S.D. (0.05)
2.6 3.4 3.1 2.9 3.2 3.4 3.3 0.25
44.6 49.1 48.4 45.4 47.2 52.6 50.2 2.6
17.8 20.7 18.3 19.1 19.0 20.6 20.5 1.1
From Russell (1985). Days basis July 1. Days basis September 1. ti Days from silk emergence to black layer. a
Diameter (cm) 4.43 4.46 4.72 4.60 4.80 4.70 4.68 0.13
Kernel depth (cm)
Ears per plant
0.83 0.9 1 0.93 0.88 0.92 0.97 0.94 0.04
1.00
1.09 1.01 0.98 1.01 1.02 I .05 0.08
Weight per 300 kernels (g)
81.0 84.7 83.4 81.8 82.9 86.5 85.6 0.9
81.6 74.7 78.3 82.7 82.8 95.1 90.9 3.1
70.2 89.4 83.9 82.2 93.0 110.6 106.1 4.7
276
W. A. RUSSELL Table VIII Simple Correlation Coefficients for 18 Plant, Ear, and Grain Traits with Grain Yield for 28 Cultivars Evaluated in Two Years, 1982 and 1983a Trait
r-values
Days to pollen shed Days to silk emergence Pollen -silk interval Date black layer Days silk emergence to black layer Plant height Ear height Stay-green Total plant weight Stover weight Rate grain-fill Harvest index Ear length Ear diameter Kernel depth Ears per plant Shelling percentage Weight per 300 kernels
0.19 -0.32 -0.76** 0.68** 0.7 1 0.09 0.03 -0.65** 0.94** 0.67** 0.77** 0.83** 0.63** 0.42* 0.57** 0.28 0.68** 0.58**
**
*’ ** Significantat 5 and 1% levels, respectively. From Russell (1985).
Some trait changes may be specific to the particular cultivars used rather than a general occurrence of all germplasm representative of similar eras. Although Russell (1985) obtained a significant increase in harvest index from 1930 to 1970 and 1980 eras, Tapper (1983) observed a significant decrease from 1930 (54.0%) to 1960 (45.1%), and then an increase to 1970 (49.7%). Tollenaar (1989) found no significant changes for harvest index at plant densities of 2,4, and 8 plants mW2,whereas significant increases were obtained at 13 plants m-2 for hybrids from 1959 to 1980. On average, Tollenaar stated that approximately 15% of the genetic gain in total grain yield could be attributed to increased harvest index. Neither Tapper (1983) nor Russell (1985) found distinct trends for rate of grain fill from older to newer hybrids (Tables IX and VII). In both studies, however, grain fill duration or days to black layer increased significantly. When significantgenetic gains for grain yield are obtained, it is appropri-
Table M Mean Values for Plant Traits of Hybrids Representative of 1930-1970 Eras'.'
(%I
Rate of grainfilling (g day- I )
Duration of grainfilling (days)
CER'
LAIC
LOVA'
52.7 52.7 48.7 44.0 49.1 1.a
3.2 3.4 3.0 3.3 3.6 0.2
59.1 59.8 66.5 64.3 63.7 1.2
40.1 38.8 39.9 40.5 37.0 3.4
3.0 3.1 3.2 3.0 3.1 0.2
27.3 26.4 25.1 30.4 43.0 1.7
Decade of hybrids
Harvest index
1930 1940 1950 1960 1970 L.S.D. (0.05)
LOVB'
Springvigor rating'
staygreen ratinp'
30.2 28.2 28.3 32.6 32.2 1.1
6.5 6.4 6.5 6.6 7.0 0.2
1.8 2.7 4.8 6.8 4.4 0.9
'Averaged over three plant densities and several environments. After Tapper (1983). Cer, Carbon dioxide exchange, mg dm-I & I ; LAI, leaf-area index; LOVA, leaf orientation value above the ear,LO W , leaf orientation value below the ear.Spring-vigor rating, 1 = lowest vigor, 9 = highest vigoq stay-green rating, 1 = least green leaf area, 9 = greatest green leaf area.
278
W. A. RUSSELL
ate to determine changes that may have occurred for physiological traits that seem important to grain yield. Tapper (1983) obtained data for several physiological traits at three plant densities in four location- year environments (Table IX). Carbon dioxide exchange rate (CER) and leaf-area index (LAI) showed no trend over eras. Russell (1985) showed a significant increase for total plant weight at harvest; this increase seems likely to have been accompanied by an increase for LA1 because he observed no trend for plant or ear heights. Tollenaar ( 1989) presented evidence of an increase of LA1 at four plant densities. Tollenaar ( 1989) also found increases from old to new hybrids for final leaf number and production of total dry matter. Whereas the production of total dry matter increased from old to new hybrids during three phases of plant development from planting to maturity (Tollenaar, 199 I), the increase was greatest during the period 3 weeks past silking to maturity. The greater dry matter accumulation of the newest hybrids seemed to be associated with increased stress tolerance. Leaf orientation above the ear (LOVA) (Table IX) showed a change toward more erect leaves, the greatest change being for the 1970 era. Rather than a general change in leaf orientation, the distinct increase of LOVA for the 1970 era is more likely a contribution by parental inbred B73, which has a pronounced erect-leaf habit. The leaf-orientation value below the ear (LOVB) also increased over eras. Spring vigor ratings increased, primarily for the 1970 era. This increase by the 1970 era for spring vigor rating also may have been caused by B73, which has been shown to have good spring seedling vigor (Mock and McNeill, 1979). M. Tollenaar (personal communication) noted that old hybrids actually had greater early vigor than did new hybrids. Tapper (1983) and Russell (1985) showed similar, significant improvements for plant health (stay-green). Tollenaar (1991) obtained average losses in stover dry weight, a reflection of plant health or stay-green, of 15.3, 1 1.3, and 8.4%for three old hybrids, four more recent hybrids, and one most recent hybrid, respectively. Hybrid X plant density interaction was not significant for loss of stover dry weight. Crosbie (1982) discussed the future role of physiology in maize breeding. To use selection of physiological traits to effect indirect increases for grain yield requires that the physiological traits have higher heritability; greater selection intensity and larger effective population size can be used; smaller number of generations or years to develop new inbred lines; and less costly selection procedures. Crosbie ( 1982) questions direct selection for yield ever being replaced by selection for physiological traits. Physiology can be used, however, to improve breeding efficiency by using such traits in the selection of source materials or in the development of breeding populations. Also, maize breeders select for a total phenotype when developing inbred lines;
GENETIC IMPROVEMENT OF MAIZE YIELDS
279
thus, it is important to know how various physiological traits contribute to grain yield and other agronomic traits. Dwyer and Tollenaar( 1989),in their study of eight hybrids in Ontario that represented the years 1959- 1988, found increasing trends for photosynthetic response, chlorophyllconcentration,crop growth rate, and LAI. These increases were similar to yield increases of the same hybrids (Tollenaar, 1989). Dwyer and Tollenaar suggested that it may be useful to consider physiological responses as selection criteria in future breeding programs. Perhaps it would be more logical to improve these traits in germplasm sources used for breeding purposes, rather than to select directly in an inbred development and evaluation program. Smith and White (1988) have given a comprehensive review of diseases occumng in maize. Selection for disease resistance has been an integral component of maize breeding for many years, yet there are few data reflecting directly on success for this selection relative to its effect on grain yield. Shurtleff (1980) estimated that yearly losses caused by maize diseases in the United States are 7 - 17%.Losses would be greater than this in most seasons if breeders had not made continuous improvements for resistance to diseases in parental inbred lines. Stay-green,which is primarily a rating for total plant health, has been evaluated in several experiments. The improvement for greater stalk strength has been well documented (Russell, 1974, 1984; Duvick, 1977, 1984),and stalk quality is dependent upon overall plant health. All experiments that have evaluated stay-green (Tapper, 1983; Duvick, 1984, 1991;Russell, 1985)have shown significant improvement for overall plant health from the early to the most recent decades. As new maize pathogens, or new races of existing species, have become important in the United States, plant pathologistsand breeders have reacted quickly so that more resistant hybrids have become available in a relatively short time. 2.mays L. is an extremely heterogeneousspecies;consequently, genes that condition resistance to plant pathogens usually can be found quickly. In some instances, single genes have been introduced into elite parental lines to give resistance to specific pathogenic species;e.g., Htl gives resistance to northern corn leaf blight [caused by Exserohifum turcicum (Pass.) Leonard and Suggs], rhm gives resistance to race 0 of southern corn leaf blight [caused by Bipofaris maydis (Nisis.) Shoem.], and R& gives resistance to races of common leaf rust (caused by Puccinia sorghi Schw.) in the United States. More generally, breeders have relied on polygenic sources of resistance as a genetic means to control diseases. During the 1950s and 1960s, breeders successfully developed cms-T (Texas-type, male-sterile cytoplasm) in many female parental lines, with pollen restorer genes in male parental lines to give pollen-fertile hybrids. This provided a very beneficial
280
W. A. RUSSELL
mechanism for hybrid seed producers. Adverse effects caused by cms-T and pollen restorer genes in single-cross hybrids were negligible, and effects may have been positive because of the nearly complete control for source of pollen in seed production fields. Suddenly, the effect of breeding was negative in 1970 because of the epiphytotic of southern corn leaf blight, race T, which infected any germplasm that had cms-T, caused an estimated yield loss of 15% in the United States. The problem was corrected quickly in subsequent years by discontinuing the use of cms-T in hybrid seed production. A review of the important insect species attacking maize has been given by Dicke and Guthrie (1988). Improvement in resistance and/or tolerantx of maize cultivars by breeding to some insect species has been reasonably successful in the United States. Probably one of the most successful has been the development of maize cultivars that have resistance to the European corn borer (Ostriniu nubilalis, Hubner). Resistance of U.S. germplasm to this pest occurs at two plant stages: resistance to whorl-leaf feeding (presilk stage, stage 2) and resistance post anthesis (stage 5). Resistances in the two plant stages have different genetic mechanisms. Resistance pre-anthesis is 1,4-bencaused primarily by DIMBOA (2,4-dihydroxy-7-methoxy-(2H)zoxazin-3(4H)-one) in leaf whorl tissue, but the cause for resistance during postanthesis is not known. Cooperative research by entomologists and breeders (U.S. Department of Agriculture Corn Insects Research Laboratory and Iowa State University) has developed several resistant inbred lines and breeding populations that have been released for breeding purposes (Russell, 1976; Russell and Guthrie, 1979,1982,1991). Although these cultivarshave not contributed directly to hybrid improvement, they can be used as breeding sources of resistance. Progress for resistance in one released breeding population, BS9, and effects of resistance to reduce yield losses are shown in Table X (Klenke et a!., 1986). Resistance was improved significantly for both first-generation (leaf-whorl tissue feeding) and second-generation (sheath and collar feeding, stalk cavities) corn borers. Cavity counts were reduced from 8.9 in BS9CO to 3.1 in BS9C4 with second-generation infestations, or from 8.0 to 3.4 for combined first- and second-generation infestations. Grain yield loss for control was 3 1.1% for BS9CO to BS9C4, but this loss was reduced to 10.0% for combined infestations of first and second generations. Compared with many cultivars of earlier eras, resistance or tolerance of modern hybrids to European corn borer is much improved, and progress will continue, probably more rapidly with the availability of better resistance sources of germplasm and of artificial techniques for rearing and infestation of the insect. Stay-green may also reflect improvement for resistance to second-generation European corn borer. Duvick (1984, 1991) showed a continuous improvement in the resistance to second-generation
Table X European Corn Borer Ratings and Grain Yield for BS9 and for Four Cycle Populations after S, Recurrent Selection for Firstand Second-Ceneration Resistanceu Treatmentsb FGI Populations BS9CO c1 c2 c3 c4
L.S.D. (0.05) Changd
SGI
cvc
FGI and SGI
ControlC (q ha-')
Rating (1 -9)=
Yield (q h - l )
Rating (1 -9)'
Yield (q ha-I)
Ratinpd (1 -9)'
Yield (q ha-I)
SGI
69.7 63.4 60.3 55.3 48.0 13.8 -31.1
3.6 3.6 2.8 2.5 2.7 0.8 - 25.0
60.7 65.0 52.4 56.2 47.9 13.8 -21.1
6.4 5.9 5.7 4.4 4.4 1.3 -31.2
53.1 47.5 44.4 40.8 44.0 13.8 -17.1
6.3 6.2 5.2 4.4 4.3 1.3 -31.8
49.0 48.7 48.7 47.8 44.1 13.8 - 10.0
8.9 8.7 5.7 3.9 3. I 2.2 -65.2
Hallauer et al. ( 1988); after Klenke et al. (1 986). FGI, First-generation infestations; SGI, second-generation infestation; CVC, cavity counts with one cavity = 2.5 cm. Populations were not infested and were sprayed with carbofuran to control natural infestation. Second-generationrating. Visual ratings on 1 (resistant) to 9 (susceptible)scale. fChange of C4 relative to CO. a From
FGI and SGI (no.) 8.0 7.1
5.0 4.8 3.4 2.2 - 57.5
282
W. A. RUSSELL
European corn borer from the 1930 to the 1980 hybrids, which would be conducive to greater stalk quality and reduced harvest loss.
VI. GENETIC GAINS VIA RECURRENT SELECTION IN POPULATIONS The primary purpose of recurrent selection is to improve a population for one or more traits and to maintain genetic variability for continued selection by intermatingprogenies for each cycle of selection. Recurrent selection and pedigree selection can complement each other in a comprehensive maizebreeding program. Genetic advance in hybrids depends on conducting systematic selection programs that develop improved germplasm sources for applied breeding programs. Continued genetic improvement of germplasm sources will contribute to realized genetic advance in lines and hybrids extracted by conventional plant-breeding methods. The effectiveness of recurrent selection procedures to improve various traits has been well documented (Eberhart et al., 1973;Russeil et al., 11973; Walejko and Russell, 1977;Moll et al., 1978;Stangland et al., 1982;Hallauer et al., 1983; Smith, 1983; Martin and Russell, 1984a,b;Darrah, 1985; Klenke et al., 1986;Nyhus et al., 1988,1989a,b;Helms et al., 1989;Homer et al., 1989;Odhiambo and Compton, 1989). A comprehensive review of recurrent selection methods and results was presented by Hallauer et al. ( 1988). Results from a few recurrent selection procedures will be reviewed to give an indication of how such programs can assist in genetic advance of commercial hybrid maize. Two maize synthetics, BSSS and BSCB 1, have been used in a long-term reciprocal recurrent selection program at Iowa State University (Eberhart et al., 1973;Helms et al., 1989).Each population has been used as tester oflines from the other population. Yield has been the primary selection trait, with the inclusion of selection for root and stalk strength and resistance to disease and insect pests. An evaluation study by Helms et al. (1989)compared performances for the original population cross and crosses of improved populations after 6 and 10cycles ofrecurrent selection (TableXI). There was significantimprovement for yield after both the sixth and tenth cycles. Root lodging was not reduced significantly, whereas stalk lodging was reduced significantly. Grain moisture was increased significantly for C6 X C6 and had a small decrease for C 10 X C 10; the increase relative to CO X CO was primarily a reflection of improved plant health. The cross of
GENETIC IMPROVEMENT OF MAIZE YIELDS
283
Table XI Means for Four Traits of Crosses of theOrigina1and Improved Populationsof BSSS and BSCBl Maize Evaluated in 10 Environments
Grain (q ha-')
Moisture (g kg-l)
Root
Crosses
Yield BSSSCO X BSCB 1CO BSSS(R)C6 X BSCBI(R)C6 BSS(R)C10 X BSCBI(R)CIO BSSSCO X Mo 17 BSSS(R)C6 X Mo I7 BSSS(R)ClO X Mo17 B73 X Mol7 L.S.D. (0.05)
51.3 63.8 73.0 67.5 75.7 84.8 85.5 5.8
236.9 253. I 245.4 250.1 260.9 262.8 255.3 12.8
16.8 16.6 13.7 11.4 7.1 9.2 12.3 6.5
Stalk
(%I 32.9 23.1 17.6 23.8 19.8 12.0 15.3 7.9
After Helms ef al. (1989).
BSSS(R)CI0 X Mo 17 yielded nearly equal to B73 X Mo 17and had less root and stalk lodging. Interpopulation crossesby using S, Lines from Cycles 0 and 9 of BSSS and BSCBl were evaluated at two locations in 1983 (0.S. Smith, unpublished observations). Although the materials were evaluated under heat and drought stress, frequency distributions for yields of 100 S, X S, crosses in each set (Fig. 14) show that important yield gains were made for C9 X C9 crosses. Some C9 X C9 crosses yielded as much as check single crosses, and there was still significant genetic variation, which indicates that further progress can be expected. BSSS has also been used in a long-term recurrent selection program in which the first seven cycles were based on performance of half-sib testcross progenies, then followed by evaluation based on Sz lines (Eberhart et al., 1973; Helms et al., 1989). The improved populations are identified as BS 13(S)CN. Walters et al. (199 1) used S, lines from BSSSCO, BSSS(R)C9, BS 13(S)C3,and C3 X C9 for an evaluation of lines per se and in crosseswith B73 and Mol7. Yield data summarized from their study are presented in Table XII. Each of the improved populations had a significant yield increase compared with CO, and C3 X C9 S, Lines had a significantly higher mean yield than did lines from C3 or C9. Testcrosses of the lines were significantly greater for all improved populations, with lines from C9 being the best. Moreover, the highest yielding individual crosses were from the C9 source.
284
W. A. RUSSELL
Grain Yield, q ha-'
Fig. 14. Frequency distributions of 100 crosses ofS3lines from BSSSCO and BSCB 1CO and of 100crossesofS,linesfromBSSS(R)C9andBSCBl(R)C9.(-),CO X CO,@ = 56.5 f 12.6; (--), c9 X C9, @ = 31.8 f 10.5.
In all instances, significant values for genetic variability indicated that further progress in yield improvement can be expected. Reciprocal full-sib recurrent selection was proposed as a breeding procedure that allows the integration of long-term and short-termobjectives(Hallauer, 1967, 1984; Hallauer and Eberhart, 1970). Progress after eight cycles of selection in two populations, BS 10 and BS 1 1, was presented by Eyherabide ( 1989).Eight cycles of selection were effective for increasing grain yield of BSlO (2.9%cycle-'), BSl 1 (1.6%cycle-'), and the population cross (6.5% cycle-') (Table XIII). Prolificacy increased, and there was improved resistance for root and stalk lodging. The inbreeding depression for grain yield of BSlOC8 (32.9%)andBSllC8 (37.9%)waslessthan that forBSlOCO(42.4%) and BSllCO (56.1%). Further evidence for yield improvement by recurrent selection in BSSS is shown by comparison of hybrid performances of inbred lines developed from the original and improved cycles (Table XIV). Trait mean values are based on agronomic data from machine-pianted and machine-harvested trials conducted at four Iowa locationsduring 1976- 1985(37 location - year environments). InbredsB14 and B37, which were released in 1953and 1958,
Table Xn Mean Yields over 50 S , Progenies and B73 and Mo17 Testcrossesof the S, Progenies, Ranges of Yields, and Estimated Genetic Varianceab B73 testcrosses (q ha-')
S, lines (q ha-')
Mo 17 testcrosses (q ha-')
Population
Yield
Range
ij?
Yield
Range
%
Yield
Range
%
BSSSCO BSI 3(S)C3 BSSS(R)C9 c 3x c9 L.S.D. (0.05)
29.2 39.2 39.1 41.9
16.8-44.4 28.6 -49.0 25.7-53.7 29.6-55.4
29 18 23 17
59.9 64.2 72.4 66.9
50.5-71.7 58.5-73.2 63.1-82.1 56.6-73.4
38 7 17 24
61.8 69.2 74.6 10.2
48.9-71.7 55.7-81.5 64.0-85.9 61.9-80.1
38 58 27 35
1.2
1.1
'S,, lines from BSSSCO and three BSSS populationsimproved for yield by recurrent selection. After Waiters el UZ. (1991). Estimated genetic variance.
1.1
W. A. RUSSELL
286
Table XlII Means for Grain Yield, Ears Plant-', and Root and Stak Lodging for BSlO and BSll Populations and Their Crossesa6 Trait Lodging Populations and crosses
BS1OCO c2 c4 C6 C8
BSl ICO c2 c4 C6 C8 CQXCO c2xc2 c 4x c4 C6 X C6 C8 X C8 L.S.D. (0.05)
Ears plant-'
Root
(q ha-')
41.5 43.8 50.9 52.4 51.3 49.2 51.5 48.6 54.7 55.6 46.5 51.2 59.6 63.6 74.6 4.4
0.83 1.01 1.08 1.07 1.11 0.93 0.97 1.02 1.09 1.12 0.90 1.10 1.10 1.14 1.20 0.07
5.2 1.7 3.5 5.1 2.6 10.7 8.8 7.5 3.8 1.9 6.5 2.1 3.2 4.9 3.6 3.1
Yield
Stalk
(96) 20.7 24.0 18.8 13.4 16.1 22.0 15.5 15.6 11.7 12.9 22.9 19.6 18.3 15.4 12.5 5.3
Data obtained in eight environments. 'After Eyherabide (1989).
respectively, came from the original BSSS population. Inbred B37 has more genetic potential than B14 to respond to higher plant densities and fertility levels. Inbred B73 evolved from the C5 population of BSSS, which is the population developed after five cycles of recurrent selection for yield improvement with a double-cross tester, and in combination with Mo17, B73 has contributed a yield gain of 10.0%compared with B37. Furthermore, inbred B84, which evolved from the C7 population of BSSS, has contributed a yield gain of 12.3%over B73. Periodically, surveyshave been conducted to determine the use of public lines by the hybrid corn seed industry. A summary by Zuber and Dmah ( 1980) showed that the commercial use of B37 as a percentage of the total U.S. hybrid seed requirement was 25.7%in 1970. Subsequently, the use of B37 had decreased to 6.8% in 1975 and 2.4% in 1979. Inbred B73 was released in 1972, and the commercial use was esti-
GENETIC IMPROVEMENT OF MAIZE YIELDS
287
Table XIV Agronomic Data for Four single Crosses Evaluated in 37 h t i o n - Y e a r Environments(1976 - 1985) in Iowa'
Cross
Yieldb (q ha-9
B14A X Mo17 B37 X Mo17 B73 X Mo17 B84 X Mo17 L.S.D. (0.05)
72.8 77.4 85.1 95.6 3.8
Grain moisture
Lodging Root Stalk
Dropped ears
(%)
(%I
(%)
20.6 22.6 22.4 22.9 0.5
11.4 16.4 15.2 12.1 4.8
7.6 16.6 9.8 10.8 3.8
1 .o
1.2 1.7 0.8
From Hallauer et al. (1 988). Yield B84 > B73 > B37 > B14 = 12.3, 10.0, and 6.3%, respectively.
matedat 3.1%in 1975 and 16.19in 1979. InbredB84wasreleasedin 1978, but has not become widely used, primarily because it has not contributed satisfactory root and stalk strength to commercial hybrids. Recurrent selectionprocedures have been used very effectivelyto improve agronomic traits, but yield has not been one of the selection criteria (Thompson, 1982;Devey and Russell, 1983;Colbert etal., 1984;Martin and Russell, 1984a,b; Johnson et al., 1986; Klenke et al., 1986; Nyhus et al., 1989a,b).Martin and Russell (1984a,b)used recurrent selection based on S, line evaluations for three cycles to improve BS 1 maize synthetic for resistance to artificial stalk rot and resistance to mechanical breakage. They evaluated 100 S, lines from the CO and C3 populations for artificial stalk rot and resistance to mechanical breakage. Frequency distributions show that, with both selection procedures, highly significant improvements were achieved for both traits; genetic variance was decreased for stalk rot ratings, but was increased for stalk strength (Figs. 15 and 16). Evaluations of the populations per se,CO to C3, crosses of the populations, and testcrosses of these populations showed significant improvement for resistance to natural field stalk rot and field stalk lodging. Grain yield, however, decreased significantly for the populations per se and for the crosses of the two sets of improved populations, and showed no change for the testcrosses. In other recurrent selection studies in which selection criteria have been for agronomic traits that have not included yield, highly significant gains have been achieved for the selected traits, but usually yield disadvantages
288
W. A. RUSSELL 35
'
I
-BSlCO
6J=42.50+7.15
Fig. 15. Frequency distributions for stalk strength of 100 S,lines from BSlCO, BSI MSC3, and BS lSRC3 (arrowsidentify population means;class interval = standard error of the difference between two means = 4.5 kg) (Martin and Russell, 1984a). (-), BSCICO, @ = 42.50 & 7.15; (---), BSlSRC3, a"c = 48.73 f 8.52; (--), BSlMSC3, = 73.86 f 12.66.
a:
have also been observed (Thompson, 1982; Devey and Russell, 1983; Klenke et al., 1986; Nyhus et al., 1989a,b).An exception has been reported by Colbert et al. ( 1984),who found highly significantgains for stalk crushing strength of two maize synthetics, and grain yield showed nonsignificant increases. Also, Johnson et al. (1986) used full-sib recurrent selection in the cultivar "Tuxpeno Crema I" to reduce plant height, but there was no selection for grain yield. Subsequently, Cycles 0, 6, 9, 12, and 15 were evaluated to determine changes that occurred for several traits (Table XV). Plant height decreased 2.4% cycle-', caused by reduced total node number and internode length below the ear. Ear height was reduced near 50%.Reductions for plant and ear heights were probably important factors in great reduction of lodged plants. Although grain yield was not a selection criterion, yield had a highly significantincrease at 4.4% cycle-'. Harvest index and ears per plant also had
GENETIC IMPROVEMENT OF MAIZE YIELDS
40
-
289
I \ - - BSlSRC3 6'Q --0 .18t0.05 I \ ...'.......'BS1 MSC3 6' -0 14t0.03-
Stalk Rot Rating
Fig. 16. Frequency distributions for stalk-rot ratings of 100 S, lines from BSICO, BS I MSC3, and BS I SRC3 (arrowsidentify populationmeans;class interval = standarderror of the difference between two means = 0.4) (Martin and Russell, 1984a). (-), BSICO, 3 = 0.36 f 0.07; (---), BSlSRC3, 3 = 0 . 1 8 f0.05;( * * * * ) , BSlMSC3,@=0.14 fO.03.
highly significant increases. Leaf-area index decreased significantly at 50,000 plants ha-', but had no significant change at optimal plant density (data not shown). Consequently, grain yield/leaf area increased from 85 g m-z at Cycle 0 to 225 g m-z at Cycle 15, or 9.5% cycle-'. Tuxpeno Creme I, Cycle 15, can be expected to be a much improved source for inbred lines compared with Cycle 0. Causes for grain yield reductions may have been inbreeding effects, change in partitioning of photosynthate when selection has been for improved stalk quality, and pleiotropy. Inbreeding effects are caused partly by the restricted number of lines selected for recombination in each cycle of recurrent selection. In selection for stalk-rot resistance and stalk strength, partitioning of carbohydrates may have been changed so that a greater amount was used for stalk development and a lesser amount for ear development. Pleiotropywould be a causativefactor for yield decreasesif some genes that condition selected traits also contribute to lower grain yields. It seems that recurrent selection procedures should be an integral part of a maize breeding program as a means to continue improvements of source
290
W. A. RUSSELL Table XV
Grain Yield and Agronomic Traits
as Mected by Cycles of Full-Sib Recurrent Selection for Reduced Height in Tuxpeno Crema I"
Cycle
Plant height* (cm)
Grain yield* (q ha-')
0 6 9 12 15 L.S.D. Change cycle-I (%)
282 219 21 1 202 179 22 -2.4**
31.7 42.9 44.8 49.3 54.0 3.0 4.4**
Harvest index'
Ears plant-1d
0.30 0.40 0.40 0.4 1
0.70 0.87 0.90 0.93 0.98 0.12 2.5**
0.45
0.04 3.1**
Lodgin8
(96)
43 12 14 9 5
6.7**
**
Significant at p = 0.0 1. "AAer Johnson et al. (1986). Data from three locations in two years; grain yield and lodging at optimal plant density. 'Data from two locations in two years. dData from two locations, measured at near optimal plant density.
*
populations. For traits that are relatively simply inherited, highly significant improvementscan be achieved in few cycles of selection. It seems, however, that yield should be either included in a selection index, when the primary purpose is to improve certain agronomic traits, or monitored so that yield decreases are avoided. Yield is improved slowly, and a program whose primary purpose is yield improvement must be considered long term. Multiple trait selection can be used, but rate of gain for any one trait w illbe less than in a program of single-trait selection.
VII. IMPROVEMENT OF INBRED LINES The replacement of double crosses by single crosses in the United States, which began about 1960, has been an important item in yield gains. This effect of the single crosses is evident in Fig. 2, which shows an accelerated yield increase, beginning about 1960. The best single crosses are expected to yield more than the best double crosses, except with certain types of epistatic
GENETIC IMPROVEMENT OF MAIZE YIELDS
29 1
gene action. Also, it is much easier to identify two inbred lines that combine to give a high-performance single cross than to identify four inbred lines that combine to give a high-performance double cross. Few single crosseswere used before 1960because the production of singlecross seed was not economically feasible for the seed industry. The situation was changed, however, with the gradual improvement of the inbred parents for seed yield and the adoption of better cultural practices. Maize breeders and hybrid seed producers are well aware of the general improvement of the inbred parental lines, but this improvement is less well documented than for hybrids. Agronomic improvements include: cold tolerance such that germination and emergence are better, resistance to plant diseases, resistance to insects, resistance to barrenness, resistance to root and stalk lodging, better pollen production, higher seed yield, and better seed quality. Duvick (1984) compared field performances for five inbreds per decade for five decades, 1930- 1970. The inbreds showed a yield gain of 0.50 q ha-' yr-', or a total gain of 20 q ha-'. Significant gains were also observed for ears per 100 plants and for resistance to root and stalk lodging. Meghji et al. (1984), using the inbred lines that Russell (1974) had used earlier, evaluated the agronomic performance of the lines. The inbreds did not show any yield improvement for the 1950-compared with 1930-erainbreds, but the 1970-era inbreds had an average yield increase of 14.5% over the 1950 lines. Furthermore, the 1970 inbreds were much improved for plant health, so that the grain-mng period was extended several days, and root and stalk lodging was much less compared with the 1930 lines. Whereas in earlier years breeders selected lines primarily on the basis of hybrid performance, now much more emphasis is given to the development of lines with greater vigor and higher seed yields. The Iowa State University maize breeding project has a collection of inbred lines that date back to the first lines developed in the 1920s. This collection, which includes lines developed from 1920sto 1980s, was grown for reproduction of seed in 1988. Summer 1988 was characterizedby severe stressduring much of the growing season caused by below-averageprecipitation and above-average temperatures. The superiority of lines developed since about 1950was very evident by such traits as greater plant vigor, better pollen production and silk emergence, much better stay-green, larger ears, and higher grain yields. Many of the older lines barely produced enough seed for cold-storage samples. This observation showed the problems that seed producers must have experienced in some earlier years to obtain single-cross seed for use in the production of double-cross seed. It was evident why single-cross hybrids were not used commer~allyto a great extent before 1960.
292
W. A. RUSSELL IL "
Master's Winners, 1954-1989
\i = 94.6 + 1.50(X-1954) r 2 = 0.87
Mean = 120.8
90
50vt I
[owa Corn Yields, 1954-1989 Y = 37.1 + 1.20(X-1954) r 2 = 0.70 Mean = 58.0
40
30f
'
20 I 1954
I
I
I
I
1960
1966
1972 Year
I I
1978
I
I
1984
5o
'
40 1990
Fig. 17. Observed yields in Iowa and predictedlinear responsesfor state average and master corn grower winners, 1954- 1989.
VIII. FUTURE TRENDS Grain yields for maize are expected to increase because of the combined effects of two items: (1) improved cultural practices and management, and (2) increased genetic potential of the hybrids. An interestingcomparison for yield increases in Iowa is shown in Fig. 17. The lower line shows the state average yield during 1954 - 1989, and the upper line showsaverage yields for the master maize-champion contest. The rates of yield increase have been 1.20 and 1.50 q ha-; yr-; for the state average and the master contest, respectively. Projected to the year 2000, the state average is 92.3 q ha-; and the master contest is 165.4 q ha-'. On the basis of predicted yields, the difference between the two sets was 57.5 q ha-' in 1954, increased to 69.4 q ha-' in 1989, and is predicted at 73.1 q ha-* in 2000. In both instances, the yield levels have continued to increase because of the combined effects of better hybrids, improved cultural practices, and better management.Further improvements of these items can be expected to contribute significantly to give higher state averages in the future. It is conceivable that the rate of gain for U.S. grain yield could become even greater than presented here (Fig. 2) because there are more breeders and
GENETIC IMPROVEMENT OF MAIZE YIELDS
293
more companies with breeding programs now than for any previous time in the United States. The evaluations by Russell (1984) and Duvick (1984) showed that an important cause for yield gain is the genetic potential of the most recent hybrids to take advantage of higher plant densities to produce more grain per unit field area. Perhaps further gains can be achieved by the development and evaluation of parental materials at higher plant densities to identify greater resistance to barrenness. Stalk lodging, however, is still a major problem and must be corrected if higher plant densities are to be used. Consequently, breeders must develop materials that have better disease resistance and stalk quality before hybrids can be planted at higher plant densities to produce more grain per unit area. Recurrent selection can be used effectively to improve stalk quality in breeding populations, but caution must be used in selection to avoid yield declines (Devey and Russell, 1983; Martin and Russell, 1984a,b). We must be aware that greater problems with plant pests may occur at any time (e.g., maize leaf epiphytoticof 1970)thus causing difficulty for breeders even to maintain the yield level now achieved. Occasionally, germplasm with resistance for some pests may have to be introduced from unadapted sources, and this could cause temporary reductions for hybrid performance until unfavorable genes linked with the resistance genes can be eliminated. We are only beginning to observe the potential for various types of recurrent selection to increase gene frequencies of favorable alleles in breeding populations. These improved populations will be the sources for new inbred lines transmittingto hybrids greater yield potential and favorable changesof other traits for which the populations have been improved. Also, we have onlyjust begun to explore in greater depth the possible utilization of germplasm from other areas such as Mexico and South America. Currently, detasseling in single-crossseed production fields has become a primary problem in the U.S. Corn Belt. Detasseling by hand is a labor-intensive operation, and labor shortages occur every year. Detasseling by machine, either by cutting or pulling tassels, is used but still requires subsequent hand-pulling to be certain that all female plants are detasseled. The Texas (T)-type, male-sterile cytoplasm (cms-T), which was used for more than 10 years to 1970, was successful because of the stability for male sterility and ease with which pollen restoration was achieved in male parental lines. After cms-T had to be discontinued because of its high susceptibilityto southern corn leaf blight [incited by Bipolaris maydis (Nisik.) Shoem., race TI, several other male-sterile types have been used but none as successfully as cms-T. Breederscan assist in relieving the problem by continuing to develop higheryielding parental lines, thus reducing the area required for seed production fields. If seed producers could use parental lines that are less highly inbred (Carlone and Russell, 1988, 1989), higher seed yield per unit area would
294
W. A. RUSSELL
occur. Perhaps breeders and molecular biologists can cooperatein the development of a procedure that will give a male sterility not likely to succumb to new pests. Also, the system would have to be readily amenable to pollen restoration. Apomixis has not been reported in maize, but has been noted in Tripsacurn ductyloides (Bashaw, 197l), which can be crossed with maize. Bashaw ( 1971) outlined a plan for use of apomixis in sorghum breeding and hybrid seed production. Bashaw (1 980) and Hanna and Bashaw ( 1987)have discussed the use of apomixis in plant breeding and hybrid seed production. It seems apomixis would be usable with maize if the right kind could be found in maize or transferred from another species. If such a system for hybrid seed were to come into use for maize, patents and protective laws would have to be relied on to assure seed companies exclusive rights to their apomictic hybrids. The role that plant biotechnology will have in maize improvement in the future can only be speculated at this time. “Genetic engineering” will not become a substitute for plant breeding; instead, the two scienceswill become an integrated program for cultivar improvement. Presently, plant biotechnology will become most useful to effect improvements for traits that are simply inherited, i.e., resistance to diseases and insects, resistance to herbicides, nutrient qualities, maturity, salt tolerance, cold tolerance, and others. Transfer to maize of genes from related species, a difficult procedure with standard genetic methods, may become much easier via genetic engineering. As techniques for genetic engineering are further developed, they may become used for traits that have a more complex inheritance. Primarily, it seems that biotechnology and plant breeding will become a more efficient total program and will reduce time and expense of field programs. Although the rate of gain for maize yield in the United States may decrease, compared with the past three decades, continually increasing maize yields for the future seem promising.
REFERENCES Bashaw, E. C. 1971. Apomoxis and its potential in sorghum breeding. Proc. Annu. Corn Sorghum Ind. Res. Conf: 26,54-59. Bashaw, E. C. 1980. Apomoxis and its application in crop improvement. In “Hybridization of Crop Plants” (W. R. Fehr and H. H. Hadley, eds.), pp. 45 -63. Am. SOC. Agron. Crop Sci. SOC. Am., Madison,Wisconsin. Brown, W. L. 1975. A broader germplasm base in corn and sorghum. Proc. Annu. Corn Sorghum Ind. Res. Conf: 30,8 I - 89. Bundy, L. G., and Carter, P. R. 1988. Corn hybrid responseto nitrogen fertilizer in the Northern corn belt. J. Prod. Apic. 1,99- 104.
GENETIC IMPROVEMENT OF MAIZE YIELDS
29 5
Cardwell, V. B. 1982. Fifty years of Minnesota corn production: Sources of yield increase. Agron. J. 74,984-990. Carlone, M. R. 1985. The response to three plant densities and four nitrogen levels for maize cultivars representingseven eras ofbreeding. M.S. Thesis, Parks Library, Iowa State Univ., Ames. Carlone, M. R., and Russell, W. A. 1987. Responseto plant densitiesand nitrogen levelsfor four maize cultivars from different eras of breeding. Crop Sci. 27,465 -470. Carlone, M. R., and Russell, W. A. 1988. Evaluation of S2 maize lines reproduced for several generations by random mating within lines. I. Comparisons between the original and maintained S2lines. CropSci. 28,916-920. Carlone, M. R., and Russell, W. A. 1989. Evaluation of S2maize lines reproduced for several generations by random mating within lines. 11. Comparisons for testcross performance of original and advanced S, and S, lines. Crop Sci. 29,899-904. Castlebeny, R. M., Crum, C. W., and Krull, C. F. 1984. Genetic yield improvement of U.S. maize cultivars under varying fertility and climatic environments. Crop Sci. 24,33-36. Colbert, T. R., Darrah, L. L., and Zuber, M. S. 1984. Effect of recurrent selection for stalk crushing strength on agronomic characteristicsand soluble stalk solids in maize. Crop Sci. 24,473-478.
Crabb, A. R. 1947. “The Hybrid-Corn Makers: Prophets of Plenty.” Rutgers Univ. Press, New Brunswick, New Jersey. Crosbie, T. M. 1982. Changes in physiologicaltraits associated with long-term breeding efforts to improve grain yield of maize. Proc. Annu. Corn Sorghum Ind. Res. Con& 37,206-223. Darrah, L. L. 1985. Evaluation of population improvement in the Kenya maize breeding methods study. To Feed Ourselves,Proc. East. Central South. Ajk Reg. Maize Workshop pp. 160-176. Derieux, M., Darrigand, M., Gallais, A., Barriere, Y.,Bloc, O., and Montalant, Y. 1987. Estimation du progies gknktique &lid chez le mais grain en France entre 1950 et 1985. Agronomie 7, 1 - 1 I. Devey, M. E., and Russell, W. A. 1983. Evaluation of recurrent selection for stalk quality in a maize cultivar and effects on other agronomic traits. Iowa State J. Res 58,207-2 19. Dicke, F. F., and Guthrie, W. D. 1988. The most important corn insects. I n “Corn and Corn Improvement” (G. F. Sprague and J. W. Dudley, eds.), 3rd Ed., pp. 767-867. Am. Soc. Agron., Madison, Wisconsin. Duncan, W. G. 1958. The relation between corn population and yield. Agron. J. 50, 82-84. Duvick, D. N. 1977. Genetic rates ofgain in hybrid maize during the last 40years. Maydica 22, 187- 196.
Duvick, D. N. 1984. Genetic contributions to yield gains0fU.S. hybrid maize, 1930 to 1980. I n “Genetic Contributions to Yield Gains of Five Major Crop Plants” (W. R. Fehr, ed.), Spec. Publ. No. 7, pp. 15-47. Crop Sci. SOC.Am., Madison, Wisconsin. Duvick, D. N. 199 1. Genetic contributions to advances in yield of U.S. maize. Maydica 36, in press. Dwyer, L. M., and Tollenaar, M. 1989. Genetic improvement in photosynthetic response of hybrid maize cultivars, 1959 to 1988. Can. J. Plunt Sci. 69, 81-91. East, E. M. 1908. Inbreeding in corn. Rep. Conn. Agric. Exp. Stn., 1907 pp. 419-428. Eberhart, S. A., and Russell, W. A. 1966. Stability parametemfor comparing varieties. Crop Sci. 6, 36-40.
Eberhart, S. A., Debela, S.,and Hallauer, A. R. 1973. Reciprocal recurrent selection in BSSS and BSCBI maize varieties and half sib selection in BSSS. Crop Sci.13,451 -456, Eyherabide, G. H. 1989. Response to eight cycles of full-sib reciprocal recurrent selection for
296
W. A. RUSSELL
grain yield and standabilityin BS 10and BS 11 maize populations. Ph.D. Thesis, Iowa State Univ., Ames. Fakorede, M. A. B., and Mock, J. J. 1982. Correlated responses to recurrent selection for grain yield in maize. Iowa Agric. Home Econ. Exp. Stn., Res. Bull. No. 596, pp. 179-206. Frey, K. J. 197 1. Improving crop yields through crop breeding.Am. Soc. Agron. Spec. Publ. No. 20, pp. 15-58. Gardner, C. A. C., Bax, P. L., Bailey, D. J., piper, T. E., Segebart, R. L., Smith,0.S., Tiffany, C. W., Trimble, M. W., and Wilson, B. N. 1990. Response of corn hybrids to nitrogen fertilizer. J. Prod. Agric. 3, 39-43. Gardner, C. 0. 1961. An evaluation of effects of mass selection and seed irradiation with thermal neutrons on yield of corn. Crop Sci. 1,241 -245. Goodman, M. M. 1985. Exotic maize germplasm:Status,prospects, and remedies. Iowa State J. Rex 59,497-527. Goodman, M. M. 1990.Genetic and germplasm stocksworth conserving. J. Hered. 8 1 , l l - 16. Hallauer, A. R. 1967. Development of singlecross hybrids from two-eared maize populations. Crop Sci. 7, 192- 195. Hallauer, A. R. 1984. Reciprocal full-sib selection in maize. Crop Sci. 24,755-759. Hallauer, A. R., and Eberhart, S. A. 1970. Reciprocal full-sib selection. Crop Sci. 10,3 15 - 3 16. Hallauer, A. R., Russell, W. A., and Smith,0. S. 1983. Quantitative analysis of Iowa Stiff Stalk Synthetic. Stadler Genet. Symp. 15, 83- 104. Hallauer, A. R., Russell, W. A., and Lamkey, K. R. 1988. Corn breeding. In “Corn and Corn Improvement” (G. F. Sprague and J. W. Dudley, eds.), 3rd Ed., pp. 469-564. Am. SOC. Agron., Madison, Wisconsin. Hanna, W. W., and Bashaw, E. C. 1987. Apomixis: Its identificationand use in plant breeding. Crop Sci. 27, I 136 - 1 139. Hayes, H. K. 1963. “A Professor’sStory of Hybrid Corn.” Burgess, Minneapolis, Minnesota. Helms, T. C., Hallauer, A. R., and Smith, 0.S. 1989. Genetic drill and selectionevaluated from recurrent selection programs in make. Crop Sci. 29,602-607. Homer, E. S., Magliore, E., and Morera, J. A. 1989.Comparison of selectionfor S, progeny vs. testcross performance for population improvement in maize. Crop Sci. 29,868 - 874. Johnson, E. C., Fischer, K. S., Edmeades, G. O., and Palmer, A. F. E. 1986. Recurrent selection for reduced plant height in lowland tropical maize. Crop Sci. 26,253-260. Jones, D. F. 1918.The effects ofinbreeding and crowbreeding upon development. Conn.Agric. Exp. Stn. Bull. No. 207. Jones, D. F. 19 19. Inbreeding in corn improvement. Breeders Gaz. 75, 1I 1 1 - 1 1 12. Kamprath, E. J., Moll, R. H., and Rodriguez, N. 1982. Effects of nitrogen fertilization and recurrent selection on performance of hybrid populations of corn. Agron. J. 74,955-958. Klenke, J. R., Russell, W. A., and Guthrie, W. D. 1986. Recurrent selection for resistance to European corn borer in a corn synthetic and correlated effects on agronomic traits. Crop Sci. 26, 864-868. Kojic, L. 1990. Maize improvement in Yugoslavia and Eastern European countries. Natl. Maize Con&. 2nd- Rex. Econ., Environ.. Grado, Italy. Kuhn, H. C., and Gevers, H. 0. 1980. Hybrid improvement in the past three decades. Proc. South Afr. Maize Breed. Symp., 4th, (H. 0. Gevers and J. G. Duplessis, eds.), Tech. Commun. No. 172, pp. 69-73. Dep. Agric. Fish., Repub. South Afr. Lamkey, K. R., and Smith, 0.S. 1987. Performance and inbreeding depression of populations representing seven eras of maize breeding. Crop Sci. 27,695-699. Lonnquist, J. H. 1964. A modification of the ear-to-row procedure for the improvement of maize populations. Crop Sci. 4,227-228.
GENETIC IMPROVEMENT OF MAIZE YIELDS
297
Martin, M. J., and Russell, W. A. 1984a.Response ofa maize syntheticto recurrent selection for stalk quality. Crop Sci. 24,331 -337. Martin, M. J., and Russell, W. A. 1984b. Correlated responses of yield and other agronomic traits to recurrent selection for stalk quality in a maize (Zeu mays L.) synthetic. Crop Sci. 24,746 - 750. Meghji, M. R., Dudley, J. W., Lambert, R. J., and Sprague, G. F. 1984. Inbreeding depression, inbred and hybrid grain yields, and other traits of maize representingthree eras. Crop Sci. 24,545 - 549. Misevic, D., Javandic, N., Dumanovic, J., Kojic, L., Greder, R. R., Vrebalov, J., and Misevic, M. 1986.Geneticratesofgaininhybridmaize yieldsin Yugoslaviainperiod 1966- 1984. I. FA0 maturity 500-700. Genefika 18, 181-204. Misevic, D., Javandic, N., Dumanovic, J., Kojic, L., Greder, R. R., Vrebalov, J., and Misevic, M. 1987a.Genetic rates ofgain in hybrid maize yields in Yugoslaviain period 1966- 1984. 11. FA0 maturity 200-400. Genefika19, 103- 119. Misevic, D., Vidakovic, M., Parlov, D., Radic, J., and Vekic, N. 1987b. Maize breeding at the beginning of 21st century. Genetika 19,263-273. Mock, J. J., and McNeill, M. J. 1979. Cold tolerance of maize inbred lines adapted to various latitudes in North America. Crop Sci. 19,239-242. Moll, R. H., Cockerham, C. C., Stuber, C. W., and Williams, W. P. 1978. Selection responses, genetic-environmental interactions, and heterosis with recurrent selection for yield in maize. Crop Sci. 18, 641 -645. Muruli, B. J., and Paulsen, G. M. 1981. Improvement of nitrogen use efficiency and its relationship to other traits in maize. Muydicu 26,63-73. Nyhus, K. A., Russell, W. A., and Guthrie, W. D. 1988. Response of two maize synthetics to recurrent selection for resistance to first-generation European corn borer (Lepid0ptera:Pyddae) and Diplodia stalk rot. J. Econ. Entomol. 81, 1792- 1798. Nyhus, K. A., Russell, W. A., and Guthrie, W. D. 1989a. Distributions among S, lines for European corn borer (Lepidoptera:Pyralidae) and stalk rot resistance ratings in two maize synthetics improved by recurrent selection. J. Econ. Entomol. 82,239-245. Nyhus, K. A., Russell, W. A., and Guthrie, W. D. 1989b.Changes in agronomic traitsassociated with recurrent selection in two maize synthetics. Crop Sci. 29,269-275. Odhiambo, M. D., and Compton, W. A. 1989. Five cyclesof replicated S, vs. reciprocal full-sib index selection in maize. Crop Sci. 29,3 14-3 19. Paterniani, E. 1990. Maize breeding in the tropics. In “Critical Reviews in Plant Sciences 11” (B. G. Conger, ed.), pp. 125- 154. CRC Press, Boca Raton, Florida. Russell, W. A. 1974. Comparative performance for maize hybrids representingdifferenteras of maize breeding. Proc. Annu. Corn Sorghum Ind. Res. Conf 29, 8 1 - 101. Russell, W. A. 1976. Registration of B75 germplasm line ofmake (Zen mays L.) (Reg. No. GP 62). Crop Sci. 16, 3 15. Russell, W. A. 1984. Agronomic performance of maize cultivars representing different eras of maize breeding. Muydicu 29,375 - 390. Russell, W. A. 1985. Evaluations for plant, ear, and grain traits of maize cdtivars representing seven eras of breeding. Muydicu 30,85 -96. Russell, W. A. 1986. Contribution of breeding to maize improvement in the United States, 1920s-1980s. Iowa SfufeJ.Res. 61, 5-34. Russell, W. A., and Balko, L. G. 1980. Response of corn inbred lines and single crosses to nitrogen fertilizer. Proc. Annu. Corn Sorghum Ind. Rex Conf 35,48-67. Russell, W. A,, and Guthrie, W. D. 1979. Registration of B85 and B86 germplasm lines ofmaize (Reg. Nos. GP 76 and GP 77). Crop Sci. 19,565.
298
W. A. RUSSELL
Russell, W. A., and Guthrie, W. D. 1982. Registration of BS9(CB)C4 maize germplasm. Crop Sci. 22, 694. Russell, W. A., and Guthrie, W. D. 1991. Registration ofBS17(CB)C4and BS16(CB)C4. Crop Sci. 31,238-239. Russell, W. A., and Machado, V. 1978. Selection procedures in the development of m&e inbred lines and the effects of plant densitieson the relationshipsbetween inbred traits and hybrid yields. Iowa Agric. Home Econ. Exp. Stn. Res. Bull. No. 585. Russell, W.A., and Teich, A. H. 1967. Selection in Zeu mays L. by inbred line appearance and testcross performance in low and high plant densities. Iowa Agric. Home Econ. Exp. 9 n . Res. Bull. No. 552. Russell, W. A., Eberhart, S. A., and Vega 0, Urbana A. 1973. Recurrent selection for specific combining ability for yield in two maize populations. Crop Sci. 13,257 -26 I. Shull, G. H. 1908. The composition ofa field ofmaize. Rep. Am. Breeders’Assoc.4,296-301. Shull, G. H. 1909. A pure-line method in corn breeding. Rep. Am. Breeders’Assoc. 5,5 1-59. Shull, G. H. 1910. Hybridization methods in corn breeding. Am. Breeders’Mug. 1,98- 107. Shurtleff, M. C. 1980. “Compendium of Corn Diseases,” 2nd Ed. Am Phytopathol. Soc., St. Paul, Minnesota. Smith, D. R., and White, D. G. 1988. Diseases ofcorn. In “Cornand Corn Improvement”(G. F. Sprague and J. W. Dudley, eds.),3rd Ed., pp. 687-766. Am. Soc. Agron., Madison, Wisconsin. Smith, 0. S. 1983. Evaluation of recurrent selectionin BSSS, BSCBI, and BS 13 maize populations. Crop Sci. 23,35-40. Stangland, G. R., Russell, W. A., and Smith, 0.S. 1982. Agronomic evaluation of four make synthetics and their crosses after recurrent selection for yield. Muydicu 27, 199-2 12. Steel, R. G. D., and Tome, J. H. 1980. “Principles and Procedures of Statistics. A Biometrical Approach,” 2nd Ed. McGraw-Hill, New York. Tapper, D. C. 1983. Changes in physiologicaltraits assoCiated with grain yield improvement in single-cross maize hybrids from 1930to 1970. Ph.D. Thesis,Iowa State Univ., Ames. (Diss. Absfr. 83-DA8316164.) Thompson, L. N. 1975. Weather variability, climatic change, and grain production. Science 188,535-541. Thompson, L. N. 1982. “Weather and Technology Trend in Corn Yields. Better Crops with Plant Food.” Potash Phosphate Inst., Atlanta, Georgia. Tollenaar, M. 1989. Genetic improvement in grain yield ofcommercial maize hybridsgrown in Ontario from 1959 to 1988. CropSci. 29, 1365-1371. Tollenaar, M. 1991. Physiological basis of genetic improvement of maize hybrids in Ontario from 1959 to 1988. Crop Sci. 31, 119- 124. Trifunovic, V. 1978. Maize production and maize breeding in Europe. In “Maize Breedingand Genetics” (D. B. Walden, ed.),pp. 41 -58. Wiley, New York. Tsai, C. Y., Huber, D. M., Glover, D. H., and Warren, H. L. 1984. Relationship ofN deposition to grain yield and N response of three maize hybrids. Crop Sci. 24,277-28 1. U.S. Department of Agriculture. 1930- 1989. “Agricultural Statistics.” U.S. Gov. Print. OK, Washington, D.C. Walejko, R. N., and Russell, W. A. 1977. Evaluation ofrecurrent selectionfor combining abirity in two open-pollinated maize cultivars. Crop Sci. 17,647 -65 1. Walters, S. P., Russell, W. A., and Lamkey, K. R. 1991. Performance and genetic variance among S, lines and testcrosses of Iowa Stiff Stalk synthetic of maize. Crop Sci. 31,76- 80. Zuber, M. S., and Darrah, L. L. 1980. 1979 U.S. corn germplasm base. Roc. Annu. Corn Sorghum Ind. Res. Corrj35234-249.
Index A Acidity, effects on zero point of charge, 223 Acid washing, effects on point of zero salt effect, 224 Adaptation for earlier maturity, maize, 247 Adsorbates, nuclear magnetic resonance S ~ ~ ~ ~ ~ O144 S C- O PY, 146 Adsorption, potentialdetermining ions, 204 Agricultural genetics, DNA markers and, 40-41 Albite, NMR studies of, 145 Angular momentum, 94,98 Angular velocity, Larmor, 95 Anion exchange capacity, 202 Anisotropic local symmetry, 128 Apedal soil materials, 2 - 3 Apomixis, in Tripsacum dactyloides, 294 Arabinose, in brnr genotypes, 169 Asymmetry parameter, I39 Autopolyploids, genetic mapping, 69 - 7 1 Axial local symmetry, 128
B Balance of surface charge, 203 Barreness, maize resistance to, 272- 274 Batch titration technique, 210-212 Black layer, maize, 274-276 brnr mutations (see Brown-midrib mutations) Breakthrough curves, I8 - 2 1 Breeding programs maize, 245 -249 grain yield, 249 - 260 nitrogen fertility effects, 262-270 plant density effects, 262 - 270 stress vs nonstress environments, 260 262 use of DNA markers for identification of genetic variation, 65-66 for recombination and selection, 66-67
Brown-midrib mutations (see also specific brnr mutation) genetics, 160- I6 1 history Of, 159- I60 BSCB I reciprocal recurrent selection program, 282-290 BSSS reciprocal recurrent selection program, 282-290 Bypass flow effect of macropores, 22 - 25 measurement of, 25 modeling, 31
C Cadmium, in macropores, 33 Caffeic acid, 162 Carbon dioxide exchange, maize, 276-278 Catchment, internal, 25 Cation exchange capacity, 202 Cellulose in pearl millet, sorghum, and maize brnr mutants, 166 in plant cell wall, 169 Channels, voids as, 5 Chemical shift anisotropy, 127- 128 tensor, 127 Chlorides, for flow velocities, 18 Chlorophyll concentration, maize hybrids, 279 Chromosomal location, of QTLs, 55 - 57 Chromosomes, recombinant, 43 -44 Chromosome walking, 79 - 80 Cinnamic acid4hydroxylase, 162 Cianamyl alcoho1:NADPH oxidoreductase, I62 activity in brnr mutants, 163 Class pedotransferfunctions, 27 -29 ems-T gene, maize, 279-280 Coagulation defined, 203 zero point of charge determination, 2 14 299
3 00
INDEX
Cold tolerance, maize, 247 Colloids, isoelectric point determination, 2 14 Common intersection point, 202,227 Competing reactions, effects on surface charge measurements, 226-2267 Complex populations, DNA marker-assisted study Of, 67-69 Continuous pedotransferfunctions,27 Core lignin, 164 cross-linking to hemicelldose, 167 Correlation time, 108- 110 pCoumaric acid in forage plants, 167 trans-cinnamic acid conversion to, 162 coupling constants, quadrupolar, I39 - 143 CP NMR (see Cross-polarizationNMR) Cracks, horizontal, effect on , A and K,,,,, 22 Crop growth rate, maize hybrids, 279 Crop improvement exotic germplasm for, 7 1-75 genetic markers and, 80-82 Crossover, for potentiometric titrations, 232 Cross-polarization-magic angle spinning NMR interrupteddecoupling, 132- I35 inversion-recovery, 135- 137 reverse-electronics, I37 variable contact time, 132 Cross-polarizationNMR,123- 126 Cultural conditions, effect on maize yields, 253 Cultural practices, effect on maize yields, 248 - 249
D Decoupling dipolar, 129- 1 30 high-power, 129- 130 Density maize response lo, 262-270 for maize yields, optimal, 250-251,254255 Detailed balancing, principle of, 102, 104 Detasseling, in single-cross maize seed production fields, 293 Detection, in pulse NMR, 1 13- 1 15 Deterministic approach, to water flow modeling, 30 Diethyl sulfate, mutagen for bmr mutations, 160
2,4-Dihydroxy-7-methoxy-( 2H)- 1,4-benZOXazin-3-(4H)-1, pre-anthesis caused by, 280 Dipolar decoupling, 129- 130 Dipolar interactions, 109- 110 direct (through-space), 128 indirect (electron-coupled or f-coupled), 128 Disease resistance BSSS and BSCBl maize synthetics, 282 maize selection for, 279-280 Displacement, of liquid, 18-2 1 Dissociated charge, 203 -204 Divalent ions, enhancement of specific adsorption effects, 228 DNA markers agricultural genetics and, 40-41 in breeding programs for identification of genetic variation, 65-66 for recombination, 66-67 for selection, 66-67 for chromosomal location of QTLs, 55 .-57 for complex populations, 67 - 69 for crop improvement utilizing exotic germplasm, 7 1 - 75 for environmental effects, 60-6 1 for epistasis, 63- 64 for gene dosage effects, 57 - 59 for hybrid vigor, 61 -62 information obtained from, 49 linkage maps in crop plants, 50 for pleiotropy, 59-60 for polyploids, 69 - 7 I for transgression,62 Dominance theory of heterosis, 6 1 Dosage effects (see Gene dosage) Double crosses maize, 246,249 replacement by single crosses (maize), 290-292 Dry matter digestibility, pearl millet, sorghum, and maize bmr mutants, 166 Dyes, for pore continuity studies, 16
E Ear-to-row breeding, maize, 246 Effective spin -spin relaxation time, 108- 109 Effective transverse relaxation time, I08
INDEX Electrokinetic techniques, for surface charge measurement, 2 13 - 2 15 Electrolyte solutions, effects on surface charge measurement, 225-231 Electroosmosis, surface charge measurement and, 2 I 3 Electrophoresis comparative surface charge measurement with, 215 for isoelectric point of colloids, 2 14 Electrophoretic measurements, pH range of, 214-215 Electrophoretic mobility specific adsorption effects, 228-229 surface charge measurement and, 2 13 -2 14 zero point of charge determination, 2 14 Environmental effects of macropore flow, 33 on QTLs, 60-61 Environmental stress, effects on maize yields, 260 - 262 Epistasis DNA markers for, 63 -64 in polyploids, 69 Equilibration, effects on surface charge measurement, 230-231 Equivalent pore-size distributions, 8 - 9 Equivalent sizes, 6 - 8 Ethylmethane sulfate, mutagen for brnr mutations, 160 European corn borer, maize cultivars resistant to, 280-281 Evolutionary relationships, genetic maps for, 75-76
F Fertilizers, nitrogen for maize, 248 - 249 maize response to, 262-270 Ferulic acid catechol-0-methyltransferase-catalyzed, I62 in forage plants, 167 Ferulic acid-5-hydroxylase, 162 Field gradients, 139 Flotation, 203 Flux density, of water, 18 Forage quality digestibility, 170- 172 in vitro rates of digestion, 172- 173
301
lignin concentrations, 163- 167 neutral cell wall sugars, I69 nitrogen concentrations, 169 - 170 phenolics, 167- 169 total fiber, 163- 167 Force on a charge moving in a magnetic field, 149 Free induction decay, in pulse NMR, 1 13115 Full width at half-height, 109
C Galactose, in bmr genotypes, 169 Gene cloning, mapbased, 76 - 77 Gene dosage effects in polyploids, 69 on QTLs, DNA markers for, 57 - 59 Gene introgression,genetic markers for, 72 Genetic linkage maps (see Linkage maps) Genetic mapping comparative, 75-76 fundamentals of, 45 Genetic markers (see also DNA markers) for agricultural productivity, 48 - 54 CIWS Of, 42-43 for control of parental genotype disrup tion, 72 - 73 and crop improvement, 80-82 for gene introgression, 72 linkages, 43 -48 with genes affecting quantitative traits, 51-52 with genes affecting simply inherited traits, 51 Genetics, agricultural DNA markers and, 40 - 4 1 engineering vs plant breeding, 294 recurrent selection in maize, 282-290 Genetic variation, DNA markers for, 65-66 Genotypes effect on maize yields, 253 parental, genetic markers for, 72 - 73 Germplasm diversity, genetic markers for, 65 exotic, DNA markers for, 7 1-75 for maize pest resistance, 293 Glucose, in bmr genotypes, 169 Grain fill rate, maize, 275-276 Grinding, effects on zero point of charge, 223 Gyromagnetic ratio, 94
INDEX
302 H
Hartmann - Hahn condition, 124 Harvest index, maize, 275 -276 Hectorite slumes, N M R studies of, 145 Heisenberg uncertainty principle, 109 Hemicellulose brnr mutation effects, 169 in pearl millet, sorghum, and maize brnr mutants, 166 in plant cell wall, 169 Heterosis, 61 -62 High-power decoupling, 129- 130 High-resolution mapping, of QTLs, 56 Horizontal air-filled cracks, effect on K,, and Kunaw, 22 Htl gene, maize, 279 Hybrid vigor, 61 -62 Hydraulic conductivity, of planar and tubular soil pore models, 13 - 16 Hydraulic conductivity at saturation measurement of, 25 prediction, 16- 18 Hydraulic conductivity near saturation effect of macropores on, 2 1 -22 measurement of, 25 Hydroxycinnamic acid CoA ligase, 162 5-Hydroxy ferulic acid, 162
I Illite, NMR studies of, 145 Image analyzers, for pore size distribution, 6 Immobile water, 30, 34 Inbred lines, maize, 290-292 Indifferent ions, adsorption, 207 - 208 Inner-sphere complex charge, 20 1 Inner-sphere complexes defined, 20 I differentiation from outer-sphere complexes, 208 dynamics of, 146 Insect resistance, maize, 280-282 Interfaces, nuclear magnetic resonance spectroscopy, 144- I 46 Internal catchment, 25 Interrupted-decouplingCP-MAS NMR, 132- 135 Interval mapping, 53
Inversion-recovery CP-MAS NMR, 135137 Inversion-recovery experiments, 1 18- 1 19 In vitro dry matter digestibility, pearl millet, sorghum, and maize brnr mutants, 166 Ion exchange, effects on surface charge measurement, 229-230 Ion retention, surface charge measurement by, 215-217 Ions classifications, 208 -209 specifically adsorbed, 227-229 Ion-saturation techniques, effects on zero point of charge, 224-225 Iowa Corn Borer synthetic No. 1 maize, 267 -270 Iowa State Corn Yield Tests, 249-250, ,259 Iowa Stiff Stalk Synthetic maize, 267 - 2'70 Isoelectric point defined, 201 measurement, 214 Isotropic chemical shift, 127 Isotropic local symmetry, 128 Isozyme markers, tomato linkage map, 46 47
K K,, (seeHydraulic conductivity at saturation) K,,, (see Hydraulic conductivity near saturation)
L Larmor angular velocity, 95 Leaf-area index, maize, 276-278 Leaf orientation value above the ear, maize, 276-278 Leaforientation value below the ear, maize, 276-278 LeaEstem ratio, pearl millet, sorghum, and maize brnr mutants, 166 Leaves, lignin concentrations, 165 Lignification, enzymes involved in, 162- 163 Lignin biosynthesis, 161 - 163 concentrations in pearl millet, sorghum, and maize brnr mutants, 166 core, 164 total fiber and, 163- 167
INDEX Likelihood intervals, for QTLs, 55 Likelihood mapping, resolution, 78 Line broadening inhomogeneous, 109 sources of, 127- 129 Linkage analyses, in polyploids, 69 - 7 I Linkage blocks, 73 Linkage maps crop plants, 50 tomato, 46-47 Linkages, between genetic markers, 43 -48 and genes affecting quantitative traits, 5 1 54 and genes affecting simply inherited traits, 51
Local symmetry, types of, I28 Lodging resistance, maize, 249, 255 Longitudinal relaxation times, 107 Longitudinal (spin-lattice)relaxation time, rotating frame, measurement, 120- 123
M Macropore flow, environmental effects, 33 Macropores classification in size classes, 1 1 distributions complementarity of morphological and physical methods, I I - 12 morphometric techniques, 5 -6 soil physical techniques, 6 - 12 effects on flow rate, 13- 16 Kmtand K near saturation, 2 1 rooting patterns, 32 saturated and unsaturated flow, 2 1 - 22 solute movement, 13-25 mobile water in, 30, 34 occurrence in different soils, 2 types Of, 2-5 vertical, effect on K, and K,,,,, 22 Macroporous soils hydraulic conductivity at saturation, 1618
planar model, 13- I6 tubular model, 13- 16 Magic angle, 130 Magic angle spin, 131 - 132 Magic-angle spinning devices, 92
303
Magnetic fields force on a charge moving in a, I49 potential energy of a magnetic moment in, 149 representation, convention for, 148 static, magnetic moment in, 93-96 superconducting magnets, 92 Magnetic moment in a magnetic field, potential energy of, 149 nuclear, 98-99 in a static magnetic field, 93 - 96 Magnetic quantum number, 98 Magnetic resonance nonequilibriurn relaxation, 100- 103 spin-lattice relaxation, 103- I04 spin temperature, 103- 104 statistical model of, 100- 105 vector model classical physics, 93 - 97 magnetic moment in a static magnetic field, 93 - 96 quantum physics, 98- 100 Maize bmr mutants cellulose concentrations, 166 hemicellulose concentrations, 166 in vitro dry matter digestibility, 166 leafistem ratio, 166 lignin concentrations, 164, 166 neutral cell wall sugars, 169 nitrogen concentrations, 166 phenolics in, 167- 169 Maize breeding, 245 -249 disease resistance, 279-280 grain yield, 249 - 260 insect resistance, 280-282 nitrogen fertility effects, 262 -270 plant density effects, 262 - 270 pure-line method of, 246 role of physiology in, 278-279 selection for disease resistance, 279 - 280 single crosses, 290-292 stress vs nonstress environments, 260- 262 Maize cultivars plant health improvement, 274-279 esistance to barreness, 272-274 root loding resistance, 270-272 stalk lodging resistance, 270-272 Tuxpeno Crema I, 288 -289 Maize yields in Brazil, 258
3 04 BSSS and BSCBl synthetics, 282-290 density effects, 262 - 270 double crosses, 246,249 ear-to-row breeding, 246 factors affecting, 248 - 249 fertilizer use, 248-249 in France, 247,257 future trends, 292-294 genetic contribution to total yield, 258259 genetic gains in, 249- 260 genetic improvement, history of, 245-247 hand- vs machine-harvested, 262 inbred Lines, 290 - 292 Iowa Corn Borer synthetic No. 1,267 - 270 Iowa State University studies, 249-252 Iowa Stiff Stalk Synthetic, 267-270 lodging resistance and, 249,255 mass selection programs, 246 nitrogen fertilizer effects, 262-270 in Ontario, Canada, 257-258,279 optimal densities, 250-251 recurrent selection in populations, 282290 in Romania, 247 single crosses, 246, 249,290-292 in South Africa, 257 So S, generations, 256-257 stress vs noustress environment effects, 260-262 total grain yield and genetic yield gain of hybrids, 259 in the US., 248 - 249 varietal hybridization, 246 world production, 247 -248 in Yugoslavia, 247,258 MAS (see Magic angle spinning) Mass selection programs, maize, 246 Maxwell-Boltzmann statistics, 100- 107 Mechanical breakage, maize synthetic resistance to, 287 -288 Methylene Blue, for pore continuity studies, 16 0-Methyltransferase, catalysis of ferulic acid, 162 Mineral addition, surface charge measurement by, 218-219 Mobile water, 30, 34 Modeling, water flow, 29 - 32 deterministic approach
INDEX morphometric data, 30 schematized porosity, 30 stochastic approach, 29 Moment of force, 94, 149 Morphometric pore analysis, water flow based on, 13- 16 Morphometric techniques, for pore size distribution, 5 -6 Multiple factor hypothesis, 5 1 - 52
N Natural abundance sensitivity. 126 Necks, in a flow system, 16 Net proton charge, 20 1 Neutral detergent fiber in brnr mutants, 164 in cell wall, 164 in pearl millet, sorghum, and maize bmr mutants, 166 Neutral sugars, in cell wall, 169 Nitrogen concentrations in pearl millet, sorghum, and maize bmr mutants, 166 lost in bypass flow, 33 Nitrogen fertilizers, maize response to, 248 249,262-270 Nonequilibrium relaxation, 100- 103 Nuclear electricquadrupole moment, 139 Nuclear g-factor, 98 Nuclear-gyromagnetic ratio, effect on nuclide sensitivity, 126 Nuclear magnetic moment, 98-99 Nuclear magnetic resonance, quantum physics of, 99- 100 Nuclear magnetic resonance spectroscopy CP-MAS (see Cross-polarization-magic angle spinning NMR) cross-polarization, 123- 126 detection, 113-115 enhancement by polarization transfer, 123-126 experiments, 115-117. 126-132 free induction decay, 113- 1 15 high-power decoupling, 129- 130 of interafaces and adsorbates, 144- 146 inversion-recovery, 1 18- 1 19 line-broadening, 127- 129 local symmetry of powder patterns, 128 materials suitable for, 147
INDEX measurements of T2,119- 120 of Tp, 120-123 of T (temperature), I I8 - I 19 rfpulse~,112-113 sample spinning, 130- 132 selective excitation, 132- 137 sensitivity, 126- 127 single-pulse excitation, 1 I7 - I 18 spectral editing, 132- 137 spin-echo, 1 19- 120 spin-lock, 120- 123 stagesof, 115-116 Nuclear magneton PN,98 Nuclear spin, 98-99 Nuclear-spin quantum number, 98 Nuclear-spin states degenerate, 98 selection rule, 99- 100 Nuclides sensitivity, 126- 127 for spin-locking, 125- 126
Ostrinia nubilalis, maize cultivars resistant to, 280-281 Outer-sphere complex charge, 201 Outer-sphere complexes defined, 20 I differentiation from inner-sphere complexes, 208 dynamics of, 146 Overdominance theory of heterosis, 6 1 Oxides, surface charge, 234
P Packing voids, 3 - 5 Pearl millet bmr mutants cellulose concentrations, 166 hemicellulose concentrations, 166 in vitro dry matter digestibility, I66 leafstem ratio, 166 lignin concentration, 164 lignin concentrations, 166 nitrogen concentrations, 166 phenolics in, 168- 169 Pedal soil materials, 2 - 3
305
Pedotransferfunctions, 27 -29 Permanent charge defined, 20 1 inorganic materials, 205 separation from variable charge, 219-222 pH effects on specific adsorption, 228 on surface charge measurements, 226 - 227 on surface charge ofcomplex materials, 234 Phenolic compounds, phenylalanine substrate, 161 Phenolics, in forage plants, 167- 169 Phenotype, cumulative effects of QTLs epistasis, 62-63 hybrid vigor, 6 I -62 transgression, 62 Phenylalanine, lignin synthesis, I6 1 Phenylalanine ammonia-lyase, lignin synthesis, 16 1 Phosphate adsorption, N M R studies of, 146 Phosphorus, lost in bypass flow, 33 Photosynthetic response, maize hybrids, 279 pH range, of electrophoretic measurements, 214-215 Phyllosilicates, N M R studies of cations absorbed by, 144- 145 Physiological traits, important for grain yield, 276-278 Piston flow, 20 Planar pore models, 13- 16 Planes, 5 Plant density maize response to, 262 -270 for maize yields, optimal, 250-251,254255 Plant health, maize, 274-279 Pleiotropy, DNA markers for, 59-60 Point counts, 5-6 Point of zero net charge, 203 determination by ion retention,2 15- 2 17 Point of zero net proton charge, 202 Point of zero salt effect, 202 acid washing effects, 224 measurement, 2 18-220 Points of zero charge, common definitions of, 204 Polarization transfer enhancement by, 123- 126 nuclide sensitivity enhancement with, 126- 127
306
INDEX
to quadrupolar nuclei, 143- 144 Pollen-silk interval, maize, 274-276 Polyploids, DNA marker-assisted study of, 69-71 Population size, linkage distance and, 7 I Pores (see Macropores) Porosity, of phanar and tubular soil pore models, 13-16 Potential-determiningions adsorption of, 204 classification problems, 208 -209 defined, 201 variations in charge due to, 205-206 Potential energy, magnetic moment in a magnetic field, 149 Potentiometric titration change with time of point of zero salt effect, 230-231 comparative surface charge measurement with,215 crossover areas, 232 defined, 202 surface charge measurement, 209 - 2 13 Precipitates, aging of, 225 Pristine point of zero charge defined, 202 overall zero point of charge and, 232 Pure-line method, of maize breeding, 246
Q QTLs (see Quantitative trait loci) Quadrupolar coupling, 129, 139- 143 Quadrupolar frequency, 140 Quadrupolar interactions, 109- 110 Quadrupolar nuclei, 138- 144 polarization transfer to, 143- 144 Spectra Of, 139- 143 Quantimet counts, 6 , 9 Quantitative trait loci comparative mapping, 75 -76 individual characterization, 52 individual, 54 - 55 chromosomal location, 55 - 57 environmental effects, 60-61 gene dosage effect, 57 - 59 mapping, 55 - 57 pleiotropy, 59-60
linkages between genetic markers and genes affecting, 5 1 - 54 location of, 52 multiple-marker association, 53 single-marker association, 53 Likelihood mapping, resolution, 78 multiple, effects on individual phenotype, 61 -64 epitasis, 63-64 hybrid vigor, 5 1 - 52 transgression, 62 substitution mapping, resolution, 78 Quantum number magnetic, 98 nuclear-spin, 98
R Radiofrequency pulses, 1 12- I 13 Rate of transition (energy states), 102 Recombinant chromosomes, 43 -44 Recombination between genotypes, DNA markers for, 66 67 in polyploids, 69 Recurrent selection, for maize, 282-290 half-sib testcross progenies, 283 -284 interpopulation crosses, 283 reciprocal full-sib, 284-289 Relaxation correlation time, 108- 110 frequency dependence of, 110- 1 12 nonequilibrium, 100- 103 physical mechanisms of, 108- 110 spin-lattice, 103- 104 vector model of, I06 - I08 Relaxation times, 107 Repulsion phase linkage, 70 Resolution, of gene manipulation methods, 78 Restriction fragment length polymorphisms, 43-44 for eliminating undesirable genome regions, 74 tomato linkage map, 46-47 Reverse-electronicsCP-MAS NMR, 137 rhm gene, maize, 279 Rooting patterns, effect of macropores, 32
INDEX Root lodging BSSS and BSCB 1 maize synthetics, 284, 286 maize resistance to, 270-272 Roots,lignin concentrations, 165 R d gene, maize, 279
S Salt titration, surface charge measurement by, 217-218 Sample pretreatment, effect on zero point of charge, 223 Sample spinng, variable-angle, 13 1 Sample spinning, in pulse NMR, 130- 132 Saturated flow, effect of macropores, 2 1- 22 Scalar chemical shift, 127 Sedimentation defined, 203 zero point of charge determination, 2 14 Selection in breeding programs, DNA markers for, 66 - 67 for disease resistance in maize, 279-280 DNA marker-facilitated, 76 recurrent, for maize, 282 - 290 Selection rule, nuclear spin states, 99- 100 Sensitivity, of nuclides, 126- 127 Serial titration techniques, 2 10-2 12 Shielding, 127 Short-circuiting, 26 Shrinkage cracks, effect on &,and K,,, 22 Silicate materials, dominating charge characteristics, 207 Silk emergence, maize, 274-276 Sinapic acid, formation, 162 Single crosses detasseling, 293 maize, 246,249,290-292 Single-dose restriction fragment polymorphism, for polyploid linkage analysis, 70 Sizing (particles), effects on zero point of charge, 223 Smectite, NMR studies of, 145 Soil matrix, immobile water in, 30,34 Solid phase, alteration of, 222-225 Solid-state single-pulse-excitation magicangle NMR spectra, 14 1 Solutes, movement, pore size effects, 13- 25
307
Sorghum brnr mutants cellulose concentrations, 166 hemicellulose concentrations, 166 in vitro dry matter digestibility, 166 leafstem ratio, 166 lignin concentrations, 164- 166 neutral cell wall sugars, 169 nitrogen concentrations, 166 phenolics in, 167- 169 Sorghum-sudangrass bmr mutants lignin concentrations, 164- 165 neutral cell wall sugars, 169 Specific adsorption, effects on surface charge measurements, 227 - 229 Spectral density function, 1 10- 111 Spin-lattice relaxation, 103- 104 Spin-locking, 129 measurement of T,,,, 120-123 nuclides for, 125 - 126 Spin -spin exchange, 109- 1 10 Spin temperature, 101, 103-104 Stalk lodging, I59 BSSS and BSCB 1 maize synthetics, 284, 286 maize resistance to, 270-272 Stalk rot, maize synthetic resistance to, 287288 Standability, maize, 247 Static dipolar coupling, 110 Statistical model, of magnetic resonance, 100- I05
Stay-green, maize, 274-276,278 Stems, lignin concentrations, 165 Stochastic approach, to water flow modeling, 29 Streaming potential comparative surface charge measurement with, 2 15 surface charge measurement and, 2 13 Stress environment, effect on maize yields, 260-262 Substitution mapping QTLs, 56 resolution, 78 Sudangrass bmr mutants, neutral cell wall sugars, 169 Sugars, in cell wall, 169 Superconducting magnets, magnetic field of, 92
INDEX
308 Surface charge applications, 234-235 defined, 20 1 measurement comparison of techniques, 2 15 electrokinetictechniques, 2 13- 2 15 ion retention, 2 15 -2 17 mineral addition, 2 18-2 19 potentiometric titration, 209-21 3 salt titration, 217-218 measurement problems interactions with electrolyte solution, 225-231 solid phase alteration, 222-225 oxides, sensitivity to pH, 234 permanent charge, 205 predictions of, 235-236 variable charge, 205 Symmetric second-rank tensors, 138- 139
T Taranakite, N M R studies of, 145 Temperature, NMR measurement, 118- 1 19 Titration techniques, 2 10- 2 12 Tomato, genetic linkage map, 46-47 Torque, 94 Total dry weight, maize, 275,278 Traits quantitative from exotic germplasm, 72 linkage between genetic markers and genes affecting, 51 -54 simply inherited from exotic germplasm, 72 linkage between genetic markers and genes affecting, 5 1 Transgression, DNA markers for, 62 Transition probability (energy states), 102 Transition rate (energy states), 102 Transverse relaxation time, 107 measurement, 119- 120 Tripsacurn dactyloides, apomixis in, 294 Trivalent ions, enhancement of specific adsorption effects, 228 Tubular pore models, 13- 16 Tuxpeno Crema I maize cultivars, 288-290 Tyrosine, phenolic compound synthesis, I6 1
U Unsaturated flow, effect of macropores, 2 I 22 V
Variable-angle sample spinning, 131 disadvantages, 142 Variable charge defined, 20 1 inorganic materials, 205 separation from permanent charge, 2 19222 Variable contact-time CP-MAS NMR, 132 Varietal hybridization, maize, 246 Vector models of magnetic resonance, 93- 100 of relaxation, 106 - 108 Voids, types of, 3-5 vughs, 5
W Water flux density, 18 immobile, in soil matrix, 30, 34 mobile, in macropores, 30, 34 Water flow macropore, environmental effects, 33 modeling, 29-32 morphometric pore analysis-based, 13- 16 saturated, effect of macropores, 2 1 -22 unsaturated, effect of macropores, 2 1- 22 velocity, chlorides for, 18 Weather effects, on maize yields, 253 World production, maize, 247 - 248
X Xylose, in bmr genotypes, 169
Y Yeast artifical chromosomes, resolution., 7880
INDEX
Z Zero point of charge acidity effects, 223 alterations in, 222-225 in composite materials, 235-236 defined, 20 1 determination,2 14 effects of ion-saturation techniques, 224 225
309
grinding effects, 223 mechanical vs coprecipitated mixtures, 233 suing effects on, 223 Zero point of titration, measurement, 220 Zero zeta potential, 20 I Zeta potential defined, 201 determination.2 13
This Page Intentionally Left Blank
