Primates of the Oriental Night

CONTENTS Foreward

v

Preface

vii

Acknowledgments

ix

Taxonomic History of the Tarsiers, Evidence for the Origins of Buffon’s Tarsier, and the Fate of Tarsius spectrum Pallas, 1778

01

Distribution and Biogeography of Tarsiers

13

A Tarsier Capture in Upper Montane Forest on Borneo

29

Distribution of Tarsier Acoustic Forms, North and Central Sulawesi: With Notes on the Primary Taxonomy of Sulawesi’s Tarsiers

35

Distribution of Tarsier Haplotypes for Some Parts of Northern and Central Sulawesi

51

A Method for Multivariate Analysis and Classification of Tarsier Tail Tufts

71

Tarsier Longevity: Data from a Recapture in the Wild and from Captive Animals

85

Eastern Tarsiers in Captivity, Part I: Enclosure and Enrichment

91

Eastern Tarsiers in Captivity, Part II: A Preliminary Assessment of Diet

97

The Conservation Status of Indonesia’s Tarsiers

105

Craniometry of Slow Lorises (Genus Nycticebus) of Insular Southeast Asia

115

Enclosure Design for Captive Slow and Pygmy Lorises

123

Confiscation, Rehabilitation and Placement of Slow Lorises: Recommendations to Improve the Handling of Confiscated Slow Lorises Nycticebus coucang Dedication

137 146

Foreword PROF. DR. ROCHADI ABDULHADI PRESIDENT OF THE INDONESIAN BIOLOGICAL SOCIETY EXECUTIVE SECRETARY INDONESIAN INSTITUTE OF SCIENCES (LIPI) I am extremely pleased to see that “Primates of the Oriental Night” has, at last, been completed. This book is a fabulous example of the benefits of true collaboration between researchers of many nations. The list of authors includes people from Austalia, England, Germany, Indonesia, Mexico, the Philippines, and the United States of America. It is wonderful what we can accomplish when we all work together. We live in a globalized world, and it is increasingly necessary that we do so. I am particularly pleased to see several Indonesian authors are involved, and that the book is being published in Indonesia, by the Indonesian Institute of Sciences. We can point to this book as proof that many Indonesians are mentally prepared for the globalized world. Indonesia is a country of vast biodiversity resources, but also vast conservation challenges. When looking for conservation solutions, too often we focus only on the large charismatic animals like orangutan, tiger, and elephant. In Indonesia we are blessed with a wealth of animals like this, to the extent that we forget that there are small charismatic animals. Tarsiers and lorises, the primates of the oriental night, also have tremendous charisma. We need only point to examples from popular movies, such as the Yoda character and the Ewoks, from the “Star Wars” movie series, to see the powerful influence of these charismatic animals on the global society. Unfortunately, we can also point to the prevalence of tarsiers and lorises for sale in the markets of Indonesia and other southeast Asian countries to offer proof that the general public is fascinated with these creatures. The publication of this book marks a sea change in the knowledge about tarsiers and lorises that is available to readers in southeast Asia. I am delighted to learn that one hundred copies of this book will be distributed free of charge to scientists and schools throughout Indonesia. I hope that this effort motivates other Indonesians to follow in the footsteps of the Indonesian authors who contributed to “Primates of the Oriental Night” and become serious biodiversity scientists, committed to the pursuit of knowledge about, and the preservation of, Indonesia’s vast and valuable biodiversity.

v

PREFACE

Some of the manuscripts contained in this volume have existed in various forms for several years now and, in order to avoid some confusion, it may help to offer an explanation as to why they have finally appeared in their current form. Plans for this volume began with the International Primatological Society Congress in Adelaide, Australia, in January 2001. Colin Groves and myself organized a symposium “Advances in Tarsier Biology”. Our intentions were to present a series of seminars that would complement the book “Tarsiers: Past, Present, and Future” edited by Patricia C. Wright, Elwyn L. Simons, and Sharon Gursky, which had been scheduled for release prior to the congress. With the delayed release of that book, and the poor attendance at Adelaide, our plans for an edited volume from that symposium were delayed. By the summer of 2002, Colin Groves, Alexandra Nietsch, and myself had been exchanging emails on the issue of tarsier taxonomy and unrecognized species within the T. tarsier species complex—what we were calling Eastern tarsiers—for several years, and were ready to meet to discuss the issue face-to-face. We located a sponsor, Dr, Willie Smits, of The Gibbon Foundation, and planned a meeting for early November 2002, in Jakarta. As the world is well aware, terrorists exploded several bombs in tourist areas of Bali on October 12, 2002, resulting in the deaths of over 200 people. Particularly affected was Australia, from where about half of the victims originated. If for no reason better than to appease our worried family members, the Tarsier Taxonomy Workshop in Jakarta was postponed. It is hard to imagine a silver lining in that tragedy, and yet the “Indonesian Prosimian Workshop”, upon which this book is ultimately based, owes it genesis to that event. Concurrent to our planning for this event, Dr. Smits opened the Schmutzer Primate Center, a world class, semi-autonomous primate facility within Ragunan Zoo in Jakarta. Within Indonesia large numbers of lorises, as well as some tarsiers, are for sale on the black market. Large numbers of these primates would need to be confiscated to stem the trade, and the Schmutzer Primate Center would require expertise in the management of nocturnal primates. Therefore, during the weeks after the bombing I asked Dr. Smits if he would like for me to contact my colleagues with loris expertise, and expand the focus of our workshop to include captive care and conservation of tarsiers and lorises. He readily agreed to an expanded budget that would allow to us invite experts who were active and specialized in those areas. To help organize the workshop, I contacted Helena Fitch-Snyder, whom I had met through our work together in the North American Prosimian Taxon Advisory Group, and Helga Schulze, whom I had met at the “Creatures of the Dark” conference in Durham, North Carolina in 1993. In the end, the foreign invitees at the workshop included three researchers active in tarsier field biology, three who were active in loris husbandry and conservation, one who had conducted field work on both tarsiers and lorises, and one taxonomist. In addition to Dr. Alexandra Nietsch and myself, tarsier field biology was represented by Irene Neri-Arboleda, who had recently completed the most thorough field study of Philippine tarsiers to date. This workshop was particularly blessed by the participation of Dr. Sharon Gursky, Assistant Professor of Anthropology at Texas A & M University who not only has amassed more hours of field data than anybody else on tarsiers, but is quite possibly the one-and-only person who has conducted systematic field studies on both tarsiers and lorises. Dr. Helena Fitch-Snyder and Helga Schulze, who edited the book “A Manuaul of Loris Husbandry”, and who between them have possibly more experience than anyone else in maintaining loris colonies outside of their native countries, represented lorises. Dr. Ulriche Streicher, D.V.M. came to the attention of the organizers from her presentation at the Beijing, IPS meetings, wherein she offered evidence that showed how coat markings in lorises change seasonally. Her participation was quite fortunate, because as a D.V.M. working at a primate rescue center in Vietnam that specializes in lorises, her experience proved to be invaluable to her Indonesian counterparts on the frontline of combating the illegal trade in lorises.

vii

The workshop benefited greatly from the participation of Dr. Colin Groves, Professor of Anthropology and Archeology at Australian National University, author of the book “Primate Taxonomy” and an expert in seemingly everything. The workshop itself led to a new collaboration with Indonesian scientists, including mammalian taxonomist, Ibnu Maryanto, who co-authored a manuscript with Dr. Groves and graciously agreed to assist in the editing of this book. Two other long time colleagues and friends, loris biologist Anna Nekaris and tarsier biologist Stefan Merker, were invited to attend but were unable to because of other obligations. Thus, the objective of this book is to pull together several manuscripts that are related to each other as least as much by history as by subject matter. As such, this book does not present as cohesive a narrative as “Tarsiers: Past, Present, and Future”, and in some respects, it is more a complement to that book. In other respects, however, this book publishes the results of the “Indonesian Prosimian Workshop”. As such, it includes information on animal rescue and captive care. This book, therefore, is written by specialists in tarsier and loris biology, and is written for the larger community of specialists—scientists, conservationists, animal keepers, monitors, and others—who require current information about our subject matter, the nocturnal primates of Asia. It is worth commenting on our choice of publishers. Given this book’s earliest origin as a symposium for the IPS Congress at Adelaide, there were discussions with academic publishing houses. In the end, by choosing the Indonesian Institute of Sciences, we wound up with great freedom in terms of manuscript length and color illustrations. Additionally, given the cost savings, we will be able to distribute 100 copies, free of charge, to schools, libraries, and other institutions throughout Indonesia. Myron Shekelle et al. Jakarta, 5 July 2008

Some participants at the Indonesian Prosimian Workshop, Schmutzer Primate Center, 21 February 2003.

viii

ACKNOWLEDGMENTS The editors are deeply indebted to Dr. Willie Smits and the Gibbon Foundation, who funded the workshop and this volume. The Schmutzer Primate Center Staff hosted many of the workshop activities, and several staff members assisted with the workshop, including Femka Den Haas, and Made Wedana. Dr. Siti Nuramaliati Prijono, director of the Indonesian Institute of Sciences’s Museum Zoologicum Bogoriense sponsored a seminar that featured the workshop’s participants and hosted other workshop activities.

Dr. Willie Smits

Dr. Siti Nuramaliati Prijono

Dr. Ibnu Maryanto

ix

Primates of The O riental Night

TAXONOMIC HISTORY OF THE TARSIERS, EVIDENCE FOR THE ORIGINS OF BUFFON’S TARSIER, AND THE FATE OF Tarsius spectrum Pallas, 1778 Colin Groves 1), Myron Shekelle 2), & Douglas Brandon-Jones 3) School of Archaeology and Anthropology, Australian National University Canberra, ACT 0200, Australia. Email: [email protected] 2) Center for Biodiversity and Conservation Studies, Faculty of Mathematics and Natural Sciences University of Indonesia, Depok 16421, Indonesia, Email: [email protected] 3) 22 Karenia Street, Bray Park, QLD 4500, Australia 1)

ABSTRACT A survey of the history of tarsier taxonomy indicates that Tarsius tarsier Erxleben, 1777 is a senior subjective synonym of T. spectrum Pallas, 1778. Buffon’s tarsier, long thought lost or possibly destroyed, has been recently rediscovered and is identified as being eastern in origin (i.e. from within what has previously been classified as T. spectrum or the T. spectrum complex). The identification of Buffon’s tarsier as an Eastern tarsier alters Hill’s taxonomy by making T. spectrum a junior subjective synonym of T. tarsier. Eastern tarsiers become the type species of the genus. Our work, conducted, before the rediscovery of Buffon’s specimen, is based on illustrations of the skin and cranium by Daubenton. Investigations of Buffon’s specimen are ongoing, but do not alter our fundamental conclusions. Daubenton’s illustration of the cranium shows a nasal profile that is not consistent with published illustrations of Philippine and Western tarsiers. Several other characteristics are argued to be consistent with Buffon’s tarsier being eastern in origin. Keywords: T. tarsier, T. bancanus, T. syrichta, Taxonomy

A BRIEF HISTORY OF TARSIER TAXONOMY - Linnaeus’s tarsier Standing last in the list of Linnaeus’s (1758) 21 species of the genus Simia is Simia syrichta, reading as follows: S. caudata imberbis, ore ciliisque vibrissatis. Syst.nat.3. Cercopithecus luzonicus minimus. Pet.gaz. 21.t.13.t.11. Habitat in Luzonum insulis. Cabrera (1923) showed that Petiver (“Pet.”) took his 1705 description in turn from the papers of a Jesuit missionary, G..J. Camel, who was clearly referring to a tarsier, and that the locality, Luzon is probably to be explained as being a place to which tarsiers were traded from nearby Samar. Cabrera (1923) used the form Tarsius syrichtus, but, as Linnaeus (1758) spelt the name Syrichta with a capital initial letter, it is probable that he was using it as a noun in apposition (he began adjectives with a small letter), so it does not change gender: see Musser & Dagosto, 1987.

Consequently, the earliest available name for a Philippine tarsier is syrichta Linnaeus, 1758. Cabrera’s (1923) mention of Samar may be taken as a restriction of the type locality. Cabrera’s (1923) opinion that Simia syrichta is a Philippine tarsier has been supported by numerous authors since that date. Hill (1953a,b), in his influential revision, accepted Cabrera’s (1923) assessment of Linnaeus’s Simia syrichta as the Philippine tarsier; so did Musser & Dagosto (1987). Remarking on its essential accuracy, Niemitz (1984) translated Camel’s description, including the comment, “it is said to live on charcoal, but this is wrong”! Meyer (1895), not knowing all this, described the Philippine tarsier as Tarsius philippinensis, with type locality Samar, while Heude (1898:164) clearly thought that he was the first to name a Philippine tarsier when he described T. carbonarius from two specimens from Mindanao: from the Gulf of Davao and the valley of the River Poulangui. He distinguished this from the only other species he knew, “Tarsius spectrum de Java” (sic), by dental characters and by the narrowness and greater length

1

Groves, Shekelle & Brandon-Jones- Taxonomic History of The Tarsiers

of the skull, adding, in a curious but quite independent echo of Camel, “On m’a dit qu’il mangeait du charbon!” Buffon’s tarsier Buffon (1749:87, and pl. 219) gave the vernacular name Tarsier or Woolly Jerboa to an animal he received from an unknown locality. Remarking on its long hind legs and other characters, he described

“the inferior part of the hind legs” as being hairless and its tail as being, like that of the jerboa, garnished with long hairs towards the tip. It is not clear exactly what is meant by “the inferior part of the hind legs” – evidently, not necessarily the whole of the tarsus, because in the plate the proximal part of the tarsus is haired (Figure 1). This, and the conspicuously tufted tail, eliminates

Figure 1: Buffon’s tarsier. This illustration has the gestalt of an Eastern tarsier in having a tail that is more hirsute than is common in Western and Philippine tarsiers, relatively small eyes, long and pointed ears, woolly appearance of the fur, and similarities of the relative leg length (from Buffon 1765).

2

Primates of The O riental Night

the almost naked-tailed Philippine tarsier (Figure 2). The head is not excessively broad, and the very long, dark, conspicuous tail tuft suggests it is not a Western tarsier. Buffon’s colleague Daubenton published, as an addendum (p.114 [N°MCCXXXV]), a description of the skull and skeleton of the tarsier (always assumed, perhaps not correctly, to be of the same specimen). A plate of the skull and skeleton was published by Fischer (1804), who called it “Daubenton’s tarsier”. Poor as they are, the figures show the relatively small orbits and long braincase of the Sulawesi tarsier (Figure 3). Other than a few specimens mounted for public display, I. Geoffroy St. Hilaire (1851:ii, fn.2) found Buffon’s collection neglected, and “ne tardait pas à être attaqué par les insects”. Nonetheless, the mounted tarsier specimen does still exist; it is the subject of a paper in preparation by CPG, in collaboration with C. Callou and J. Cuisin. Suffice it to say that it is indeed a Sulawesi tarsier, and we cannot understand why Elliot (1910) baldly claimed that its “bare tarsi and nearly naked tail” (sic!) shows it came from the Philippines. Elliot refused to fix a type locality and considered it indeterminable, adding that as Pallas’s spectrum was based upon it, “Tarsius spectrum must be dropped from the list of recognized species”. Buffon was the main source for Erxleben’s (1777:72) name Lemur tarsier, which consequently is not a Philippine tarsier as listed by Hill (1953a), but is the earliest available name for a Sulawesi tarsier. Schreber (1778:554) cited Buffon and Erxleben for the description of his Didelphys? macrotarsos, which he inclined to think was actually a marsupial. Gmelin (1788) cited Buffon first, and Pennant second, for his Didelphis macrotarsus. Link (1795) likewise based the name Macrotarsus buffonii on Buffon’s tarsier, as did Audebert (1797) for Tarsius daubentonii.

probably Ambon”, and is called podje by the Macassans. Pallas described its teeth and its general external appearance, ending with “Cauda nudiuscula”, meaning “Tail virtually naked”. According to Smit et al. (1986), Schlosser ’s cabinet may have been purchased by Boddaert, but thereafter nothing is known about it. Hill (1953a) argued that the description was that of a Sulawesi tarsier, and fixed the type locality as Macassar (=Makassar, until recently known as Ujung Pandang, but the name recently reverted to its original, Makassar), because of Pallas’s mention of its Macassan name, podje. Possibly influenced by Sody (1949), Niemitz (1984:13) influentially gave the type locality as Minahasa (far northern Sulawesi). In the absence of any evidence to the contrary, Hill’s fixation is better substantiated. As Cabrera (1923) recognized, Lemur tarsier Erxleben, 1777, based on Buffon’s description (see above), predates Lemur spectrum Pallas, 1778 and Didelphys? macrotarsos Schreber, 1778. Unfortunately, therefore, the well-known name Tarsius spectrum, commonly used for Sulawesi lowland tarsiers, must be superseded by Tarsius tarsier (Erxleben, 1777). Geoffroy (1796) later redescribed Pallas’ specimen as Tarsius pallassii.

Pallas’s tarsier Pallas (1778) also based Lemur spectrum in Buffon’s tarsier, but referred to it a specimen in the museum of his late friend Schlosser, which in effect formed the bulk of his description. He stated that it came from “the furthest islands of the Indian Ocean,

Citing Pallas (1778), Pennant (1781) gave Amboina (=Ambon) as the locality, and podje as the Macassan name. Later Kerr (1792) paraphrased Pennant’s (1781) description, naming it Lemur podje. “Doctor Hunter” is almost certainly the famous surgeon John Hunter (1728-1793), whose

-Pennant’s tarsier Pennant (1771) redescribed Buffon’s tarsier, and ten years later (Pennant, 1781) he described “two fine specimens from the cabinet of Doctor Hunter” as having A pointed visage… hairs on the legs and feet short, white, and thin; tail almost naked: the greater part round and scaly, like that of a rat; but grows hairy towards the end, which is tufted.

3

4

Philippine tarsier

T. tarsier (unknown prov.)

T. tarsier Togian Islands

Eastern tarsiers

T. dentatus

Figure 2: Diagnostic value of tarsier tails. The Philippine tarsier in this figure has a very slight, almost invisible tuft of fur on the distal few centimeters of the tail. The tuft fur is sparse, very short, and light in coloration. Western tarsiers have a clearly apparent tuft of fur on the distal end of the tail, in appearance like that of a feathered arrow shaft. The fur on the tail of Eastern tarsiers is altogether different than the other two and gives the tail of an Eastern tarsier an appearance similar to that of a bottle brush. The tail fur of Eastern tarsiers is generally longer, darker, and the tuft typically covers as much as 1/3 to 1/2 of the length of the tail, but it develops gradually along the length of the tail such that its actual length is difficult to measure accurately. In T. sangirensis, the tail tuft is in the same position as it is in other Eastern tarsiers, but the fur is shorter, sparser, and lighter in color. (adapted from Shekelle 2003)

T. b. bancanus T. s. fraterculus T. sangirensis Tarsier photos © Myron Shekelle, 2002, except T. syrichta, Sheena Hynd

Western tarsier

Groves, Shekelle & Brandon-Jones- Taxonomic History of The Tarsiers

Primates of The O riental Night

Buffon’s tarsier

Eastern tarsiers

Philippine tarsier

Western tarsier

Figure 3a: Buffon’s tarsier is most similar to Eastern tarsiers in the shape of the nasal profile (adapted from Musser and Dagosto 1987, Fischer 1804, and Shekelle 2003).

Buffon’s tarsier

Eastern tarsiers

Philippine tarsier

Western tarsier

Figure 3b: Buffon’s tarsier is most similar to Eastern tarsiers in that the shape of the cranium is oblong and has a slight postorbital constriction (adapted from Musser and Dagosto 1987 and Fischer 1804). Note: Tarsius dianae is a junior synonym of T. dentatus

Buffon’s tarsier

Eastern tarsiers

Philippine tarsier

Western tarsier

Figure 3c: Buffon’s tarsier is most similar to Eastern Tarsiers in the relative height of the orbits and the braincase (adapted from Musser and Dagosto 1987 and Fischer 1804).

5

Groves, Shekelle & Brandon-Jones- Taxonomic History of The Tarsiers

anatomical preparations and manuscripts were bequeathed to the nation under the trusteeship of the Royal College of Surgeons collection upon his death, forming the basis for what became the Hunterian Museum. The Royal College of Surgeons took a direct hit during the Second World War, and many valuable specimens were destroyed, presumably including the tarsiers. Although the description is not absolutely clear-cut, the “pointed visage” and the description of the tail strongly suggest a Sulawesi tarsier, so Lemur podje Kerr, 1792 is best regarded as a junior synonym of Lemur tarsier although it is of course available to anyone who might consider Hunter’s specimens as distinct. Fischer’s revision of tarsiers Fischer (1804) considered that he had evidence for not one but three species of tarsier: Tarsius pallassii (adopting Geoffroy’s renaming of Lemur spectrum Pallas, 1778), T. daubentonii (an explicit renaming of Lemur tarsier Erxleben, 1777 and Didelphis macrotarsus [sic] Gmelin, 1788, apparently unaware of Audebert’s similar action), and a new species T. fuscus or fuscomanus. This new species was said to have come from Madagascar, probably (as he remarked) an error for Macassar. His description of the tail could only be that of a Sulawesi tarsier. Differences between his three species, which in effect, therefore, are all Sulawesi tarsiers, are based on stages of wear on the incisors and other features. Hill (1953a) presumed that the type of Fischer’s T. fuscus was lost, and there seems no contrary evidence. Tarsius pallassii and T. daubentonii are objective junior synonyms of Lemur spectrum and Lemur tarsier, respectively. Given the type locality, Macassar, for all three of Fischer’s taxa, T. fuscus (which he also called T. fuscomanus) must rank as a subjective junior synonym of Lemur tarsier, despite the fact that Miller and Hollister (1921), obscurely, referred to “Tarsius fuscus fuscus from northeastern Celebes”. Desmarest (1804) combined Fischer’s first two species into one, which he called “Tarsier aux mains rousses, Tarsius spectrum Geoffroy” [sic],

6

including Buffon’s, Pallas’s and Pennant’s tarsiers; and renamed the third “Tarsier aux mains brunes, Tarsius fischerii = T. fuscomanus Fischer”. Other 19th century revisions of tarsiers Horsfield (1824) described the first welllocalised tarsier, Tarsius bancanus from Jeboos (=Jebus), Bangka. It was described as being dark, lacking upper central incisors (but the type was evidently an infant, and having rounded, horizontal ears, shorter than the head, and a flatter facial profile (Figure 4). Horsfield (1824) compared it to Fischer’s (1804) three species, citing T. fuscus as T. fischeri. Fitzinger’s (1870) revision differed little from Horsfield’s (1824) arrangement, recognising four species: Tarsius spectrum (based on Pallas’s tarsier, and from Ambon), T. fuscomanus (Fischer’s tarsier, but said to be from Mindanao and Bohol), T. daubentonii (including Buffon’s, Pennant’s, Schreber’s and Erxleben’s tarsiers, from Celebes, Selayer and Borneo), and Tarsius bancanus (from Sumatra and Bangka). His nomenclature aside, his revision is noteworthy for being the first to compare Philippine and Sulawesi tarsiers, and to widen the known distribution by the inclusion of some new localities. Meyer (1897) had a different view of the nomenclature. He used Tarsius fuscus Fischer for the Sulawesi tarsier, T. philippensis Meyer (1895) for the Philippine species, and T.spectrum (Pallas) for the one from Borneo (which he also knew from the Karimata and Natuna Is., Belitung and Sumatra). He described a new species, T. sangirensis from the Sangihe Is. This represents a further step towards a modern understanding of the species of tarsiers. 20th century revisions Elliot (1910) described two new species, Tarsius saltator from Belitung and Tarsius borneanus from the Landak River, West Kalimantan. Chasen (1940) regarded these as subspecies of the only species he recognized, Tarsius tarsier Erxleben, and described a further subspecies, T. t. natunensis from Sirhassen, in the South Natuna Is.

Primates of The O riental Night

Figure 4: Buffon’s and Horsfield’s tarsiers. The illustration of Buffon’s tarsier (upper left) does not have the gestalt of an infant, as seen in the illustration Horsfield’s tarsier (upper right), and the mother / infant photo (right) (from Buffon 1765 and Horsfield 1824). Tarsier photo (c) Myron Shekelle 2008.

Miller (1910:404) described a new species, Tarsius fraterculus, from Bohol in the Philippines, said to resemble T. philippensis but smaller in size. Miller & Hollister (1921) described two new tarsiers from Sulawesi: Tarsius fuscus dentatus, from Labuan Sore in the lowlands near Parigi, and Tarsius pumilus, from Rano Rano, in the central highlands.

To T. pumilus they referred two specimens from Gimpu, in the nearby lowlands, in addition to the type. Tarsius pumilus is a valid species, but the Gimpu specimens do not belong to it (Musser & Dagosto, 1987); more recently, Merker & Groves (2006) have included them in their new species, Tarsius lariang.

7

Groves, Shekelle & Brandon-Jones- Taxonomic History of The Tarsiers

Sody (1949:138-143) adamantly assigned Sulawesi and Western tarsiers to separate species, but could not decide whether those from the Philippines were identical to either or neither of them. He used T. fuscus for the Sulawesi tarsier, and T. bancanus for the Western. In T. bancanus he recognised T. b. bancanus (synonym saltator) and T. b. borneanus, without mentioning Chasen’s natunensis. In T. fuscus he recognized T. f. fuscus from Minahassa, T. f. dentatus from central Sulawesi (“very weak”), a new subspecies T. f. pelengensis from Peleng (“we must acknowledge that we are describing a very poor race”), and T. f. sangirensis from Great Sangir (“a very good race”). It was Hill (1953a,b) who first definitively split Tarsius into three species, revising their nomenclature (Hill, 1953a) and describing in detail the striking differences in their tails (1953b) His revision was followed for over 30 years, and has been the basis for all subsequent treatments. Niemitz (1984a) reviewed the differences between the three species, commenting on the synonymy and refusing to recognize any subspecies except for T. s. spectrum and T. s. pumilus, and T. b. bancanus and T. b. borneanus. Contrary to Hill (1953a), however, Niemitz indicated Minahassa as the type locality of T. spectrum (1984:13). Musser & Dagosto (1987) further reviewed some of the nomenclatorial history of tarsiers. They accepted that Linnaeus’s Simia syrichta was a Philippine tarsier, but were skeptical of the association of Pallas’s Lemur spectrum with Sulawesi tarsiers although provisionally prepared to accept Hill’s (1953a) opinion. They revised the differences between T. bancanus, T. syrichta and T. spectrum, indicating that the first two are somewhat more closely related than either is to the third. Most importantly, Musser & Dagosto (1987) pointed out that the type specimen of Tarsius pumilus represents a valid species, and that Miller & Hollister (1921) had muddied the waters by inexplicably assigning to it the two Gimpu specimens (which are juveniles of what Musser and Dagosto regarded as the one and only lowland species, T. spectrum). Their demonstration that T.

8

pumilus is a valid species raised the number recognized since Hill (1953a) from three to four. Feiler (1990) proposed raising T. sangirensis Meyer, 1897 to specific rank. Niemitz et al. (1991) described a new species, Tarsius dianae, from central Sulawesi without, however, comparing their new taxon to T. fuscus dentatus, even though the type localities of these two taxa are separated by only about 80 km. Groves (1998) produced a preliminary revision of tarsiers, corroborating the provisional findings of Musser & Dagosto (1987) that the Philippine species, T. syrichta, is closer to T. bancanus, but nonetheless very distinct, while T. pumilus is closer to T. spectrum. Discriminant analysis of admittedly small samples confirms that, within T. spectrum (understood broadly), samples from the mainland, Selayar (a single specimen), Peleng and Sangihe are all distinct, and the latter at least is probably a distinct species. On the mainland, northern and central Sulawesi samples separate, if less clearly. Finally Merker and Groves (2006) described another new species from Sulawesi: Tarsius lariang from Gimpu, Central Sulawesi. This finally resolved the true identity of the two juveniles from Gimpu which had been spuriously associated by Miller and Hollister with Tarsius pumilus, and unceremoniously expelled from that species by Musser and Dagosto. Generic names for tarsiers The earliest generic name for tarsiers is Tarsius Storr, 1780, based on Lemur tarsier Erxleben, 1777, hence on Buffon’s tarsier, a Sulawesi tarsier (see above). The next name, Macrotarsus Link, 1795, was likewise based on Buffon’s tarsier. E. Geoffroy St. Hilaire (1812) also listed Tarsius, in his Strepsirrhini, citing no sources for the generic name but listing two species, T. spectrum (referring to Buffon and to Pallas) and T. fuscomanus (referring to Fischer). Gray (1821) included tarsiers in his order Heteronychae of the class Quadrumana; other members of the order were the lemurs and lorises. His tarsiers belonged to the family Loridae, and were placed in two genera:

Primates of The O riental Night

Tarsier, Tarsius. Geoff. Lemur tarsium Pallas. Rabienus. Gray. Lemur spectrum Pallas. The first of these two genera is incorrectly ascribed to E. Geoffroy and awarded, as type species, a non-existent name of Pallas’s. Rabienus, based on Pallas’s genuine name, is a junior subjective synonym of Tarsius. Swainson (1835) described a genus Cephalopachus, and Lesson (1840) described Hypsicebus, both erected for T. bancanus Horsfield, 1824. These names are available if the Western tarsier is regarded as generically distinct from Sulawesi tarsiers. Presuming Lemur tarsier Erxleben, 1777 was correctly identified as a Philippine tarsier, Groves (1998) concluded Tarsius Storr is available for a genus containing Philippine tarsiers, leaving Rabienus Gray for Sulawesi tarsiers; but as L. tarsier Erxleben actually refers to a Sulawesi tarsier (see above), Groves’s (1998) conclusion is error. No generic name is available for Philippine tarsiers. DISCUSSION Given this history, Tarsius spectrum is a junior objective synonym of T. tarsier. At first glance, the loss of this name, that has been used for tarsiers for over two centuries, seems regrettable. There is a silver lining, however, since the name T. spectrum is also associated with more taxonomic ambiguity than any other tarsier nomen. At one time or another, virtually all species of tarsier were referred to as Tarsius spectrum, though not always at the same time. Consequently, Clark’s (1924) “Notes on the living tarsier (Tarsius spectrum)” is an account of T. bancanus borneanus. Likewise, Woollard’s (1925) monograph “Anatomy of Tarsius spectrum” is not an Eastern tarsier, but a Western tarsier. Adding further to the confusion, Eastern Tarsiers were not generally referred to as T. spectrum until Hill (1955). Prior to that time, Eastern tarsiers were referred to by a plethora of names, but most often as T. fuscus. Thus, it is no overstatement to say that, prior to Hill, the majority of references to T. spectrum referred to something other

than an Eastern tarsier, while nearly all references to Eastern tarsiers used a name other than T. spectrum. A further consolation is that it turns out that not one, but a number of different species inhabit Sulawesi, each restricted to a particular part of the island. It will possibly cause less confusion if the name Tarsius spectrum applies to none of them than if it is applied to different species, by different authors. ACKNOWLEDGEMENTS The authors were introduced by way of the workshop Primate Taxonomy for the New Millenium and thanks are due to the organizers, Russell Mittermeier, Don Melnick, and John Oates, for bringing the three of us together to discuss tarsier taxonomy. Marian Dagosto reviewed a draft of this manuscript. C.G. offers grateful thanks to Boeadi, Paula Jenkins, Chris Smeenk, Dick Thorington and Guy Musser for access to material under their care; to Guy Musser for useful discussions; and to Marian Dagosto for sending me measurements of additional specimens. M.S. acknowledges the following: This material is based on work supported by the National Science Foundation under Grant No. INT 0107277. Much of the work for this was completed while receiving funding from the Margot Marsh Biodiversity Foundation. Additional collaboration between between M.S. and C.G. was facilitated by the Gibbon Foundation, which sponsored the “Indonesian Prosimian Workshop” in Jakarta, Indonesia. Portions of this appeared in the dissertation of M.S. and I thank Washington University in St. Louis and my thesis committee. REFERENCES Audebert, JB. 1797. Histoire Naturelle des Singes et des Makis. Paris: Desray. Bearder, SK, PE. Honess & L. Ambrose. 1994. Species diversity among galagos with special reference to mate recognition. Pp 1-22 in Alterman, L, Doyle, GA & Izard, MK (eds.),

9

Groves, Shekelle & Brandon-Jones- Taxonomic History of The Tarsiers

Creatures of the Dark: the Nocturnal Prosimians. New York: Plenum Press. Buffon, GLL. Comte de. 1765. Histoire Naturelle, Générale et Particulière. Vol. 13. Paris: Imprimerie du Roi. Cabrera, A. 1910. On the specific names of certain Primates. Annals & Magazine of Natural History (8) 6:617-618. Cabrera, A. 1923. On the identification of Simia syrichta Linnaeus. Journal of Mammalogy, 4:89-91. Chasen, FN. 1940. A handlist of Malaysian mammals. Bulletin of the Raffles Museum, Singapore, 15:I-xx, 1-209. Clark, W. & E. LeGros 1924. Notes on the living tarsier (Tarsius spectrum). Proceedings of the Zoological Society of London, 217-223. Desmarest, A. 1804. Tarsier. Nouveau Dictionnaire d’Histoire Naturelle, Appliquée aux Arts, à l’Agriculture, à l’Economie rurale et domestique, à la Médecine, etc. par une Société de Naturalistes et d’Agriculteurs. Paris: Deterville. Elliot, DG. 1910. On the genus Presbytis Esch., and ‘le Tarsier’ Buffon, with descriptions of two new species of Tarsius. Bulletin of the Mamerican Museum of Natural History, 28:151-154. Erxleben, JCP. 1777. Systema Regni Animalis per Classes, Ordines, Genera, Species, Varietates, cum Synonymia et Historia Animalium. Classis I. Mammalia. Leipzig: Weygand. Feiler, A. 1990. Ueber die Säugetiere der Sangiheund Talaud-Inseln der Beitrag A.B.Meyers für ihre Erforschung (Mammalia). Zoologische Abhandlungen der Staatlisches Museum für Tierkunde in Dresden, 46:75-94. Fischer, G. 1804. Anatomie der Maki. Frankfurt a.M.: Andrea. Fitzinger, LJ. 1870. Revision der Ordnung der Halbaffen oder Aeffer (Hemipitheci). II. Abtheilung. Familie der Schlafmaki’s (Stenopes), Galago’s (Otolicni) und Flattermaki’s (Galeopitheci). Sitzungsberi-

10

chte der Mathematisch Naturwissenschaftlichen Classe der Kaiserlichen Akademie der Wissenschaften, Wien, 62, 1:685-783. Geoffroy St.Hilaire, E. 1796. Mémoire sur les rapports naturels des makis Lemur, L. et description d’une espèce nouvelle de mammifère. Magasin encyclopèdique (2) 1:20-50. Geoffroy St.Hilaire, E. 1812. Suite au tableau des Quadrumanes. Seconde Famille. Lémuriens. Strepsirrhini. Annales du Muséum d’Histoire Naturelle, Paris, 19:156-170. Geoffroy St.Hilaire, I. 1851. Muséum d’Histoire Naturelle: Catalogue méthodique de la Collection des Mammifères, de la Collection des Oiseaux et des Collections annexes. Paris: Gide & Baudry. Gmelin, JF. 1788. Caroli a Linné, Systema Naturae, 13th edition. Leipzig: G.E.Beer. Gray, JE. 1821. On the natural arrangement of vertebrose animals. London Medical Repository, 15:296-310. Groves, CP. 1998. Systematics of tarsiers and lorises. Primates, 39:13-27. Heude, PM. 1898. Etudes odontologiques: Quatrièeme Partie. Quadrumanes. Chap.I, Lémuriens, Tarsiens, Galéopitheciens et Cébiens. Mémoires concernant l’Histoire Naturelle de l’Empire Chinois, 4:155-208. Hill, WCO. 1953a. Note on the taxonomy of the genus Tarsius. Proceedings of the Zoolo-gical Society of London, 123:13-16. Hill, WCO. 1953b. Caudal cutaneous specializations in Tarsius. Proceedings of the Zoological Society of London, 123:17-26. Hill, WCO. 1955. Primates: Comparative Anatomy and Taxonomy. II. Haplorhini: Tarsioidea. Edinburgh: Edinburgh University Press. Horsfield, T. 1824. Zoological Researches in Java, and the Neighbouring Islands. London: Kingsbury, Parbury & Allen. International Commission on Zoological Nomenclature. 1999. International Code of Zoological Nomenclature, Fourth Edition.

Primates of The O riental Night

London: International Trust for Zoological Nomenclature. Kerr, R. 1792. The Animal Kingdom, or Zoological System, of the Celebrated Sir Charles Linnaeus. Class I. Mammalia. London: J.Murray and R.Faulder. Lesson, RP. 1840. Species des Mammifères bimanes et quadrumanes. Paris. Link, HF. 1795. Beyträge zur Naturgeschichte, Rostock & Leipzig, 1974-1801, Bd. I. Stck. 2, pp. 65-66. Linnaeus, C. 1758. Systema Naturae per Regna Tria Naturae, secundum Classes, Ordines, Genera, Species, cum Synonymis, Locis. I. 10th ed. Stockholm: Laurent Salvi. MacKinnon, J. & K. MacKinnon. 1980. The behavior of wild Spectral Tarsiers. International Journal of Primatology, 1:361-379. Merker, S. & CP. Groves. 2006. Tarsius lariang: a new Primate species from Western Central Sulawesi. International Journal of Primatology, 27: 465-485. Meyer, AB. 1897. Säugethiere vom Celebes- und Philippinen-Archipel, I. Abhandlungen und Berichte der Kaiserlich Zoologische und Anthropologische-Ethnologische Museum zu Dresden, 6:I-VIII, 1-36. Miller, GS. 1910. Descriptions of two new genera and sixteen new species of mammals from the Philippine Islands. Proceedings of the United States National Museum, 38:391404. Miller, GS. & N. Hollister. 1921. Twenty new mammals collected by H.C.Raven in Celebes. Proceedings of the Biological Society of Washington, 34:93-104. Musser, GG. & M. Dagosto. 1987. The identity of Tarsius pumilus, a pygmy species endemic to the montane mossy forests of Central Sulawesi. American Museum Novitates, 2867:1-53. Musser, GG. & M. Dagosto. 1987. The identity of Tarsius pumilus, a pygmy species endemic to the montane mossy forests of Central

Sulawesi. American Museum Novitates, 2867:1-53. Niemitz, C. 1979. Relationships among anatomy, ecology, and behavior: a model developed in the Genus Tarsius, with thoughts about phylogenetic mechanisms and adaptive interactions. In M. E. Morbeck, H. Preuschoft, & N. Gomberg (Eds.), Environment, Behavior, and Morphology: Dynamic Interactions in Primates (pp. 119-137). New York: Gustav Fischer. Niemitz, C. 1984a. Taxonomy and distribution of the genus Tarsius Storr, 1780. Pp 1-16 in C. Niemitz, ed., Biology of Tarsiers. Stuttgart: Gustav Fischer Verlag. Niemitz, C. 1984b. Vocal communication of two tarsier species (Tarsius bancanus and Tarsius spectrum). In C.Niemitz (ed.), Biology of Tarsiers, 129-141. Gustav Fischer Verlag, Stuttgart & New York. Niemitz, C, A. Nietsch, S. Warter & Y. Rumpler 1991. Tarsius dianae: a new primate species from Central Sulawesi (Indonesia). Folia primatologica, 56:105-116. Nietsch, A. & ML.Kopp. 1998. Role of vocalization in species differentiation of Sulawesi Tarsiers. Folia primatologica, 68(suppl.1): 371-378. Nietsch, A. & C. Niemitz 1993. Diversity of Sulawesi Tarsiers. Deutsche Gesellschaft für Säugetierkunde, 67. Hauptversammlung, 4546. Pallas, PS. 1778. Novae Species Quadrupedum e Glirium Ordine cum Illustrationibus Variis Complurium ex Hoc Ordine Animalium. Erlangen: Wolfang Walther. Pennant, J. 1771. Synopsis of Quadrupeds. Chester: J.Monk. Pennant, J. 1781. History of Quadrupeds. London: B.White. Schreber, JCD.von. 1778. Die Säugethiere in Abbildungen nach der Natur. Leipzig: T.D.Weigel. Shekelle, M., SM. Leksono, LLS. Ichwan, & Y. Masala. 1997. The natural history of the tarsiers of

11

Groves, Shekelle & Brandon-Jones- Taxonomic History of The Tarsiers

North and Central Sulawesi. Sulawesi Primate Newsletter, 4(2):4-11. Smit, P., APM. Sanders & JPF. van der Veer. 1986. Hendrik Engel’s Alphabetical List of Dutch Zoological Cabinets and Menageries. 2nd ed. Amsterdam: Rodopi B.V. Sody, HJV. 1949. Notes on some Primates, Carnivora, and the Babirusa from the Indo-Malayan and

12

Indo-Australian regions. Treubia, 20:121190. Storr, GLC. 1780. Prodromus Methodi Mammalium. Tubingen. Swainson, W. 1835. On the Natural History and Classification of Quadrupeds. London. Woollard, HH. 1925. The anatomy of Tarsius spectrum. Proceedings of the Zoological Society of London, 70, 1071-1184.

Primates of The O riental Night

DISTRIBUTION AND BIOGEOGRAPHY OF TARSIERS Myron Shekelle Center for Biodiversity and Conservation Studies, Faculty of Mathematics and Natural Science University of Indonesia, Depok 16421, Indonesia. Email: [email protected] ABSTRACT Three clearly distinct taxa of tarsiers each inhabit a distinct biogeographic region: Western tarsiers, from island areas of Sundaland; Philippine tarsiers from Greater Mindanao; and Eastern tarsiers from Sulawesi and nearby islands. Multiple species and / or subspecies have been described from each region, and continued investigations into the alpha taxonomy of each group are warranted. Within each region tarsiers currently have discontinuous distributions, at least partly the result of anthropogenic habitat alterations. Their presence in a variety of primary and secondary habitats indicates that the historical distribution of tarsiers may have once been much more continuous, perhaps limited more by elevation and ocean barriers than by variation among lowland habitats. The distributions of Philippine and Eastern tarsiers conform well to Ice Age landmasses. The distribution of Western tarsiers does not, and is hypothesized to indicate a Holocene range expansion. A model of historical biogeography is here presented, wherein events in the Miocene led to the isolation of the three species groups. Tarsiers’ last appearance in the fossil record of mainland Asia also occurs during the Miocene.

Keywords: Tarsius, Taxonomy

INTRODUCTION All known tarsier taxa are distributed either allopatrically or parapatrically; there is not a single known case of sympatric tarsiers. Understanding their geographic distributions, therefore, is crucial for understanding tarsier taxonomy. Extant tarsiers have a curious distribution on a scattering of southeast Asian islands. They are found on both sides of Wallace’s Line, which approximates the separation between the Asian and Australian biotic communities, and although fossil tarsiers are found on mainland Asia, none exist there today (Hill 1955, Niemitz 1984, Musser & Dagosto 1987). Hill (1953, 1955) classified tarsiers into three species, all in the genus Tarsius and each endemic to a distinct biogeographic region: Tarsius syrichta Linnaeus, 1758, from islands of the southern Philippines; Tarsius spectrum Pallas, 1778 from Sulawesi and surrounding islands; and Tarsius bancanus Horsfield, 1824, from various islands of the Sunda Shelf including Borneo, southern Sumatra, Bangka, Belitung, the Karimata Islands, the South Natuna Islands, and several smaller islands. BrandonJones et al. (2004) provided an argument, which was greatly elaborated on by Groves et al. (this volume), that T. tarsier Erxleben, 1777 is a senior subjective synonym of T. spectrum (Fig. 1).

There are several conspicuous anatomical features that are diagnostic of each species group. Relative eye size is largest in Western tarsiers, smallest in Eastern tarsiers, and intermediate in Philippine Tarsiers, while relative ear length is largest in Eastern tarsiers, smallest in Western tarsiers, and intermediate in Philippine tarsiers (Niemitz 1984). The mid-tarsal segment appears naked or nearly naked in Philippine tarsiers, but is well-furred in both Eastern and Western tarsiers. The tails of all tarsiers are long and superficially rat-like, but the amount of fur varies among the species groups. Eastern tarsiers have the most fur on the tail, its appearance being almost like a bottlebrush. Philippine tarsiers have the least fur on the tail, with very short, sparse hairs that are almost invisible unless viewed at close range (although the tail tufts in a large collection of T. syrichta carbonarius from Mindanao, in Chicago’s Field Museum, approach the condition seen in T. bancanus, unpublished data). Western tarsiers are intermediate, having a noticeable tuft of fur on the tail, but not nearly so much as do Eastern tarsiers (Fig. 2). Hill (1955) accepted subspecies within each of the three species he recognized, but remarked that some of these were of dubious distinctiveness. Niemitz (1984) also used the three species taxonomy, but he synonymized the bulk of Hill’s subspecies,

13

Myron Shekelle - Distribution and Biogeography of Tarsiers

Figure 1: Distribution of Extant Tarsiers. The northwestern boundary of tarsiers in Sumatra is figured here as the Musi River, but this is speculative. The actual distribution is less continuous than figured here, tarsiers being rare or nonexistent in high elevations and tarsiers having gone locally extinct in many areas of intense human usage.

accepting only two subspecies of T. bancanus and two subspecies of T. tarsier (=T. spectrum). Several authors have noted that, based upon acoustic and biogeographic evidence, numerous other unrecognized taxa of tarsiers are likely to exist within T. tarsier (MacKinnon & MacKinnon 1980; Niemitz et al. 1991; Nietsch & Niemitz 1993; Nietsch & Kopp 1998; Nietsch 1999; Shekelle 2003). The trend after Niemitz has been to recognize each of Hill’s subspecies of T. tarsier as a distinct species, including: T. pumilus Miller & Hollister, 1921, (Niemitz 1985; Musser & Dagosto 1987; Groves 1998, 2001), T. sangirensis (Feiler 1990; Shekelle et al. 1997; Groves 1998, 2001), and T. pelengensis (Groves 2001). Additionally,

14

Niemitz et al. (1991) described a new taxon, T. dianae, from central Sulawesi. However, Shekelle et al. (1997) surveyed tarsiers at the type localities of both T. dianae and T. dentatus, Miller & Hollister, 1921 and found the same acoustic form at both locations, indicating that T. dianae is quite likely a junior subjective synonym of T. dentatus. Other forms, not included in Hill’s list of T. tarsier subspecies have since been described, including T. lariang (Merker and Groves 2006) from central Sulawesi and T. sp. (Shekelle et al. in press), an insular population from Siau Island, North Sulawesi (Table 1). Tarsiers from Sulawesi are now recognized as a species complex with at least seven species and

Primates of The O riental Night

Western tarsier

Philippine tarsier

Eastern tarsier

Figure 2: Anatomical variation among tarsier species groups. Western tarsiers have the largest eyes, shortest ears, and longest legs and hands. Eastern tarsiers have the smallest eyes, longest ears, and shortest legs and hands. Philippine tarsiers are intermediate in all of these. Eastern tarsiers have the furriest tail, Philippine the least furry, and Western tarsiers are intermediate. The mid-tarsal segment of the hind foot is nearly naked in Philippine tarsiers, but is wellfurred in both Western and Eastern tarsiers (adapted from Shekelle 2003). Photos Myron Shekelle (c) 2008, except where noted.

15

Myron Shekelle - Distribution and Biogeography of Tarsiers

Table 1: Review of Tarsier Taxonomy. Hill 1955 T. syrichta syrichta T. s. carbonarius T. s. fraterculus T. bancanus bancanus T. b. borneanus T. b. saltator

T. syrichta

Musser & Dagoso 1987 T. syrichta

*

**

*

**

T. b. bancanus

T. bancanus

T. b. bancanus

T. b. borneanus *

**

T. b. borneanus T. b. saltator

T. b. natunensis T. spectrum T. s. sangirnesis T. s. pumilus

*

**

T. spectrum *

T. spectrum **

T. b. borneanus T. b. saltator T. b. natunensis T. spectrum T. sangirnesis

T. b. natunensis T. spectrum T. sangirnesis

T. b. borneanus T. b. saltator T. b. natunensis T. spectrum T. sangirnesis

T. s. pumilus * *

T. pumilus

T. pumilus

T. pumilus

T. pumilus

** **

*** T. pelengensis

T. dentatus T. pelengensis

T. dentatus T. pelengensis

T. s. dentatus T. s. pelengensis

Niemitz 1984

**

Groves 2001 T. syrichta

Brandon-Jones 2004 T. syrichta

T. syrichta

T. s. carbonarius T. s. fraterculus T. b. bancanus

T. s. carbonarius T. s. fraterculus T. b. bancanus

T. dianae***

This Paper

T. lariang T. sp. (Siau)

* **

Niemitz found museum specimen variation to be insignificant among several taxa accepted by Hill Musser and Dagosto found museum specimen variation to be an insufficient basis for determining the validity of several taxa accepted by Hill *** cited Shekelle et al. 1997 for noting a likely conflict between T. dianae and T. dentatus. **** according to Brandon-Jones et al., a taxon “whose recognition is doubtful and requires further investigation”

probably more, and it now seems plausible that each of Hill’s three tarsier species may be a cluster of closely related taxa. To sidestep the current debates on the numbers of taxa and their formal names, I refer to Hill’s three species—T. syrichta, T. tarsier (i.e. T. spectrum), and T. bancanus—by the common names, Philippine, Eastern, and Western tarsiers, respectively, with the assumption that each of these is monophyletic and that each might be a constellation of related taxa, that is, species groups. The assumption of monophyly within species groups has not been exhaustively examined, but is consistent with the results of Musser & Dagosto (1987) and Groves (1998). Musser & Dagosto (1987) indicated that their morphologic analysis of museum specimens supported two distinct clades of tarsiers, a T. bancanus / T. syrichta clade, and a T. tarsier / T. pumilus clade. Groves (1998) went further and

16

suggested generic separation of the Eastern tarsiers from a Philippine-Western tarsier clade based upon his own analyses of morphological variation. Genetic data (Shekelle et al. 2001; Meireles et al. 2003) indicate that generic separation of Philippine and Western tarsiers might also be warranted if one were to accept a time-based classification scheme, such as proposed by Goodman et al. (1998) and Groves (2001). The value of designating new tarsier genera is that each species group could be addressed by a formal name (as opposed to the common names, Eastern, Western, and Philippine tarsiers, for example); and, if taxonomy is to be an information retrieval system, it might symbolically emphasize the underappreciated variation among tarsier species groups in taxonomy, behavior, and ecology (see Shekelle 2003). Nevertheless, for the time being I prefer to retain a single genus for extant tarsiers for the reasons that:

Primates of The O riental Night

(1) Tarsius as used by Hill is a clearly defined monophyletic clade with only three species; (2) taxonomic variation in addition to that accepted by Hill, and the question of whether or not Hill’s subspecies should be elevated to full species level is the focus of much ongoing research and debate; and (3) the monophyly of each of Hill’s three species groups, although seemingly sensible, has not been rigorously tested. Thus, the dangers of classifying tarsiers in one or two new genera at this point are that ongoing research might show either or both of them to be either monotypic or not monophyletic. Previous Estimates Before estimating the historical distribution of tarsiers, I review previous work to examine how current knowledge helps resolve discrepancies, errors, and omissions. Regarding erroneous reports there are claims of tarsiers having originated from outside the range listed by Hill (1955), Niemitz (1984), and Musser & Dagosto (1987), including the islands of Luzon, Ambon, Savu, Java, and even Madagascar. Cabrera (1923) questioned the accuracy of Camel’s (1705) report of tarsiers from Luzon, and there are no longer serious discussions of Luzon being within the historical range of tarsiers. Fischer (1804) reported that tarsiers in his study were allegedly from Madagascar, but he went on to infer that Madagascar had been confused with Makassar, a port city on Sulawesi. Hill (1955) discounted Pallas’s (1778) report of tarsiers from Ambon. Reports of tarsiers from Java and nearby Savu are still sometimes treated seriously by researchers (e.g., Niemitz 1984). Savu (= Sabu) is south of Flores in the Lesser Sunda Island chain. The nearest known tarsier populations are those in South and Southeast Sulawesi, from which Savu is separated not only by hundreds of kilometers of open ocean in the Flores Sea, but also by the island of Flores itself. It is separated from Sumatran tarsiers and hypothetical Javan tarsiers by the islands of Bali, Lombok, Sumbawa, and Sumba. Thus any dispersal for tarsiers to Savu would not only have to cross expanses of open ocean, but would also have to skip over one or

more intervening islands along the route. It would give tarsiers a strikingly discontinuous distribution – much more so than they already have. Given the lack of any additional reports of tarsiers from Savu, it seems prudent to assume that Savu is outside the range of tarsiers. Jentink (1892) provides two Javan locales for tarsiers, Surabaya and Preanger (near Bandung). The possibility of tarsiers on Java is more plausible than Savu and warrants careful consideration. Tarsiers are present on Sumatra. Sumatra and Java were a single landmass as recently as the last ice age. So, it is conceivable that tarsiers had a historical distribution on Java. Contrary to this, however, there is no evidence of tarsiers on Java today. This leaves three possibilities: 1.Tarsiers are present on Java today but their presence has gone unnoticed. 2.Tarsiers are locally extinct on Java today, but were present in Jentink’s time. 3.The historical distribution of tarsiers does not include Java and Jentink’s records are erroneous, possibly owing to specimens transported from elsewhere and purchased at the localities provided by Jentink. I argue that the weight of the evidence argues for the third possibility. Regarding the first possibility listed above, tarsiers are small, nocturnal, and cryptic by nature, such that the possibility that their presence has gone unnoticed in one or more areas should not be discounted— at least until an experienced tarsier field biologist has conducted surveys. Tarsiers are recorded in the wet lowlands of Way Kambas National Park, just across the Sunda Straits from Java in the province of Lampung on the island of Sumatra (Yanuar and Sugardjito 1993). The most extensive pristine lowland habitat on Java today is Ujung Kulon National Park, at the extreme western end of Java, surrounded on three sides by the Sunda Straits and less than 200 km from Way Kambas. If the historical distribution of tarsiers were to include Java, it is highly probable that their distribution would include the Ujung Kulon area. Ujung Kulon National Park has been the subject of several field surveys, including some that have used

17

Myron Shekelle - Distribution and Biogeography of Tarsiers

nocturnal camera trapping, without any evidence of tarsiers (Whitten et al. 2002). Indeed, one experienced tarsier field biologist has been surveying Ujung Kulon National Park for more than 20 years with no evidence whatsoever of tarsiers, while he was able to easily locate them in Bukit Barisan Selatan National Park, less than 200 km from Ujung Kulon on the southern end of Sumatra; his opinion is that the evidence is overwhelming that tarsiers are not present in Ujung Kulon (Haerudin R. Sadjudin of Yayasan Cipta Citra Lestari, personal communication). The possibility that tarsiers have gone locally extinct on Java might seem plausible, even likely, if one adopted a misconception that tarsiers are relictual taxa clinging to survival in their isolated island homes. Several lines of evidence counter this assumption, however. First, nearly 40 years of field research contradict the misconception that tarsiers, being a relictual taxon, are therefore teetering on the precipice of extinction. On the contrary, tarsiers have been found to be a weedy animal that exists in high densities in a remarkable array of habitats and varying human use (Niemitz 1984, Merker 2003, Shekelle 2003). Indeed, Island Biogeography Theory predicts that it is unlikely for tarsiers to have gone locally extinct on an island as large as Java, while they persist on tiny islands such as Sangihe, Siau, Serasan, and Subi. Nevertheless, the human population density of Java is among the highest in the world, and several Holocene extinctions are known to have occurred there, such as the Javan tiger. The wet lowland habitat of Ujung Kulon is suitably pristine to support the only remaining population of the Javan rhinoceros on Java. It seems unlikely that tarsiers, with a home range of a few hectares at most, could have gone locally extinct from Ujung Kulon, and indeed all of Java, while rhinoceroses remain. Thus, strong evidence exists that tarsiers are currently absent from Ujung Kulon National Park. By comparison with a much more vulnerable taxon, the Javan rhino, we can extrapolate that it is unlikely that tarsiers exist on Java today, or did so in the past. The most reasonable assumption is that the historical distribution in Sundaland stops at the Sunda Straits, and the two records of tarsiers on Java in Jentink—

18

the only evidence of tarsiers on Java— are erroneous, and direct examination of the specimens might shed further light. Hill’s (1955) distribution map of the genus Tarsius (upon which most subsequent research has been based) includes a few discrepancies from what is known today. For instance, he shows T. bancanus natunensis as being from the North Natuna Islands, and absent from the South Natuna Islands, when in fact, the actual distribution is the reverse (Chasen 1940). Hill omits Basilan, off the southwestern tip of Mindanao, from the range of the Philippine Tarsier, but Musser & Dagosto (1987) list specimen 35256 from the Museum of Comparative Zoology, Harvard University (MCZ) as being from Basilan. Likewise, Biliran Island, off the northern tip of Leyte is excluded from the range of Philippine Tarsiers, but NeriArboleda et al. (2002) list Biliran as having tarsiers. Hill’s map shows a spotty, discontinuous distribution of Eastern tarsiers that entirely omits the southwestern peninsula of Sulawesi, even though Hill himself identified the type locality of T. spectrum as Makassar (= Ujung Pandang). Indeed, evidence from museum specimens indicates that the distribution of Eastern tarsiers is far more continuous than appears in Hill’s map (Musser & Dagosto 1987), and subsequent field surveys have found tarsiers almost everywhere they have been looked for (e.g. MacKinnon & MacKinnon 1980; Nietsch 1999; Nietsch & Kopp 1998; Nietsch & Babo 2001, Nietsch & Burton 2002; Shekelle 2003, Shekelle & Leksono 2004). Likewise, Tarsius bancanus borneanus is marked by Hill (1955) as present in coastal regions of Borneo, but not the central regions (generally of higher elevation), perhaps because this species was thought to be present only in extreme lowlands (e.g. Clark 1924). Gorog and Sinaga’s (this volume) capture of a tarsier from the montane interior of Borneo contradicts both Clark and Hill. Distribution maps in Hill (1955), Niemitz (1984), and Musser and Dagosto (1987) all show the distribution of tarsiers on the west coast of Sumatra as extending northward about to the city of Bengkulu. On the east coast, however, Niemitz (1984) and Musser and Dagosto (1987) show the distribution stopping at what appears to be the Musi River, while Hill’s map

Primates of The O riental Night

shows it extending further to what is possibly the Hari River. All three depict a predominantly coastal distribution. Hard evidence to resolve this discrepancy is lacking, but anecdotal reports from locals indicate that tarsiers are not present in the vicinity of the Hari River. Musser & Dagosto (1987) omit the Buton Island chain from the distribution of Eastern Tarsiers, but subsequent field surveys by Nietsch & Burton (2002) reported tarsiers from Kabaena and Buton. They are presumably on Muna, as well, but owing to deforestation, Burton was unable to locate any (Nietsch, personal communication). Musser & Dagosto (1987) omit Subi Island, although it is biogeographically linked to Serasan (see Banks 1949). Suroso Leksono reported that during surveys in 2003, inhabitants of Subi claimed tarsiers existed there, although he did not see them himself (personal communication). An Hypothesis of The Historical Extent of Occurence In the following sections, I hypothesize the historical extent of occurrence for tarsiers. These hypotheses are meant to be refutable statements that can be corrected by future field surveys and more careful examination of available museum specimens. The hypotheses are based on the assumption that historical distributions on landmasses known to contain tarsiers are limited more by elevation than by habitat type, but regions of high elevation are not specifically excluded from the distribution maps for reasons of practicality. Tarsier densities vary by habitat type (see Merker 2003) and the hypothesis that the presence of tarsiers in various sub-optimal habitats, such as agroforestry and alang-alang, could represent sink populations has not been examined (Wright 2003). In many cases, evidence is lacking (e.g. the precise limit of the distributions of Western and Philippine tarsiers in the Jolo archipelago, the precise eastward limit of Eastern tarsiers in the Banggai Islands, etc.). In other cases, the data (or distribution maps, rather) are conflicting (e.g. the northwestern boundary of Western tarsiers on Sumatra being the Musi River or the Hari River, etc.).

Eastern Tarsiers To the east of Wallace’s Line, I hypothesize a historical distribution of Eastern tarsiers that is nearly continuous on Sulawesi and surrounding islands, including all land areas that were exposed during the Ice Ages (Fig. 3). Tarsius pumilus, a montane endemic, is known from three specimens collected between 1800-2200 m in elevation (Miller & Hollister 1921, Musser & Dagosto 1987, Maryanto & Yani 2004). Records of other Eastern tarsiers, of the T. tarsiercomplex, exist from sea level to 1500 m (MacKinnon & MacKinnon 1980; Musser & Dagosto 1987; Shekelle 2003). Eastern tarsiers have been recorded in almost all habitats except areas with dense human populations, areas of intensive agriculture where all potential sleeping sites have been cleared, and areas where pesticides and / or herbicides are used intensively (MacKinnon & MacKinnon 1980; Shekelle et al. 1997; Leksono et al. 1997). Any current gaps in the distribution of Eastern tarsiers are hypothesized to be recent events owing to human activity, such as habitat destruction. The range of these tarsiers extends to the offshore island groups of Togian (Nietsch & Niemitz 1993; Shekelle et al. 1997), Banggai (Sody 1949), Selayar (Musser & Dagosto 1987; Nietsch & Babo 2001), and Buton (Nietsch & Burton 2002). A chain of volcanic islands leads north from the northernmost tip of Sulawesi, and of these, at least Siau and Greater Sangihe are known to have tarsiers (Meyer 1897). The neighboring limestone islands of the Talaud chain do not, however (Yunus Masala, Forest Ranger, North Sulawesi, personal communication). Nor does it seem that tarsiers have crossed the narrow Salue Timpaus Strait to the Sula Islands—Taliabu, Mangole, and Sanana—as there are no records of tarsiers from there, although Whitten et al. (2002) indicate that these are geologically part of Sulawesi. Similarly, there is a chain of islands, the Bonerate Islands, which run southeast from Selayar into the Flores Sea toward the Lesser Sunda Islands, from which there are no reports of tarsiers. The same is true for the Tukang Besi Archipelago, southeast of Buton Island. Several species of Eastern tarsiers have been described. Reviews are found in Hill (1955), Niemitz

19

Myron Shekelle - Distribution and Biogeography of Tarsiers

(1984), Musser & Dagosto (1987), Groves (1998, 2001), Shekelle (2003) and Brandon-Jones et al. (2004). Excluding known synonyms, these include Tarsius tarsier Erxleben, 1777 (type locality most probably Makassar, South Sulawesi); T. sangirensis Meyer, 1897 (type locality, Greater Sangihe Island*, North Sulawesi); T. pumilus Miller & Hollister, 1921 (type locality, Rano Rano, Central Sulawesi); T. dentatus Miller & Hollister, 1921 (type locality, Labua Sore, Central Sulawesi); T. pelengensis Sody, 1949 (type locality, Peleng Island, Central Sulawesi), T. lariang Merker and Groves, 2006 (type locality, Gimpu, Central Sulawesi), and T. sp. (Shekelle et al. in press) (type locality, Siau Island, North Sulawesi) (Fig. 4). Several other populations probably warrant taxonomic separation (MacKinnon & MacKinnon 1980; Nietsch & Niemitz 1993; Shekelle et al. 1997; Nietsch & Kopp

1998; Groves 1998; Shekelle 2003; Brandon-Jones et al. 2004). Notably, Sulawesi is the only place with parapatrically distributed tarsiers (e.g. T. dentatus and T. lariang). *Sangi and Sangir are alternate spellings of Sangihe. **Laboea Sore, Labuan Sore, and Laboean Sore are all alternate spellings of Labua Sore. Philippine tarsiers The current distribution of Philippine tarsiers is thought to be spotty, but I hypothesize a historical distribution that was nearly continuous on islands that made up the ice age landmass of Greater Mindanao (Heaney 1985) (see also Neri-Arboleda 2002) (Fig. 5). This includes Mindanao, Samar, Leyte, and Bohol, as well as the smaller islands of Siargo, Dinagat, Basilan,

Figure 3: Distribution of Eastern Tarsiers—schematic (left) and satellite (right) views. The hypothesized historical range of these tarsiers extends as far west and north as Wallace’s Line (1), and includes the island of Sulawesi and the offshore island groups of Togian (2), Banggai (3), Selayar (4), Buton (5), and Sangihe (6). The northwestern limit of the range of Eastern Tarsiers lies between Sangihe Island and the Talaud Island chain to the northeast (7). Eastwards, it appears that the range is bounded by the Salue Timpaus Strait (8). From the southwestern peninsula, the distribution stops between Selayar Island and the Bonerate Islands (9), and from the southeastern peninsula, the boundary lies between the Buton Islands and the Tukang Besi Islands (10).

20

Primates of The O riental Night

Figure 4: Map of Eastern Tarsier type localities, with the names as they appeared in the original descriptions. The type localities of both T. dentatus and T. dianae are illustrated for reference, although they are now treated as synonyms.

Biliran, and many others. Although commonly thought of as an exclusively lowland taxon (e.g. Fulton 1939; Wharton 1950; Dagosto & Gebo 1997; Neri-Arboleda et al. 2002), Gorog & Sinaga (this volume) review evidence for Philippine tarsiers above 800 m. They have been recorded from a variety of primary and secondary habitats (Rickart et al. 1993; Dagosto & Gebo 1997; Neri-Arboleda et al. 2002). I find no records of tarsiers from Palawan, or other islands that extend from the northwest corner of Borneo, and neither Hill (1955), nor Niemitz (1984), nor Musser & Dagosto (1987) indicate that these regions have (or have ever had) tarsiers. Tarsiers have not crossed the narrow straits that separate Samar from Luzon, Leyte from Masbate, nor Bohol from Cebu. The Jolo Archipelago makes a logical dispersal corridor between Borneo and Mindanao, but I can find no records of tarsiers there, other than from Basilan near the southwestern tip of Mindanao (Musser & Dagosto 1987). Dagosto et al. (2003) state that tarsiers are absent from this archipelago, except for Basilan. Three subspecies of Philippine tarsiers have been described, but their taxonomic distinctiveness has been questioned (Hill 1955; Niemitz 1984; Musser

& Dagosto 1987; Groves 2001; Brandon-Jones et al. 2004). These include: Tarsius syrichta syrichta Linnaeus, 1758, (type locality, Samar); Tarsius syrichta fraterculus Miller, 1910, (type locality, Bohol), and T. s. carbonarius Heude, 1899 (type locality, Mindanao). Tarsiers from other islands are classified as T. s. syrichta, but this gives that taxon an illogically disjunct distribution, and other classifications are more probable. For instance, should T. s. carbonarius warrant taxonomic separation, then tarsiers from Basilan Island are more likely referable to T. s. carbonarius. Western tarsiers The Western tarsier has, perhaps, the most curious distribution of all in that it has an incomplete distribution on the ice age landmass, Sundaland (Fig. 6). Western tarsiers are recorded from several parts of Borneo, where I hypothesize a historical distribution that is nearly continuous. They have also been found in southern areas of Sumatra and a few smaller islands including Bangka, Belitung, the Karimata Islands, and the South Natuna Islands (Chasen 1940, Hill 1955, Musser and Dagosto 1987). Like other tarsiers,

21

Myron Shekelle - Distribution and Biogeography of Tarsiers

Figure 5: Distribution of Philippine Tarsiers—schematic (left) and satellite (right) views. Philippine Tarsiers have been found on islands that composed the ice age landmass, Greater Mindanao. These include Mindanao (1), Samar (2), Leyte (3), Bohol (4), and many smaller islands. Tarsiers have not been reported from Palawan (5) nor other islands that extend from the northwestern tip of Borneo. The northern extent of the range of these tarsiers is the narrow strait that separates Samar from Luzon (6). Philippine Tarsiers have not crossed the narrow expanse of ocean that separates Leyte from Masbate (7) to the north and Cebu (8) to the west. Similarly, tarsiers from Bohol have not crossed the narrow straits that separate Bohol from Cebu (9). Tarsiers are recorded from Basilan, off the southwestern tip of Mindanao, but their presence in the Jolo archipelago (between 10 and 11) is uncertain, although this is a logical dispersal corridor between Borneo and Mindanao. Wallace’s Line (12) demarcates the southern and eastern limits of the Philippine Tarsiers.

Western tarsiers have been recorded from a variety of primary and secondary habitats (Clark 1924, Fogden 1974, Niemitz 1979, Crompton and Andau 1986). As with Philippine tarsiers, Western tarsiers are most often recorded as a lowland species from sea level to 100200 m (e.g. Clark 1924), but Gorog and Sinaga (this volume) report a tarsier capture from 1200 m on Borneo. The entire extent of occurrence of Western tarsiers is on the Ice Age landmass Sundaland, but not all of Sundaland has tarsiers. For instance, tarsiers are absent from the Asian mainland, all areas of Sumatra except the southernmost tip, and the North Natuna Islands. Also, there are no credible accounts of tarsiers on Java or Bali. I hypothesize the northwestern boundary of tarsiers in Sumatra to be the Musi River, as did Musser and Dagosto (1987), but this has not been confirmed with surveys and other possibilities exist, such as the Hari River (Hill 1955). Four subspecies of Western tarsiers have been described, but again, their taxonomic

22

distinctiveness has been questioned (Hill 1953, 1955, Niemitz 1984, Musser and Dagosto 1987, Groves 1998, 2001, Brandon Jones et al. 2004). These include: Tarsius bancanus bancanus Horsfield, 1821, (type locality, Jebus (=Jeboos), on the northwest tip of the island of Bangka); Tarsius bancanus borneanus Elliot, 1910, (type locality, Borneo), T. b. saltator Elliot, 1910, (type locality, Belitung (=Beliton), and T. b. natunensis Chasen, 1940, (type locality, Serasan Island). Biogeographic Inferences Several authors have inferred that extant tarsiers are closely related taxa (e.g. MacKinnon and MacKinnon 1980, Musser and Dagosto 1987, Simons 2003), but that otherwise reasonable inference appears to be contradicted by genetic data that indicates that the origin of all three clades dates to the Miocene. Meireles et al. (2003) used nDNA for a molecular clock estimate of 5.6 mya for the split between Western and Philippine tarsiers. Preliminary results of mtDNA sequence data reported by Shekelle et al. (2001) that

Primates of The O riental Night

Figure 6: Distribution of Western Tarsiers—schematic (left) and satellite (right) views. Western Tarsiers presumably have a nearly ubiquitous historical distribution on Borneo (1). They have also been found on southern parts of Sumatra (2), Bangka (3), Belitung (4), and the South Natuna Islands (5). They have not been recorded from the Asian mainland (6), most of Sumatra (7), the North Natuna Islands (8), nor are there reliable reports from Java (9). The eastern extent of their distribution is Wallace’s Line (10). To the north, their distribution is limited somewhere in the Jolo Archipelago (11, 12), and somewhere before the Philippine island of Palawan (13). Their westward distribution corresponds with the drowned riverbed of the ice age Sunda River (14), and one of its modern day tributaries, shown here as the Musi River (15). Western Tarsiers did not cross the Sunda Straits (16), nor did they disperse south of Borneo (17).

found Philippine, Eastern, and Western tarsiers to be an unresolved trichotomy that most likely dates to the middle Miocene. In the preliminary analysis, genetic distances among Philippine, Eastern, and Western tarsiers were nearly as great as the average genetic distances among hominoids. The average genetic distance among the three tarsier species groups was 0.1157, while the average genetic distance for Hylobates vs. Homo and Pan was 0.1272. Even allowing for unequal rates of evolution, it is most likely that Philippine, Eastern, and Western tarsiers are not particularly closely related taxa. Morley (1998) found evidence of biotic exchange across the Makassar Straits in palynological data at 17 mya, 14 mya, 9.5 mya, 3.5 mya, and 1 mya. The genetic data reported by Shekelle et al. (2001) are most consistent with tarsiers crossing the Makassar Straits at any of the three older dates identified by Morley, and least consistent with the two younger dates. Hall (2001) identified the most likely time for faunal exchange across the Makassar Straits as being about 10 mya, based upon reconstructions of tectonic activity. Mercer and Roth (2003) used a molecular

clock date to estimate the arrival of squirrels to Sulawesi as approximately 11.5 mya, indicating that at least some of the small terrestrial mammals from Sulawesia are as ancient as I hypothesize for tarsiers. The last known record of tarsiers from mainland Asia is a Miocene fossil from Thailand (Ginsburg and Mein 1986). There is an intriguing synchronicity, therefore, between the fossil record, the estimated divergence of extant tarsiers based on mtDNA, the molecular clock date for squirrels, and two independent predictions for biotic exchange across the Makassar Straits, all during the middle Miocene. It is worth noting that both Musser and Dagosto (1987) and Groves (1998) argued that Philippine and Western tarsiers form a clade relative to Eastern tarsiers based on morphological data. Such a tree topology is consistent with this hypothesis. Results of the preliminary mtDNA analysis found an unresolved trichotomy, but this does not imply a trifurcation. Some inferences about tarsier biogeography within species groups can be drawn from the estimated historical distributions. Notably, the incomplete distribution of Western tarsiers on Sundaland

23

Myron Shekelle - Distribution and Biogeography of Tarsiers

contrasts sharply with the distributions of Philippine and Eastern tarsiers that cover all of Greater Mindanao and Sulawesi, respectively. At first glance, one might invoke the paradigm that tarsiers are a relictual taxon and are experiencing a gradual range reduction in Sundaland. An alternative hypothesis, however, is that Western tarsiers experienced a marked range expansion at the end of the last ice age. Banks (1949) hypothesized that the ice age Sunda River, which flowed northward from Sumatra and between the North and South Natuna Islands, formed an east-west faunal boundary in Sundaland. Brandon-Jones (1996) further hypothesized that large tracts of Sundaland were too dry to support habitat suitable for tropical primates in much of the Pleistocene. His hypothesis includes the prediction that many tropical primates experienced marked Holocene range expansions throughout Sundaland as Holocene climatic patterns permitted the spread of wetter, more suitable habitats (see also Meijaard 2003). The Western tarsiers appear to fit the predictions of both Banks and BrandonJones. The historical distribution of Philippine tarsiers is, essentially, a perfect fit with the ice age landmass, Greater Mindanao. They are typically regarded as a lowland species, but have been recorded above 800 m. Other than elevation, there are no known ecological constraints that would restrict their historical distribution throughout much of Greater Mindanao. It is not known when tarsiers dispersed to the Philippines, nor from where, but it is probable that they have been genetically isolated from other tarsiers since the Miocene, as mentioned previously. The subspecies of Philippine tarsiers recognized by Hill (1955) are allopatric populations on islands that separated after the end of the last ice age, implying a relatively short time frame for differentiation of Philippine tarsier populations. Thus, the current best guess for diversification within Philippine tarsiers is that a panmictic population on Greater Mindanao became fragmented as ocean levels rose at the end of the last ice age. Other possibilities should not be overlooked, however, given the several million years that Phillipine tarsiers were isolated from Western and Eastern tarsiers, including the possibility of deeper

24

biogeographic patterns on Greater Mindanao, such as are present on Sulawesi. Dagosto et al. (2003) cited evidence that the Zamboanga peninsula is geologically part of Sundaland, and accreted onto the main body of Mindanao about 5 mya, with the first evidence of emergent land in the late Miocene. These events seem to be conspicuously coincidental to the 5.6 mya molecular estimate for the origins of the Philippine tarsier clade. Additionally, they point out that changing ocean levels have subsequently rearranged the configuration of emergent land in Greater Mindanao several times, creating possibilities for vicariance events that could promote taxonomic diversity, such as is seen in Sulawesi. Unfortunately, the museum specimens of Philippine tarsiers are heavily concentrated from the Davao region of Mindanao (51 out 60 specimens examined by Dagosto et al. 2003), and field studies are hampered the Philippine tarsiers’ lack of a duet call with which to survey them and provide an initial estimate of population subdivision. The historical distribution of Eastern tarsiers includes all areas of Sulawesi that were exposed during the ice ages, as well as several landmasses that probably were not contiguous with Sulawesi during those times, including the Sangihe Islands, Banggai Islands, Togian Islands, and Selayar. There is a montane form recorded from three localities between 1800-2200 m. Lowland forms are common up 1100 m (Merker 2003), and have been recorded up to 1500 m (MacKinnon and MacKinnon 1980). It is not known when tarsiers migrated to Sulawesi, nor from where, but it is probable that they have been genetically isolated from other tarsiers since the Miocene, and almost certainly predate the coalescence of Sulawesi into a single landmass from a diverse archipelago (Hall 2001). The parapatric tarsiers on Sulawesi are hypothesized to be evidence of this ancient archipelago (Shekelle and Leksono 2004). Eastern tarsiers, therefore, are like the reverse of Philippine tarsiers; several allopatric populations were isolated on different islands, and then brought together by tectonic activity in the last 1-2 million years to form parapatric populations.

Primates of The O riental Night

Several generalizations can be drawn from the historical distributions of tarsiers hypothesized above. First, tarsiers are found in almost all lowland habitats that have not been severely degraded. Second, tarsier distributions are limited by elevation, the maximum reported elevation for Eastern tarsiers is 2200 m, for Western tarsiers is 1200 m, and for Philippine tarsiers is in excess of 800 m. Some of these differences could be sampling error, but an endemic montane species is recorded from Sulawesi, as opposed to Sundaland and the Philippines for which there is no known endemic montane tarsier. Mountain ranges limit dispersal to higher elevations, but are not known to form barriers to dispersal around the flanks. Third, other landforms, such as rivers are not effective barriers to tarsier dispersal over geologic time. In only a single instance is a river hypothesized to be a species boundary, i.e. the northern boundary of T. bancanus on Sumatra. Fourth, the chief barrier to dispersal in almost all instances is open ocean. Only on Sulawesi are parapatric tarsier populations known, and many of the contact zones are interpreted as evidence of an ancient archipelago, one that is predicted by geologic data. Finally, the presence of tarsiers on both sides of Wallace’s Line might suggest a certain aptitude for rafting across open ocean over geologic time intervals. The fact that tarsiers have not crossed numerous narrow ocean straits in the Philippines and Sulawesi, however, conflicts with this assessment, and successful dispersal by rafting is probably very rare for tarsiers. ACKNOWLEDGEMENTS This material is based on work supported by the National Science Foundation under Grant No. INT 0107277, and by a grant from the Margot Marsh Biodiversity Foundation. Portions of this appeared in the dissertation of M.S. and I thank Washington University in St. Louis and my thesis committee. Erik Meijaard, Vincent Nijman, Antonio Gorog, Colin Groves, and Matt Richardson reviewed this manuscript and offered helpful criticism.

REFERENCES Banks, E. 1949. Bornean Mammals. Kuching, Malaysia: The Kuching Press. Brandon-Jones, D. 1996. The Asian Colobinae (Mammalia: Cercopithecidae) as indicators of Quaternary climatic change. Bio. J. Linn. Soc. 59:327-350. Brandon-Jones, D, AA. Eudey, T. Geissmann, CP. Groves, DJ. Melnick, JC. Morales, M. Shekelle, & CB. Stewart. 2004. Results from the Workshop ‘Primate Taxonomy for the New Millennium’: Asian Primate Classification. Inter. Journal of Primatology 25(1):97-164. Camel, G.J. 1705. Philos. Trans. xxv. 2197-2204. Cabrera, A. 1923. On the identification of Simia syrichta Linnaeus. Journal of Mammalogy 4:89-91. Chasen, FN. 1940. A handlist of Malaysian mammals. Bulletin of the Raffles Museum 15:1-209. Clark, WEL. 1924. Notes on the living tarsier (Tarsius spectrum). Proc. Zool. Soc. London (1924): 217-223. Crompton, RH & PM. Andau. 1986. Locomotion and habitat utilization in free-ranging Tarsius bancanus: a preliminary report. Primates 27(3): 337-355. Dagosto, M & D. Gebo. 1997. A preliminary study of the Philippine tarsier in Leyte. Asian Primates 6:4-8. Dagosto, MDL, DL. Gebo & CN. Dolino. 2003. The natural history of the Philippine tarsier (Tarsius syrichta). In Tarsiers: Past, Present, and Future. Wright PC, Simons EL, Gursky S (eds) pp:237-259. New Brunswick, New Jersey: Rutgers University Press. Feiler, A. 1990. Über die Säugetiere der Sangihe- und Talaud-Inseln – der Beitrag A. B. Meyers für ihre Erforschung (Mammalia). Zoologische Abhandlungen des Staatlichen Museums für Tierkunde Dresden, 46: 75–94. Fischer, G. 1804. Anatomie der Maki und der ihnen verwandten Thiere. Frankfurt am Main. Fulton, JF. 1939. A trip to Bohol in quest of Tarsius. Yale J. Biol. Med. 11:561-573.

25

Myron Shekelle - Distribution and Biogeography of Tarsiers

Ginsburg, L & P. Mein. 1986. Tarsius thailandica nov. sp., Tarsiidae (Primates, Mammalia) fossile d’Asie. C. R. Academie of Science (Paris) t.304, ser. II, (19):1213-1215. Goodman, M, CA. Porter, J. Czelusniak, SL. Page, H. Schneider, J. Shoshani, G. Gunnell & C. Groves. 1998. Toward a phylogenetic classification of Primates based on DNA evidence complemented by fossil evidence. Mol. Phyl. Evol. 9:585-598. Gorog, AJ, & MH. Sinaga. 2008. A tarsier capture in montane forest on Borneo. In Primates of the Oriental Night edited by Shekelle M, I. Maryanto, C. Groves, H. Schulze, H. FitchSnyder. (eds) (this volume). Groves,CP. 1998. Systematics of tarsiers and lorises. Primates 39:13-27. Groves, C. 2001. Primate Taxonomy. Washington D.C.: Smithsonian Institution Press. 350 p. Hall, R. 2001. Cenozoic reconstructions of SE Asia and the SW Pacific: changing patterns of land and sea. In Faunal and Floral Migrations and Evolution in SE Asia-Australia, Metcalf I, Smith J, Morwood M, Davidson I. (eds) pp:35-56. Lisse: Swets and Zeitlinger Publishers. Heaney, LR. 1985. Zoogeographic evidence for Middle and Late Pleistocene land bridges to the Philippine islands. Mod. Quaternary Res. SE Asia 9:127-143, figs. 1-3. Hill, WCO. 1953. Notes on the Taxonomy of the Genus Tarsius. Proceedings of the Zoological Society of London 123:13-16. Hill,WCO. 1955. Primates: Comparative Anatomy and Taxonomy. II. Haplorhini: Tarsioidea. Edinburgh: Edinburgh University Press. Jentink, FA. 1892. Catalogue systématique des mammif res. Muséum d’Histoire Naturelle des Pays-Bas. Tome XI. E.J. Brill, Leide. Leksono, SM, Y. Masala & M. Shekelle. 1997. Tarsiers and agriculture: thoughts on an integrated management plan. Sulawesi Primate Newsletter 4(2):11-13.

26

MacKinnon, J, K. MacKinnon. 1980. The behavior of wild spectral tarsiers. International Journal of Primatology 1(4):361-379. Maryanto, I & M. Yani. 2004. The third record of pygmy tarsier (Tarsius pumilus) from Lore Lindu National Park, Central Sulawesi, Indonesia. Tropical Biodiver-sity. 8(2):79-85. Meijaard, E. 2003. Mammals of south-east Asian islands and their Late Pleistocene environments. Journal of Biogeography 30(8):1245-1257. Meireles, CM, J. Czelusniak, SL. Page, DE. Wildman, & M. Goodman. 2003. Phylogenetic position of tarsiers within the order Primates: evidence from γ–globin DNA sequences. In Tarsiers: Past, Present, and Future. Wright PC, Simons EL, Gursky S. (eds) pp:145-160. New Brunswick: Rutgers UP. Mercer, JM, & VL. Roth (2003). The effects of Cenozoic global change on squirrel phylogeny. Science 299:1568-1572. Merker, S. 2003. Vom Aussterben bedroht oder anpassungsfaehig? - Der Koboldmaki Tarsius dianae in den Regenwaeldern Sulawesis. PhD-Dissertation, University of Goettingen, Germany. Merker, S. & CP. Groves. 2006. Tarsius lariang: A new primate species from western central Sulawesi. Int. J. Primatol. 27: 465–485. Meyer, AB. 1897. Säugethiere vom Celebes- und Philippinen-Archipel, I. Abhandlungen und Berichte des Kaiserlich Zoologis-che und A n t h r o po l o g i s c h e - E t hn o l o gi s - c h e n Museums zu Dresden, 6: I–VIII, 1–36. Miller, Jr. GS, & N. Hollister. 1921. Twenty new mammals collected by H. C. Raven in Celebes. Proceedings of the Biological Society of Washington 34:93-104. Morley, RJ. 1998. Palynological evidence for tertiary plant dispersals in the SE Asian region in relation to plate tectonics and climate. In Biogeography and Geologi-cal Evolution of SE Asia. Hall R, Holloway JD (eds) pp:211234. Leiden: Backhuys.

Primates of The O riental Night

Musser, GG & M. Dagosto. 1987 The identity of Tarsius pumilus, a pygmy species endemic to the montane mossy forests of Central Sulawesi. American Museum Novitates 2867:1-53. Neri-Arboleda I, P. Stott & NP. Arboleda. 2002. Home ranges, spatial movements, and habitat associations of the Philippine tarsier (Tarsius syrichta) in Corella, Bohol. J. Zool. Lond. 257:387-402. Niemitz, C. 1979. Results of a field study on the Western tarsier (Tarsius bancanus borneanus Horsfeld, 1821) in Sarawak.” Sarawak Museum Journal 27 (1979a): 171228. Niemitz, C. 1984. Taxonomy and distribution of the genus Tarsius Storr, 1780. In Biology of Tarsiers. Niemitz C. (ed) pp:1-16. New York: Gustav Fischer Verlag. Niemitz, C. 1985. Der Koboldmaki - Evolutionsforschung an einem Primaten. Naturwiss Runsch 38:43-49. Niemitz, C, A. Nietsch, S. Warter, & Y. Rumpler. 1991 Tarsius dianae: A new primate species from Central Sulawesi(Indonesia). Fol ia Primatologica 56:105-116. Nietsch, A. 1999. Duet vocalizations among different populations of Sulawesi tarsiers. Int. J. Primatol. 20(4):567-583. Nietsch, A & C. Niemitz. 1993. Diversity of Sulawesi tarsiers. Deutsche Gesellschaft fur Saugetierkunde 67:45-46. Nietsch, A & ML. Kopp. 1998. Role of vocalization in species differentiation of Sulawesi Tarsiers. Folia primatologica 68(suppl.1):371-378. Nietsch, A & N. Babo. 2001. The tarsiers of South Sulawesi. In Konservasi Satwa Primata. pp:114-119. Yogyakarta: Fakultas Kedokteran Hewan dan Fakultas Kehutanan Universitas Gajah Mada University - Yogyakarta. Nietsch, A & J. Burton. 2002. Tarsier Species in Southwest and Southeast Sulawesi. Abstracts, The XIXth Congress of the International Primatological Society (IPS), 49 Aug. 2002, Beijing, China: 20-21.

Pallas, PS. 1778. Novae species quad e glirium ordine cum illustrationibus variis complurium ex hoc ordine animalium. Erlangen: W. Walther. Rickart, EA, LR. Heaney, PD. Heideman, & RCB. Utzurrum. 1993. The distribution and ecology of mammals on Leyte, Biliran and Maripipi Islands, Philippines. Fieldiana Zoology 72:162. Shekelle, M. 2003. Taxonomy and biogeography of Eastern Tarsiers. Doctoral thesis. Washington University, St. Louis. Shekelle, M, SM. Leksono, Ichwan LLS, & Y. Masala. 1997. The natural history of the tarsiers of North and Central Sulawesi. Sulawesi Primate Newsletter, 4 (92):4-11. Shekelle, M, JC. Morales & DM. Melnick. 2001. Genetic and acoustic evolution among Eastern Tarsiers of northern and central Sulawesi. Presented at the International Society of Primatologists, 14th Congress, Adelaide, Australia. January 7-12, 2001. Shekelle, M, & SM. Leksono (2004) “Rencana Konservasi i Pulau Sulawesi: Dengan Menggunakan Tarsius Sebagai ‘Flagship Taxon’”. Biota 9 (1):1-10. Shekelle, M, C. Groves, S. Merker & J. Supriatna. In Press. Tarsius tumpara: A New Tarsier Species from Siau Island, North Sulawesi. Primate Conservation 2008 (23). Simons, EL. 2003. The fossil record of tarsier evolution. In Tarsiers: Past, Present, and Future. Wright PC, Simons EL, Gursky S (eds) pp:934. New Brunswick, New Jersey: Rutgers University Press. Sody, HJV. 1949. Notes on some Primates, Carnivora, and the babirusa from the Indo-Malayan and indo-Australian regions. Treubia 20:121-185. Wharton, CH. 1950. The tarsier in captivity. Journal of Mammalogy 31(3):260-268. Whitten, A, M. Mustafa, & G. Henderson. 2002. The Ecology of Sulawesi. 2 nd ed. Singapore: Periplus. Yanuar, A & J. Sugardjito. 1993. Population survey of primates in Way Kambas National Park, Sumatra, Indonesia. Tiger Paper. 30-36

27

Primates of The O riental Night

A TARSIER CAPTURE IN UPPER MONTANE FOREST ON BORNEO

1

Antonia J. Gorog1 & Martua H. Sinaga2 Museum of Zoology and Department of Ecology and Evolutionary Biology, 1109 Geddes Ave. University of Michigan 2 Indonesian Institute of Sciences (LIPI) & Museum Bogoriense, Cibinong, West Java 16911, Indonesia ABSTRACT

In November of 1998, we captured a tarsier above 1200 m elevation in West Kalimantan, Indonesia. This capture is the highest recorded elevation for a Western tarsier and is furthermore unusual in that the Bornean tarsier (Tarsius bancanus) is generally described as a lowland species. In this paper, we briefly summarize the geographic distributions and habitat associations of the seven recognized species of tarsiers, report on our high elevation capture, and discuss the implications of our finding. Keywords: Tarsius bancanus, T. syrichta, T. tarsier, T. spectrum, biogeography, habitat, elevational distribution.

INTRODUCTION The recognized species of Tarsius are restricted to Southeast Asia. T. tarsier group (or Eastern tarsiers), comprising T. tarsier (=T. spectrum, see Groves et al. this volume), T. pumilus, T. dentatus, T. pelengensis, T. sangirensis, T. lariang, and T.sp is endemic to Sulawesi and nearby small islands (Groves 2001; Shekelle this volume). Tarsius tarsier shows great geographic variation in cranial morphology (Groves 1998) and vocalization (MacKinnon & MacKinnon 1980; Shekelle et al. 1997), and is likely a complex of several species. Brandon-Jones et al. (2004) identified T. tarsier as a senior subjective synonym of T. spectrum, and accepted Makassar (Ujung Pandang) in the southwestern peninsula as the type locality. It can be inferred from Groves (1998) that the tarsier in the northern peninsula of Sulawesi, which has been the focus of most behavioral and ecological research on Eastern tarsiers (e.g., Gursky 1998; MacKinnon & MacKinnon 1980), is an unnamed species. Tarsius bancanus, the Western tarsier, is found on Borneo, Sumatra and some of the interlying islands (e.g, Banka, Belitung and Serasan). Tarsius syrichta is restricted to the Philippine islands of Mindanao, Bohol, Samar, and Leyte (Groves 2001). The Philippine and Western tarsiers, T. syrichta and T. bancanus, are generally described as lowland species (e.g., Musser & Dagosto 1997; Sussman 1999) but range into lower montane forest as well. Shekelle (personal communication) noted that

two Philippine specimens in the Field Museum of Natural History (FMNH56159 and FMNH67744) from the island of Mindanao (Davao City, Mt. McKinley, east slope, and Zamboanga, Sigayan, Katipunan, respectively) include provenence data stating that they were collected at 2500 ft (758 m). Tarsius bancanus is found in a variety of forest types including secondary growth (Le Gros Clark 1924; Niemitz 1979). Payne and Francis (1998) describe the range of this species as extending “above 900 m in the Kelabit Uplands of northern Sarawak”, in addition to lowlands of other regions of Borneo. Their upland reference probably corresponds to a single specimen in the Sarawak Museum that is identified as T. bancanus and whose associated information (Kool & Nawi 1995) includes only a collection date and a locality (12 August 1949, Bario Kelabit). The elevation at Bario is approximately 1100 m; thus, this record may represent the highest and the only previous montane record for this species. Tarsius syrichta is found in many habitats including early-mid successional forest, late successional secondary forest, agroforestry systems, primary lowland forest, and montane forest from 50 to 800 m (Dagosto & Gebo 1997; Neri-Arbodela et al. 2002; Rickart et al. 1993). Eastern, or Sulawesian, tarsiers fall into two broad ecological groups—one that occupies a diversity of habitats from lowland forests and agroecosystems to lower montane forest, and another restricted to mountain tops. The members of the T. tarsier complex have been reported from all major

29

Gorog & Sinaga - A Tarsier Capture in Upper Montane Forest on Borneo

forest formations and some types of cultivated vegetation from sea level to 1500 m (MacKinnon & MacKinnon 1980) and are represented by museum specimens from this range of elevations (Musser & Dagosto 1987). Tarsius dentatus has been captured at sea level (Shekelle et al. 1997) and from 650 to 1100 m, and detected audibly from 500 to 1200 m (S. Merker, pers com), indicating an elevational distribution similar to that of the T. tarsier complex (although studies focusing on montane areas may show that it occurs higher as well). The pygmy tarsier, T. pumilus, alone is restricted to high elevation mossy forest; the species is known from only three specimens, one from 1800 m and two from 2200 m, taken in the central region of Sulawesi (Musser & Dagosto 1987; Maryanto & Yani 2004). No overlap has been reported in the altitudinal distribution of T. pumilus and other Eastern tarsiers. Findings During a survey of small mammal diversity in 1998, we captured a tarsier above 1200 m elevation in Bukit Baka-Bukit Raya National Park (BBNP) in West Kalimantan, Indonesia. To our knowledge, this represents the highest capture of a tarsier on the island of Borneo reported to date. BBNP lies in the Schwaner Range, part of the axial chain of mountains that bisects Borneo. The park spans the border between West and Central Kalimantan, lying approximately between 112°15’113°E and 0°1’-0°29’S (Jarvie et al. 1998). Bukit Baka (1600 m) and Bukit Raya (2278 m) are the two tallest mountains in Kalimantan, the Indonesian part of Borneo, and occur within the boundaries of BBNP. Our survey of small terrestrial mammals and bats was conducted on the northern slopes of Bukit Baka in an area known locally as Gunung Lubang Tedung. We accessed the region by foot from the village of Nanga Juoi, which lies within the park and to the north of Lubang Tedung. Terrestrial mammals were trapped and bats were mist-netted in each major forest formation from lowland forest at 350 m to mossy forest at 1550 m, the peak of Lubang Tedung. The KKP logging concession borders BBNP to the north and west of Bukit Baka. Thus, lowland

30

forest surrounding the park is heavily disturbed or has been converted to agriculture. Lowland forest at the foot of Gunung Lubang Tedung is utilized by local Dayak people, who occasionally harvest fruits, firewood, pig, and deer in the area. However, the lower slopes of Bukit Baka in the region of Lubang Tedung have not been deforested, and support a large zone (from approximately 350 to 700 m) of healthy, intact lowland and hill dipterocarp forest dominated by canopy trees in the genus Shorea. At about 700 m, the forest undergoes a transition from lowland dipterocarp forest to montane forest dominated by Fagaceae and patchy stands of Agathis. The mossy forests of the highest elevations (occurring above ~1500 m on Lubang Tedung) are dominated by Ericaceae and support high elevation specialists such as Nepenthes and podocarps in the genera Dacrydium and Phyllocladus (Jarvie et al. 1998). On 23 November 1998, during the portion of our survey conducted in upper montane forest (Figure. 1), we captured a tarsier in a mist net (Figure 2). The elevation of the capture site was between 1200 and 1250 m, a short walk (300 m) downslope of our camp at 1300 m. The net was set on a mossy ridge in low canopy forest 15-20 m in height. The bottom of the net hung approximately 1.5 m from the ground, and the animal was removed from the net at a height of 2.5-3.0 m. We photographed it at our camp and released it at the capture site. Other small mammals captured in the upper montane habitat of Lubang Tedung include the murid rats Leopoldamys sabanus, Maxomys whiteheadi, and M. surifer, the shrew Crocidura monticola, and the montane squirrel Dremomys everetti and bat Aethelops alecto. As on several other mountains in West Kalimantan, fewer montane endemics are present than on mountains of northern Borneo. In addition, the ecological ranges of several species that in the north are generally considered to be lowland species extend well into the uplands in southwestern Borneo, demonstrating a regional lack of altitudinal zonation in most species (Gorog 2003).

Primates of The O riental Night

Figure 1. Montane forest in BBNP near the capture site of a montane tarsier.

Figure 2: Western Tarsiers. (clockwise from upper left) montane tarsier from BBNP, montane tarsier from BBNP, T. bancanus borneanus from Sebangau (Central Kalimantan), captive T. b. bancanus from Taman Safari, captive T. b. bancanus from Taman Safari, montane tarsier from BBNP. Although the photos of the montane tarsier are slightly blurry, the tail tuft from the montane tarsier in this report appears to have fur that is shorter, sparser, and extends further along the tail than is seen in other Western Tarsiers. Appearance of the tail tuft is a key identifying feature in tarsier taxonomy, and is functionally related to leaping performance among other things. (Note: the color of the fur of captive tarsiers fades to shades that are not seen in wild animals).

Antonia Gorog © 2004, except where noted

31

Gorog & Sinaga - A Tarsier Capture in Upper Montane Forest on Borneo

Implications Bornean mammals can be grouped into those that vary regionally (e.g., the subspecies of the treeshrews Tupaia minor, T. picta, and T. glis and of the tri-colored squirrel Callosciurus prevostii and white-fronted langur Presytis frontata) or altitudinally (e.g., Maxomys spp. and Sundamys spp.), and those that show little variation across the island (e.g., the macaques Macaca fascicularis and M. nemestrina, the rusa Cervus unicolor, and the sunbear Helarctos malayanus) (Medway 1977). Current taxonomy and distributional data suggest that Bornean T. bancanus falls into this latter category; as of yet, only one morphological subspecies of the Western tarsier (T. bancanus borneanus) is recognized on Borneo (Groves 2001). In their 1987 paper on the identity of T. pumilus, Musser & Dagosto examined distinguishing characteristics of the tarsier species and provided the last comprehensive review of tarsier morphology. Despite their broad sampling of Bornean specimens, which included specimens from Sabah, Sarawak, and West Kalimantan, the authors’ goal was not to evaluate variation in Bornean T. bancanus, and their study did not address the possibility of multiple differentiated forms on this island. This new record at Bukit Baka suggests a greater morphological diversity in tarsiers on Borneo than is currently recognized. Sharp differences in tail morphology of the Eastern, Western, and Philippine tarsier species provide useful diagnostic characters for identifying species (summarized in Musser & Dagosto 1987; Shekelle 2003). The tail of the Philippine species is covered by relatively sparse, short hairs. The tails of the Sulawesian species are the most heavily haired, with the terminal tuft covering one third to almost one half of the tail. The Western tarsier possesses a more distinct tuft of long hairs confined to the distal third of the tail. The tail of the tarsier we captured at Bukit Baka was scruffier and more hirsute than that of a typical Western tarsier, but less hirsute than the tails of Eastern, or Sulawesian, tarsiers. In combination with its brownish coloration, which is generally greyer in the Bornean T. bancanus, the Bukit Baka individual does not look like a typical Western tarsier (C. Groves & M. Shekelle, pers com).

32

The high elevation tarsier we captured may represent a differentiated morph of T. bancanus and an altitudinal extension of the habitats thought to be occupied by this species. Alternatively, the individual captured at Bukit Baka may represent a new tarsier form. This is an interesting possibility given the existence of a distinct montane species, T. pumilus, on Sulawesi, but remains to be tested with specimens or genetic materials, neither of which could be collected from this animal. High elevation regions of Kalimantan represent an increasing proportion of remaining natural habitat on Borneo, as lowland forest is rapidly reduced by large-scale logging operations (Curran et al. 1999). Many mammal species endemic to montane habitat have been recorded in northern (Malaysian) Borneo (Medway 1977), and new species have been described recently (e.g., Emmons 1993) from the relatively well-studied north. In contrast, few high elevation surveys have been conducted in Kalimantan, and the fauna of Indonesian Borneo is almost certainly richer than is currently thought. We hope with this report to inspire additional survey work and research in montane areas of Kalimantan. ACKNOWLEDGEMENTS We thank LIPI (the Indonesian Institute of Sciences), the Museum Zoologicum Bogoriense, and PHKA (the Indonesian Department of Forest Protection and Nature Conservation) for supporting our field work in West Kalimantan. Special thanks to Gary Paoli, Agustinus Suyanto, Maharadatunkamsi, Rahmadi, and Medang. REFERENCES Curran, LM, I. Caniago, GD. Paoli, D. Astiani, D. Kusneti, M, Leighton, CE. Nirarita & H. Hacruman. 1999. Impact of El Niño and logging on canopy tree recruitment in Borneo. Science 286: 2184-2188. Dagosto M. & D. Gebo. 1997. A preliminary study of the Philippine tarsier in Leyte. Asian Primates 6:4-8.

Primates of The O riental Night

Emmons, LH. 1993. A new genus and species of rat from Borneo (Rodentia: Muridae). Proc. Biol. Soc. Wash 106: 752-761. Gorog, AJ. 2003. Historical Biogeography of Small Mammals in the Sunda Region of Southeast Asia. Dissertation, University of Michigan, Ann Arbor, Michigan. Groves, C. 1998. Systematics of tarsiers and lorises. Primates 39: 13-27. Groves, C. 2001. Tarsiiformes. In: Primate Taxonomy. Eileen D’Araujo (ed.). Smithsonian Institution Press, Washington, D.C., pp. 121-125. Groves, C. This volume. Getting to Know the Tarsiers: Yesterday, Today and Tomorrow. Gursky, S. 1998. Conservation status of the spectral tarsier Tarsium spectrum: population density and home range size. Fol. Prim. Supplement 1: 191-203. Jarvie, JK., Ermayanti, U. Mahyar, U. Church & Ismail. 1998. The habitats and flora of Bukit BakaBukit Raya National Park. Tropical Biodiversity 5: 11-56. Kool, KM, & Y. Nawi. 1995. Catalogue of Mammal Skins in the Sarawak Museum. Universiti Malaysia Sarawak, Kuching, Sarawak, Malaysia. Le Gros Clark, WE. 1924. Notes on the living tarsier (Tarsius spectrum). Proceedings of the Zoological Society of London 1924: 216-217. MacKinnon, J, & K. MacKinnon. 1980. The Behavior of wild spectral tarsiers. International Journal of Primatology 1: 361-379. Maryanto, I & M. Yani. 2004. The third record of pygmy tarsier (Tarsius pumilus) from Lore

Lindu National Park, Central Sulawesi, Indonesia. Tropical Biodiversity 8(2): 79-85. Medway, G.H. 1977. Mammals of Borneo.: Printed for MBRAS by Perchetakan Mas Sdn. Bhd., Kuala Lumpur, Malaysia. Musser, G. & M. Dagosto. 1987. The identity of Tarsius pumilus, a pygmy species endemic to the montane mossy forests of central Sulawesi. Amer. Mus. Novit. 2867: 1-53. Neri-Arbodela, I., P. Stott, & NP. Arbodela. 2002. Home ranges, spatial movements and habitat association of the Philippine tarsier (Tarsius syrichta). J. Zool, London 257: 387-402. Niemitz, C. 1979. Results of a field study on the western tarsier (T. bancanus bancanus Horsfield 1821) in Sarawak. Sarawak Mus. J. 27: 171-228. Payne J. & CM. Francis. 1998. A Field Guide to the Mammals of Borneo. The Sabah Society, Kota Kinabalu, Malaysia. Rickart, EA, LR. Heaney,PD. Heideman & RCB. Utzurrum. 1993. The distriubtion and ecology of mammals on Leyte, Biliran and Maripipi Islands, Philippines. Field. Zool. 72: 1-62. Shekelle, M. 2003. Taxonomy and Biogeography of Eastern Tarsiers. Ph.D. thesis. Washington University in Saint Louis Shekelle, M., Leksono, SM, LLS. Ichwan & Y. Masala. 1997. The natural history of tarsiers of North and Central Sulawesi. Sulawesi Primate Newsletter 4: 4-11. Sussman, R. 1999. Tarsiiformes. In: Primate Ecology and Social Structure. Pearson Custom Publishers, Needham Heights, MA.

33

Primates of The O riental Night

DISTRIBUTION OF TARSIER ACOUSTIC FORMS, NORTH AND CENTRAL SULAWESI: WITH NOTES ON THE PRIMARY TAXONOMY OF SULAWESI’S TARSIERS Myron Shekelle Center for Biodiversity and Conservation Studies (CBCS-UI), Faculty of Mathematics and Science (F-MIPA), University of Indonesia, Depok 16424, Republic of Indonesia, Email: [email protected] ABSTRACT This study uses field surveys of wild tarsier populations to assess the relationship between described tarsier taxa and acoustic forms. Surveyed regions of North and Central Sulawesi contain eight acoustic forms and five described taxa. There is synonymy between acoustic form and taxonomic form in two instances, minimally, and two of the taxonomic forms appear to be synonymous with each other. Thus, if tarsier acoustic forms are distinct taxa, as has been hypothesized (MacKinnon and MacKinnon 1980, Niemitz et al. 1991, Nietsch and Kopp 1998), then as many as ten tarsier taxa may be present in the region that has been surveyed thus far. Acoustic forms were identified with playback tests using populations of wild tarsiers and corroborated with heuristic spectrographic analyses. Spectrograms of three previously undescribed acoustic forms of tarsiers are presented, along with spectrograms from four acoustic forms that were already known. The distributions of tarsier taxa and acoustic forms are presented. Comments are made regarding the nature and validity of each taxon. Tarsius sangirensis is recognized as a distinct species. Tarsius dianae is likely a junior synonym of T. dentatus. The whereabouts and location of the type locality of T. pumilus are discussed. Keywords: Distribution, Tarsius spp, vocalization

INTRODUCTION The primary taxonomy of Sulawesi’s tarsier has been in question for the past twenty years. Results of collecting expeditions from the eighteenth century up to 1949 led to the description of five distinct tarsier taxa from Sulawesi and its surrounding islands. By 1984, however, Niemitz had simplified the taxonomy, classifying Sulawesi’s tarsiers into a single species, Tarsius spectrum, with two subspecies, T. spectrum, and T. pumilus (Table 1). Tarsius spectrum is now accepted as a junior synonym of T. tarsier (see Groves et al. this volume). Just prior to this, MacKinnon and MacKinnon (1980) reported the first observations of wild Tarsius tarsier, wherein they commented that there were clearly several distinct taxa of tarsiers on Sulawesi. At the same time, the classification and taxonomy of nocturnal primates was undergoing substantial revision. Prompted in part by the Recognition Concept of Species (Paterson, 1985), numerous cryptic sibling species were ultimately identified where once taxonomists had recognized only a single species. This led to the argument that the

number of species in nocturnal taxa, in particular, had been underestimated (e.g. Bearder et al, 1995; Masters, 1998). In addition to this taxonomic revision of Sulawesi’s other primate, the macaque, showed as many as seven species where once there was thought to be just two (e.g. Fooden, 1980; Groves, 1980). This created awareness of Sulawesi’s interesting biogeographic phenomenon of colonization and radiation (Whitten et al. 1987), thereby lending further credence to the hypothesis that several cryptic sibling species exist among Sulawesi’s tarsiers. Prior to 1984, five distinct tarsier taxa had been described from the region of Sulawesi and its surrounding islands (Figure 1). These are: Tarsius tarsier Erxleben, 1777, type locality Makasar (see discussion); T. sangirensis Meyer, 1897, from Greater Sangihe Island, North Sulawesi; T. pumilus Miller and Hollister, 1921, from Rano Rano, Central Sulawesi; T. dentatus Miller and Hollister, 1921, from Labuan Sore, Central Sulawesi; and, T. pelengensis Sody, 1949, from Peleng Island. The taxonomic history of tarsiers, and specifically Sulawesian tarsiers, is confusing and riddled with nomenclatural arguments and instability.

35

Shekelle- Distribution of Tarsier Acoustic Forms

Table 1. Taxonomic Revisions of Sulawesi’s Tarsiers 1949-1984

Taxon T. tarsier (=spectrum) T. sangirensis T. pumilus T. dentatus T. pelengensis

Sody 1949 species subspecies (not mentioned) subspecies subspecies

Hill 1955 species subspecies subspecies subspecies subspecies

Eastern Tarsier Taxa (prior to this study)

Niemitz et al. 1991

Niemitz 1984a species jr. synonym subspecies jr. synonym jr. synonym MacKinnon & MacKinnon 1980

Sampling Localities

Figure 1: Location of taxa, acoustic forms, and sampling localities. MacKinnon and MacKinnon (upper right) recognized three acoustic forms. Niemitz et al. (lower left) named T. dianae based upon comparisons with tarsiers from the northern tip of Sulawesi. Neither of these works addressed in detail the other described taxa of Eastern tarsiers (upper right). This study sampled a transect at approximately 100 km intervals.

36

Primates of The O riental Night

spots= sampling localities x= capture sites of T. pumillus

Figure 2. Distributions of Eastern tarsier taxa and acoustic forms.

functions, and which co-inhabit biogeographic faunal zones with Sulawesi’s macaques. Included with this hypothesis was the possibility of an altitudinal variant, the small-bodied montane tarsier of central Sulawesi, T. pumilus. They concluded by stating “there is clearly much more taxonomic work to be done to sort out the Sulawesi tarsiers, but we would predict that there are more forms to be found in southern Sulawesi, and on the offshore island groups of Selayar, Peleng, and Sangihe-Talaud.” Following the suggestion of MacKinnon and MacKinnon, several research teams began investigations into the primary taxonomy of

Sulawesian tarsiers. On one front, field primatologists began to collect data from wild Sulawesian tarsiers. Niemitz (1984b) surveyed tarsiers in Central Sulawesi and published preliminary spectrograms of tarsiers from Marena. Niemitz et al. (1991) conducted more surveys in North and Central Sulawesi that led to the description of a new species from the highlands of Central Sulawesi, Tarsius dianae (Niemitz et al. 1991). They made similar observations to those of MacKinnon and MacKinnon, stating “there may be a constellation of nocturnal tarsier species paralleling the set of closely related diurnal macaque species in Sulawesi” (Figure 1).

37

Shekelle- Distribution of Tarsier Acoustic Forms

Nietsch (1993, 1994) and Nietsch and Kopp (1998) continued investigations into the primary taxonomy of Sulawesian tarsiers, concentrating on the role of tarsier vocalizations. Nietsch documented the T. dianae acoustic form at Kebun Kopi (north of Palu along the Tawaeli-Toboli Road) and at Lake Poso, in addition to the type locality at Kamarora. She also reported hearing duets that sounded like the T. dianae acoustic form at Ampana and Morowali, but these were unconfirmed by playback or spectral analysis (Nietsch, 1994). Nietsch and Kopp (1998) also provided experimental evidence that “differences in duet structure are reliable indicators of taxonomic differentiation”, in that duets of heterospecifics were less effective in prompting responses from caged animals than were the calls of conspecifics. Based on analogies with song function in gibbons, she concluded that tarsiers from Tangkoko, Kamarora, and the Togian Islands were each likely to be a separate species. While fieldwork in Sulawesi was progressing, analyses of museum specimens were also being undertaken. Musser and Dagosto (1987) recommended the re-elevation of T. pumilus to full specific status as a small-bodied tarsier endemic to the montane forests of the central highlands. Feiler (1990) argued for the re-elevation of tarsiers from the Sangihe Islands to full specific status, T. sangirensis, based on cranial measurements. Groves (1998) concluded that the distinctive characters of Feiler were insignificant when compared with a larger sample, but he also argued for reviving T. sangirensis based upon morphometric variation in dental and cranial characteristics. Groves further mentioned the possible justification for separating the Sulawesian tarsiers from T. syrichta and T. bancanus at the generic level. In spite of the advances made by field surveys and analyses of museum specimens, some nagging issues remained, and new ones had been created. First, populations of, T. pumilus and T. pelengensis were not included in the recent field surveys. Second, the surveys of MacKinnon and MacKinnon (1980) and Niemitz et al. (1991) did not make thorough comparisons between acoustic forms and pre-existing taxa, nor did either study employ

38

systematic surveys. While it is understandable that preliminary work be ad hoc, two issues of confusion were generated by the advances that these two studies made: (1) what were the relationships between acoustic forms and the five described taxa from Sulawesi? and, (2) what variation existed between the widely separated sample points of the field surveys? This project began with reconnaissance surveys in 1994, and collecting expeditions followed in 1995 and 1996. The goal was to use a transect to “connect-the-dots”, as it were. A transect that encircles Tomini Bay, with a few minor auxiliary transects, would pass through all of the type localities of all described tarsier taxa from Sulawesi, with the notable exception of T. tarsier, as well as all of the previously identified acoustic forms (Figure 2). Based on evidence available at the time, 200 km sampling intervals were initially indicated. The results of initial surveys, however, indicated greater diversity than predicted by the previous studies, and sampling intervals were reduced to 100 km. Even this proved insufficient in the region from Gorontalo to Palu. Preliminary findings included: (1) identification of seven acoustic forms in the areas of North and Central Sulawesi that were partially surveyed by MacKinnon and MacKinnon (1980) and by Niemitz et al. (1991); (2) additional evidence for the re-elevation T. sangirensis to full-specific status; and, (3) the possibility of a nomenclatural conflict between T. dianae and T. s. dentatus (Shekelle et al. 1997). METHODS Data were collected on wild tarsiers from 12 localities in North and Central Sulawesi (Figure 1). Nine of the localities lie at approximately 100 km intervals along a transect that connects Tangkoko (the reference population for T. tarsier), and Kamarora, the type locality of T. dianae. The other three localities lie along two auxiliary transects that connect island populations with the main transect: (1) Great Sangihe Island, about 200 km north of Sulawesi; and (2) two locations in the Togian Islands of Tomini Bay, Batudaka Island and Malenge Island.

Primates of The O riental Night

Recordings of wild tarsier vocalizations were made using either a Sony WMD 6C with a Sennheiser MKE 300 microphone, or a Sony TR-600 Hi8 camcorder with either the internal microphone or with a Sennheiser MKE 300. Many recordings were made of naturally occurring vocal duets at dusk or dawn. Other recordings were made of tarsiers that were baited to sing by playing the recording of a conspecific. The current study does not discriminate natural from baited vocalizations, although this difference may be important for understanding variation within an acoustic form. For spectral analysis, recordings were replayed on the same machine that made the original field recording, and these were converted to digital “*.wav” files using an analog to digital converter. This was facilitated by the computer program “Cool 96” (copyright 1992-1996, Syntrillium Software). The following options were employed: sample rate (44100 Hz); resolution (1024 bands); windowing function option (Blackmann); spectral plot style (logarithmic energy plot with 120 dB range); dither transform results (on). Spectrogram files were pasted into Photoshop 5.0 (copyright 1989-1998 Adobe Systems Inc.) and were transformed to standardize the x- and y-axes, as well as to improve legibility. For playback experiments, a test tape was made in the field of unaltered wild tarsier recordings, one call per locality, each call being separated by a gap of silence. The duration of each duet call and the duration of each period of silence were standardized. Positioning of the speakers (Sony SRS 77G) and volume were standardized as well as could be done (speakers were positioned at a distance of 10-20 meters from the sleeping site, volume was set to full). An unhabituated group of tarsiers was exposed to the tape between 8 a.m. and 3 p.m. A positive response was recorded if the tarsiers in question began vocalizations that led to a duet call while a recorded duet was being played, or during the period of silence that followed. The time of day (8 a.m. to 3 p.m.) was chosen because tarsiers are normally sleeping and are not normally vocalizing. Thus any positive response should be the result of the experiment, not some other

factor. Also, a recording of the local duet call was played approximately one hour before and after the playback test. The playback test is, therefore, bracketed by positive responses, and this helps to ensure that any negative responses were not caused by other factors (e.g. the tarsiers did not hear the recording, the tarsiers had moved off to another site, the tarsiers were frightened into silence by the presence of researchers, etc.). If, during the course of the playback test, a positive result was recorded (i.e. the tarsiers began to sing), the test was paused for a minimum of 1/2 hour in order to let the tarsiers settle down. Thus, subsequent positive responses are likely to be the result of subsequent recordings, not because of continued excitability as a result of the previous positive response. Playback tests are time consuming. Typically, only one formal test was performed per locality. Results of the formal tests were often corroborated, however, through numerous informal tests. Informal tests occurred during survey and trapping, when trial-and-error of playback was used to locate tarsier nests and to bait the tarsiers into the nets. Since the playback tape was made in the field from recordings collected during survey, not all reciprocal tests could be conducted. Tarsiers from localities that were visited early in the study, such as Molibagu, were not exposed to recordings of tarsiers from localities that were visited subsequent to fieldwork at Molibagu. In other words, if tarsier recordings from locality x were played for tarsiers at locality y, it is not necessarily true that recordings from locality y were played for tarsiers at locality x. When conducted, however, the results of reciprocal tests were typically equal, i.e., the graph is symmetrical along the axis x = y. RESULTS Playback experiments reveal the presence of seven acoustic forms in the study area with geographically-structured variation (Figure 3). From north to south along the transect, the first acoustic form (site 1), the Sangihe form, is unique

39

Spectrograms

Shekelle- Distribution of Tarsier Acoustic Forms

40

Primates of The O riental Night

Figure 3. Results of Playback Tests Spectrograms for Seven Acoustic Forms of Sulawesian Tarsiers

and found only on Greater Sangihe Island. The second acoustic form (sites 2-5), the Manado form, is found at Tangkoko, Ratatotok/Basaan, Molibagu, and Suwawa. The third acoustic form (site 6), the Libuo form, is unique to Libuo. The fourth acoustic form (site 7), the Sejoli form, is unique. The fifth acoustic group (site 8), the Tinombo form, is unique. The sixth acoustic form (sites 9-10), the T. dianae form, is found at Marantale and Kamarora. The seventh acoustic group (sites 11-12), the Togian form, is found at

Malenge Island and Batudaka Island, in the Togian Islands. The seventh acoustic group is interesting. Results of reciprocal tests were unequal. Recordings of the duet call from all sites successfully baited these tarsiers to respond, i.e. a positive response. When recordings of the Togian Island form were played at other sites, however, the recordings did not bait the tarsiers to sing, i.e. a negative response. These results were tested and re-tested, always to the same effect.

41

Shekelle- Distribution of Tarsier Acoustic Forms

Togian Island tarsiers seem not to discriminate among acoustic forms as do other tarsiers in the study group. The playback test from site 5, Suwawa, had an anomaly. During the course of the formal playback test, the target group responded to all calls. During informal tests, prior to the formal test, however, tarsiers from this locality did not respond to the duet calls of other acoustic forms. The formal test was conducted on the last day of fieldwork under sub-optimal conditions. Various conditions of the test were not met (e.g. the minimum pause of ½ hour between positive results was not observed). Time constraints did not allow for the formal test to be repeated. This study tentatively rejects the results of the formal test for those of the informal tests. Further testing is warranted. Recordings of wild tarsiers in this study were made under field conditions and are generally of poor quality. They are not well suited for spectral analysis. The following are very general descriptions of the vocal duets. These descriptions are presented with the principal goal of facilitating fuller understanding of what is illustrated in the spectrograms. They are not meant to be definitive statements of the characteristics of the given acoustic forms. The descriptions use the following terms: unit = one part of a multi-part call; call = one coordinated, repeated vocalization. The difference between a unit and a call is somewhat arbitrary, but it is nevertheless a useful distinction for describing tarsier vocal duets. The Sangihe form is a previously undescribed acoustic form. It is characterized by a two-unit female call and a rapid series of male calls (spectrogram 1a). The female contributions include the standard call, and at least one variant that occurred at the end of a duet (spectrogram 1b), both of which have two units. The first unit of the standard female call is a long whistle, over one second long with energy concentrated around 8 kHz (spectrogram 1b, far left). The second unit rapidly descends from over 14 kHz to below 6 kHz. The variant is shorter, and does not ascend above 10 kHz (spectrogram 1b, far right). Spacing between the female calls varies, but is typically 8-10 seconds from the start of one call to the start of the next (not pictured). The male calls are

42

wide-band chirps that rapidly rise and descend (7 kHz, to over 10 kHz, to less than 7 kHz) in a span of only 0.15 seconds. The male produces these calls at the rate of about 4-6 calls per second. The Manado form was originally described by MacKinnon and MacKinnon (1980), and was further examined by Niemitz et al. (1991), each of whom used recordings made at Tangkoko (spectrograms 2a and 2b) (Table 2). The duet is a series of synchronized male and female calls that precede a crescendo of ascending whistles by the female. Initially, the female call is a descending whistle that drops from around 10 kHz to below 6 kHz over the span of approximately 0.3 seconds. The female makes these calls about once per second. The female calls decrease the degree to which they descend in pitch, gradually flattening out, and finally, they begin to ascend. During this portion, the call can become highly synchronized (e.g., spectrogram 2d, seconds 6-8). In between, and on top of the female calls, are male calls. These are wideband chirps that rise and descend rapidly (6 kHz, to 13 kHz, to less than 6 kHz), all in a span of only 0.2 seconds. The male produces these at a rate of about 2 per second. The overall length of one such vocal phrase varies greatly but they do not usually repeat any faster than about once per 10 seconds. MacKinnon and MacKinnon (1980) originally described the Libuo form as the Gorontalo form (spectrogram 3a). This acoustic form is characterized by a two- or three-unit female call accompanied by male calls (spectrogram 3). The female call has a standard form (spectrogram 3b) and at least one variant (spectrogram 3c). The standard female call is a long whistle that descends from around 13 kHz to about 7 kHz in two or three units. A variant, in spectrogram 3c is similar to the standard call, but with 5-8 units that may take up to 7 seconds total, each unit being about 0.6-1.0 seconds and separated by a momentary pause. The female calls repeat with a minimum periodicity of around 6 seconds from the start of one call to the start of the next. The male calls, wide-band chirps that rise and descend (6 kHz, to 1012 kHz, and back to 6 kHz) in about 0.15 seconds, punctuate the spaces in the female call (e.g.

T. sangirensis Meyer 1897 T. dentatus Miller and Hollister 1921

T. pumilus Miller and Hollister, 1921 T. pelengensis Sody, 1949

T. lariang Merker & Groves 2006

T. tarsier population T. tarsier population

T. tarsier population

T. tarsier population

T. tarsier population

2.

4.

6.

7.

9.

10.

11.

8.

5.

3.

T. tarsier Erxleben 1777

1.

Taxon name

Togian form

Tinombo form

Sejoli form

Gorontalo form = Libuo form

Manado form

Palu form

Peleng form

?

T. dianae form

Sangihe form

Bantimurung form

Acoustic form

this paper; Nietsch, 1994; 1998; Shekelle et al, 1997

this paper; Shekelle et al, 1997

this paper; MacKinnon and MacKinnon, 1980 this paper; MacKinnon and MacKinnon, 1980; Shekelle et al 1997 this paper; Shekelle et al, 1997

MacKinnon and MacKinnon, 1980; Niemitz, 1984; Merker and Groves, 2006

Niemitz, 1985; Musser and Dagosto, 1987

Feiler, 1990; Shekelle et al, 1997; Groves, 1998 Niemitz et al, 1991; Shekelle et al, 1997

Unpublished data

Reference

Table 2: Provisional Assessment of Primary Taxonomy of Sulawesian Tarsiers

Malenge Is., Batudaka Is.

Tinombo

Sejoli

Tangkoko, Ratatotok, Molibagu, Suwawa Libuo

Gimpu, Palu Valley, including Marena

Labuan Sore, Kamarora, Marantale, Kebun Kopi, Lake Poso, Ampana(?), Morowali(?) Rano Rano, Latimojong Mts., Mt. Rorekatimbu Peleng Is.

Gr. Sangihe Is.

Makassar

Distribution (type locality = bold face)

Living specimens remain unobserved to science. Notable similarities between the duet form of this species and that of T. dentatus may indicate a close relationship, possibly at the subspecific level. Niemitz's (1984) T. pumilus from Marena is not T. pumilus, but more likely, the Palu form of MacKinnon and MacKinnon (1980) Most field studies of T. spectrum refer to this population Distribution reported here much smaller than that reported by MacKinnon and MacKinnon (1980) replaces MacKinnon and MacKinnon's (1980) Gorontalo form in the area of Sejoli replaces MacKinnon and MacKinnon's (1980) Gorontalo form in the area of Tinombo has been argued to be a distinct taxon based on experimental playback evidence (Nietsch, 1998), strongly supported by genetic data (Shekelle et al, this volume)

Ongoing investigations of the recently rediscovered type specimen may allow for a more accurate localization of the type locality The insular population will likely be shown to be taxonomically distinct. apparent nomenclatural conflict with T. dianae

Issue

Primates of The Oriental Night

43

Shekelle- Distribution of Tarsier Acoustic Forms

spectrogram 3b, seconds 0-2), and then continue at the rate of about one per second between female calls. The Sejoli form is a previously undescribed acoustic form. The recordings from this locality are of particularly low quality, but even so, some resemblances to the Libuo form are apparent. This acoustic form has a four-unit female call accompanied by male calls (spectrogram 4a). The female call begins at about 12 kHz and descends in a series of four whistles to about 5 kHz. The first two units are rather smooth in their descent, but the final two have fine oscillations in pitch (spectrogram 4b, seconds 1.5-2.5). The entire call lasts about 2.5 seconds. The male calls, very faint in this spectrogram, are wide-band chirps that rise and fall from about 7 kHz to 11-12 kHz in about 0.15 seconds. The male repeats his calls with a periodicity of about 0.6 to 1.1 seconds. The period seems to gradually increase when the female is not calling. Relative to other tarsier duets, the duet in this spectrogram does not seem to be particularly well synchronized between the male and female calls. Few recordings were made from this locality, and it cannot be concluded whether this is characteristic of the Sejoli form. The Tinombo form is a previously undescribed acoustic form. It is remarkable for the vocal diversity of the female. In its simplest form, it bears similarities to the Togian form (see below), but the female repertoire is far greater than in Togian Island tarsier. Structurally, it is a simple duet with one female call followed by two to four male calls (spectrogram 5a). The female call is typically in two units, but is sometimes a single unit (not figured here), particularly in the early part of a duet. The basic female unit is a hook-shaped whistle that descends from around 12 kHz to about 5 kHz in about 0.4 seconds. Two of these units are sometimes given in rapid succession, the intervening gap being only 0.1 seconds (spectrogram 5b, far left). More often, the hookshaped unit is preceded by a peculiarly modulated whistle that begins at around 13 kHz, descends to below 10 kHz, rises again to a point higher than the initial frequency, and finally, descends to nearly 5 kHz, all in the span of only 0.3-0.4 seconds (spectrogram 5b, middle). Another female variant, much rarer than

44

the first two described forms, is a series of five hookshaped units that occur in rapid succession (spectrogram 5b, far right). Each of the units has a terminal frequency of about 5 kHz, but the maximum frequency of each unit gradually descends, the first having a maximum frequency of about 14 kHz and the last having a maximum frequency of just 10 kHz. The male calls are also quite interesting. They are wideband chirps, like other male tarsier calls, but they modulate from 5 kHz to over 13 kHz, and back again to 5 kHz, in about 0.10 seconds. The male produces these calls at a rate that varies from about 0.5-1.0 seconds, the interval lengthening when the female is not calling. The T. dianae form was originally described by Niemitz et al. (1991) based on recordings made from Kamarora. They published renderings of a spectrogram that has been re-scaled here for comparative purposes (spectrogram 6a). This acoustic form features a longer duet that is characterized by many repeated calls. It is a challenge to make accurate, legible spectrograms of these duets due to their length. The re-scaled version of the figure from Niemitz et al (1991) offers a glimpse at the overall character of the duet. In the early portion of the duet, the female gives fast, repetitive wide-band chirps that descend from around 15 kHz to around 5 kHz in only 0.3 seconds or less, at the rate of 2-3 calls per second. The rate of repetition gradually increases to about 7 calls per second, as the range of frequency modulation gradually narrows to a band centered at about 10-11 kHz (spectrograms 6b, seconds 6-12; 6c, seconds 49). The female gives about 7 relatively narrow-band calls that are between 10-11 kHz, and last about 0.7 seconds each. Following this, she begins again to give rapid chirps, first in a narrow band centered around 10 kHz, then gradually increasing the frequency range of the calls until they descend from about 12 kHz to 8 kHz. The male call is also a wideband chirp that descends from around 8 kHz to around 5 kHz. These are repeated at the rate of about two per second. The frequency range of these calls widens slightly as the female’s calls flatten out. It is during this portion (spectrograms 6b, seconds 6-12; 6c, seconds 4-9) that the duet becomes more highly

Primates of The O riental Night

synchronized, with the male and female calls beginning in almost perfect unison (spectrograms 6d, seconds 4.0-6.5; 6e, seconds 4-7). Following this, synchronization of male and female calls remains tight with the female giving either two or three calls per male call (spectrograms 6d, seconds7-10; 6e, seconds 7-10). Spectrograms of the Togian form were originally published by Nietsch and Kopp (1998) (spectrogram 7a). Structurally, this is a simple duet with one female call being followed by two or three male calls. Like the Tinombo form, the female call is a hook-shaped whistle, although the top of the hook is less smooth in the Togian tarsier. The maximum pitch is about 12-13 kHz (although the figure by Nietsch shows this to be only about 10 kHz), and there is a sharp descent to around 6 kHz. The call lasts about 0.5 seconds and is repeated at the rate of about 1 call per 1.5 seconds. The male call is shaped like a temporally compressed version of the female call. Maximum pitch is around 10-11 kHz, and the call descends precipitously to 5 kHz or lower. The male call lasts only about 0.15 seconds. Male calls are performed as a series of 2 or 3 calls that gradually ascend in maximum pitch. DISCUSSION Acoustic Data: Results from this study reveal seven acoustic forms of tarsiers in the study area. Within a form, there are self-evident similarities in duet structure. There is also excellent interobserver reliability between MacKinnon and MacKinnon (1980), Niemitz et al. (1991), Nietsch and Kopp (1998), and myself, evidenced in the spectrograms. Spectrograms recorded from tarsiers at Tangkoko in the late 1970’s, late 1980’s, and late 1990’s show stability of form that is reassuring that form of tarsier call persists over time. Mate Recognition theory predicts that tarsier vocalizations are a species-specific system that evolves in such a way as to advertise the fitness of mating partners. It follows, then, that so too must the tarsier auditory system evolve in such a way as to receive and process those vocalizations, filtering

signal from noise. Thus, for the purposes of this study, an acoustic form of tarsiers is the most inclusive group of animals that have a similar and appropriate behavioral response to a given vocalization (e.g. groups of tarsiers that respond to a recorded vocal duet with the performance of a vocal duet). Hypothetically, each acoustic form is a distinct taxon, but this remains to be verified. It is easy to see from the spectrograms, that there is quantitative variation in the vocal duets within an acoustic form, and qualitative variation among acoustic forms. Some of the most obvious quantitative variation is the frequency (in kHz) of the female calls seen in spectrogram 2. For example, the final note of the female’s crescendo varies from around 9 kHz in spectrograms 2b and 2c, to 11 kHz in spectrogram 2a and 2e, to fully 15 kHz in spectrogram 2c. Since 2a, 2b, and 2c are all from one locality, Tangkoko, it is probable that this is variation that normally occurs within an acoustic form. In contrast, however, it is more difficult to make meaningful quantitative comparisons among the acoustic forms. What are the homologies, for example, between the crescendos of the Manado form and the hook-shaped female calls from the Togian Islands? The nature of quantitative comparisons of differences is that they are, in fact, comparisons of similarities. Things that are different are, simply, different. Thus it is with tarsier acoustic forms that some characters suitable for quantitative comparisons within a form are unsuitable for comparisons among forms. Homologies for quantitative comparison likely do exist among acoustic forms, but finding them will require further study. Notes on Tarsier Taxonomy: Several issues trouble the taxonomy and classification of Sulawesian tarsiers. Primary among these was, until recently, the lack of a type specimens for Tarsius tarsier (see Groves et al. this volume). Second to this is the likelihood of unrecognized cryptic taxa. Investigations into acoustic variation have uncovered eight distinct acoustic forms: the seven described in this study plus MacKinnon and MacKinnon’s (1980) Palu form. Third, there is a large

45

Shekelle- Distribution of Tarsier Acoustic Forms

problem between two independent lines of investigation: surveys of museum specimens are based primarily on analyses of skeletons and pelts, while field surveys rely heavily on acoustic data and data from wild tarsiers that are trapped and released. While complementary studies often lead to dynamic advances in understanding, the unfortunate situation at hand is that many diagnostic characters used in the museum studies are impractical to score on a living tarsier (e.g. relative inflation of the auditory bulla anterior to the carotid foramen), and museum specimens cannot be scored for acoustic forms. Thus, the kernel of the problem is complementary taxonomic investigations have produced data sets that cannot be compared with each other. A clear taxonomic statement that identifies T. tarsier, clarifies the relationship between known acoustic forms and described taxa, and which states the distribution of these forms, where known, is needed. Tarsius tarsier: This taxon is based on Buffon’s tarsier and is believed to come from Makassar (Groves et al. this volume). Makassar was not sampled in this study, but subsequent surveys show a highly distinctive duet from near to Makassar (unpublished data). Thus, each of the acoustic forms discussed here are likely to be distinct taxa. Tarsius sangirensis: Although this taxon was treated as a subspecies by both Sody (1949) and Hill (1955) and synonymized with T. spectrum by Niemitz (1984a), there is general agreement between Feiler (1990), Shekelle et al. (1997), and Groves (1998) that this is a valid species. The type locality is Greater Sangihe Island and the type specimen, according to Hill (1955) is in the Dresden Museum. This taxon is distributed sparsely throughout Greater Sangihe Island. There are reports that tarsiers exist on other islands in the Sangihe Island group (e.g. Siau Island). Subsequent surveys indicate tarsiers are still present on Siau, although they are quite rare, and the acoustic form is distinct from that of T. sangirensis. The local name is sengkasi (bahasa Sangihe). One family referred to tarsiers as higo. Sangihe Island tarsiers can be easily and confidently diagnosed from T. tarsier by the tail and

46

the tarsi, both of which are more sparsely haired with hairs that are shorter than in T. tarsier—the characters on which this taxon was based (Meyer, 1897). These distinctive characteristics of T. sangirensis are intermediate between T. tarsier and T. syrichta, which is curious because Greater Sangihe Island is about midway between Sulawesi and the Philippine island of Mindanao. Tarsius sangirensis, nevertheless, is clearl related to the T. tarsier-complex. Regarding Sangihe Island tarsiers, Musser and Dagosto (1987) state that “the tail, although less densely haired than in typical T. spectrum, is not at all similar to the sparsely haired tail of T. syrichta.” While this is true, it does not argue against the uniqueness of T. sangirensis. They also state that, “the tarsus is sparsely haired, but some individuals of T. spectrum do resemble T. syrichta in this regard”. I disagree. None of the mainland Sulawesian tarsiers in this study approximated the condition in T. syrichta, nor did they resemble T. sangirensis. Furriness of the tarsus changes with age: infants have densely haired tarsi with hair that extends onto the hands and feet. Adults gradually lose some of that hair, particularly on the hands, feet and ventral/superior aspect of the tarsi, which may appear nearly nude in adults (unpublished data). The dorsal/inferior aspect of the tarsi of T. tarsier are always haired. The same is true of T. sangirensis, but the hair is shorter and less dense. Groves (1998) found that T. sangirensis is distinctive from all other Sulawesian tarsiers in having large and broad skulls, long toothrows, and short lateral incisors. Shekelle et al. (1997) found them to be distinguished from other Sulawesian taxa in having a higher average body weight, a unique acoustic form, and several unusual behavioral characteristics including a tendency toward less sociality while sleeping, and a preference for more exposed sleeping sites—which may be due, in part, to habitat degradation. Some authors including Groves (1998) have indicated that T. sangirensis may have a less distinctive postauricular white spot (a synapomorphy that links all Sulawesian taxa), and finer, less wooly fur. To the former, I disagree. The fur of T. sangirensis is perhaps finer and less wooly than that of T. tarsier,

Primates of The O riental Night

but it is also most certainly lighter. The dorsal aspect of the body and limbs is a light, milk chocolate-like brown. It has less of the distinctive brown and black mottled appearance of T. spectrum. The underside is almost pure white as is the postauricular white spot (see photo of T. sangirensis in Rowe, 1996, p.55—the postauricular white spot is visible at the base of the ear, and, faintly, at the top of the ear). It may be that the museum specimens bear some artifacts of preservation or aging that have turned the pelts dark. Tarsius dentatus: This taxon was treated as a questionably valid subspecies by both Sody (1949) and Hill (1955), and was synonymized with T. spectrum by Niemitz (1984). The infant tarsier in Rowe (1996) that is labeled as T. dentatus is misidentified. It is a tarsier from Libuo. The type locality of T. dentatus is Labua Sore. Dr. Lenora Bynum studied macaques in that region for several years in the early 1990’s and her surveys indicate that the site, which is also important for macaque taxonomy, goes by the name Labuan Sore (pers. comm.). When I visited there in April and May of 1996 local people agreed that there is a coastal site called Labuan Sore. [This is, perhaps, a corruption of labuhan sore, or, “evening anchorage”]. It is coastal and treeless, so I surveyed the nearby agricultural lands around the village of Marantale and used these as my reference sample for T. dentatus. Tarsiers at Marantale exhibited the vocal duets of T. dianae. Thus, it raises the strong possibility of nomenclatural conflict. Tarsius pumilus: There is agreement that T. pumilus is a valid taxon (Niemitz 1985, Musser and Dagosto 1987, Groves 1998, and Maryanto and Yani 2004). It is known from only three localities: the type locality, Rano Rano, at 1800 m in Central Sulawesi, the Latimojong Mountains at 2200 m in South Sulawesi, and from 2200 m on Mt. Rorekatimbu, but it is thought to be distributed throughout the mossy montane forests of the central region of Sulawesi at elevations over 1800 m. It is distinctive in its very small body size, head and body length being about 75% that of T. tarsier (Musser and Dagosto, 1987). The controversy with this taxon is not its validity, but its whereabouts. Tremble et al. (1993)

surveyed the locality listed in Musser and Dagosto (1987) as being the type locality for over one month without finding any sign of tarsiers (Yakob Muskita, pers. comm.). An experienced six-member team spent a chilly night at 1800 m on the flank of Mt. Nokilalaki for this study without finding evidence of tarsiers. Musser spent considerable time in the montane forests of Central Sulawesi in the 1970s without encountering tarsiers (Musser and Dagosto, 1987). The question remains where is T. pumilus? Tarsius pelengensis: This taxon was described by Sody (1949) but considered weak by Hill (1955) and synonymized with T. spectrum by Niemitz (1984a). Musser and Dagosto (1987) noted that museum samples from Peleng were distinctively large, but they refrained from confirming its status as a valid subspecies, conclusions that were also reached by Groves (1998). Acoustic surveys were conducted by James Burton (Nietsch and Burton 2002), and the resulting spectromgrams show clear similarities with those of T. dentatus. Tarsius dianae: Described as a distinct species by Niemitz et al (1991) based upon a unique vocal duet, several minor characters of appearance, some behavioral differences, and possibly a unique karyotype. As mentioned previously, there is a likely nomenclatural conflict between this taxon and T. dentatus. Manado form: This form exists as an acoustic variant and not yet a recognized taxon. It is synonymous with the Manado form of MacKinnon and MacKinnon (1980) and Shekelle et al (1997). Based on acoustic data, the distribution is from the northeastern tip of Sulawesi to the faunal break at Gorontalo. It includes Tangkoko Nature Reserve, from where most research on wild Tarsius spectrum originates. The most common local names are tangkasi (Bahasa Minahasa) in the northern part of the range, and mimito (Bahasa Gorontalo) in the southern part. Libuo form: This form exists as an acoustic variant and not yet a recognized taxon. It is synonymous with the Gorontalo form of MacKinnon and MacKinnon (1980) and the Libuo form of Shekelle et al (1997). Acoustic surveys presented here suggest

47

Shekelle- Distribution of Tarsier Acoustic Forms

that it has a much more restricted range than that presented by MacKinnon and MacKinnon, who report the range as being the same as that of M. hecki, i.e. the entire northern peninsula from Gorontalo to just north of Palu. My acoustic surveys indicate that the distribution of this acoustic form is limited to an area no greater than that bounded by Gorontalo in the east, and Moutong/Molosipat in the west. The local name of tarsiers in this region is mimito (Bahasa Gorontalo). Sejoli form: This form is an acoustic variant and not yet a recognized taxon. It is synonymous with the Sejoli form of Shekelle et al. (1997), and replaces the Gorontalo form of MacKinnon and MacKinnon (1980) in the area around the North Sulawesi-Central Sulawesi provincial boundary. Sejoli is a small village in the vicinity of Moutong and Molosipat. Acoustic surveys indicate that the distribution of this acoustic form is limited to an area no greater than that bounded by Tanjung Panjang in the east, and Tinombo in the west. Tinombo form: This form is an acoustic variant and not yet a recognized taxon. It is synonymous with the Tinombo form of Shekelle et al. (1997), and replaces the Gorontalo form of MacKinnon and MacKinnon (1980) in the vicinity of Tinombo. Acoustic surveys indicate that the distribution of this acoustic form is limited to an area no greater than that bounded by Sejoli in the east, and Marantale in the south. Togian form: This form is an acoustic variant and not a recognized taxon. It is synonymous with the Togian form of Shekelle et al. (1997), and the Togian tarsiers of Nietsch and Kopp (1998). The confirmed distribution of this acoustic form is limited to the islands of Malenge and Batudaka, but it is reasonable to assume that its distribution extends to all of the Togian Islands (except perhaps Una Una), as the Togian Islands were a single land mass and were possibly connected to Sulawesi as recently as the last Ice Age (Whitten et al. 1987). Nietsch and Kopp (1998) has argued for the taxonomic separation of this acoustic form on the strength of experimental evidence from playback tests where captive T. tarsier were exposed to recordings of the vocalizations of the

48

Togian tarsiers, as well as those of T. dentatus, and conspecific T. tarsier. Several language groups exist in the Togian Islands and there are likely to be many local names. I recorded a few including: bunsing, tangkasi, and podi. Palu form: MacKinnon and MacKinnon (1980) originally described this acoustic form. They list the distribution as being the valley of the Palu River. This form has recently been described as T. lariang by Merker and Groves (2006). Other acoustic forms: MacKinnon and MacKinnon (1980) reported that a colleague, Dr. Dick Watling, recorded three additional tarsier acoustic forms, all from Central Sulawesi, but they did not state where in Central Sulawesi he made these recordings. It is not necessary to assume that Watling’s forms are all new and unpublished. For example, it could be that his forms are, say, T. dianae, the Tinombo form, and the Togian form, all of which come from Central Sulawesi. Nevertheless, additional acoustic forms almost certainly exist, undiscovered, on Sulawesi. ACKNOWLEDGEMENTS I thank Dr. Jatna Supriatna, Dr. Noviar Andayani, Suroso Mukti Leksono, Luluk Lely Soraya Ichwan, and Yunus Masala. Permits and logistical support were arranged for by LIPI; Department of Forestry; KSDA in Manado, Bitung, Gorontalo, and Palu; Lore Lindu National Park; and Dumoga-Bone National Park. Special thanks are due to Ibu Wati (LIPI), Pak Endang (SBKSDA Manado), Pak Rolex Lameande (KSDA Palu). Also notable is the assistance and advice provided by Dr. Joe Erwin, Dr. Ed Mulligan, and Dr. Mitchell Sommers. Terry Gleason read drafts of this paper and made important contributions to the section on biogeography. Jacinta Beehner and Kellie Glasscock also read drafts and made useful comments. Financial support was provided by the following organizations: National Science Foundation, Primate Conservation, Inc., Washington University Department of Anthropology, L.S.B. Leakey Foundation, Explorer ’s Club, Wenner Gren Foundation, Garuda Indonesia Airlines.

Primates of The O riental Night

REFERENCENS Banks, E. 1949 Bornean Mammals. Kuching, Malaysia: The Kuching Press. Bearder, SK, PE. Honess & L. Ambrose. Species diversity among galagos with special reference to mate recognition. In: Alterman L et al., editors. Creatures of the Dark: The Nocturnal Prosimians. New York, Plenum press. p 331-352. Feiler, A. 1990. Ueber die Saugetiere der Sangiheund talaud-Inslen- der Beitrag AB Meyers Fur ihre Erforschung (Mammalia). Zoologische Abhandlungen Staatliche Museum fur Tierkunde in Dresden 46:75-94. Fooden, J. 1980. Classification and distribution of living macaques (Macaca Lacepede, 1799). In Lindburg DG, editor. The Macaques: Studies in Ecology, Behavior, and Evolution. New York: Van Nostrand Reinhold Company. p1-9. Groves, CP. 1980. Speciation in Macaca: The view from Sulawesi. In Lindburg DG, editor. The Macaques: Studies in Ecology, Behavior, and Evolution. New York: Van Nostrand Reinhold Company. Groves, C. 1998. Systematics of tarsiers and lorises. Primates 39(1):13-27. Gursky, S. 1994. Infant care in the spectral tarsier. International Journal of Primatology 15(6):843-855. Gursky, S. 1995. Group size and composition in the spectral tarsier, Tarsius spectrum: implications for social organization. Tropical Biodiversity 3(1):57-62. Gursky, S. 1997. Modeling maternal time budgets: the impact of lactation and infant transport on the time budget of the spectral tarsier, Tarsius spectrum. Dissertation Thesis. SUNY Stony Brook. Gursky, S. 1998. Conservation status of the spectral tarsier Tarsius spectrum: population density and home range size. Folia Primatoligica 1998:69(suppl 1):191-203.

Hill, WCO. 1953. Notes on the taxonomy of the genus Tarsius. Proceedings of the Zoological Society of London 123:13-16. Hill, WCO. 1955. Primates: Comparative Anatomy and Taxonomy. II. Haplorhini: Tarsioidea. Edinburgh: Edinburgh University Press. Horsfield, T. 1821. Zoological Researches in Java. London: Black, Kingsbury, Parbury, Allen. Clark, WEL. 1924. Notes on the living tarsier (Tarsius spectrum). Proceedings of the Zoological Society of London p217-223. MacKinnon, J & K. MacKinnon. 1980. The behavior of wild spectral tarsiers. International Journal of Primatology 1(4):361-379. Maryanto, I & M. Yani. 2004. The third record of pygmy tarsier (Tarsius pumilus) from Lore Lindu National Park, Central Sulawesi, Indonesia. Tropical Biodiversity 8(2): 79-85. Masters JC. 1998. Speciation in the lesser galagos. Folia Primatologica 69(suppl 1):357-370 Miller Jr., GS, & Hollister N. 1921. Twenty new mammals collected by H. C. Raven in Celebes. Proceedings of the Biological Society of Washington 34:93-104. Musser, GG, & M. Dagosto. 1987 The identity of Tarsius pumilus, a pygmy species endemic to the montane mossy forests of Central Sulawesi. American Museum Novitates (2867):1-53. Niemitz, C. 1984a. Taxonomy and distribution of the genus Tarsius Storr, 1780. In Niemitz C, editor. The Biology of Tarsiers. New York: Gustav Fischer Verlag. p1-16. Niemitz , C. 1984b. Vocal communication of two tarsier species (Tarsius bancanus and Tarsius spectrum). In Niemitz C, editor. The Biology of Tarsiers. New York: Gustav Fischer Verlag. p129-142. Niemitz, C. 1985 Der Koboldmaki-Evolutionsforschung an einem Primaten. Naturwiss Runsch 38:43-49. Niemitz, C, A. Nietsch, S. Warter, & Y. Rumpler. 1991. Tarsius dianae: A new primate species from Central Sulawesi(Indonesia).” Fol ia Primatologica 56:105-116.

49

Shekelle- Distribution of Tarsier Acoustic Forms

Nietsch A. 1993. Vocal acoustics and social behavior in tarsiers. In Creatures of the Dark: The Nocturnal Prosimians Conference in Durham, North Carolina. Nietsch A. 1994. A comparative study of vocal communication in Sulawesi tarsiers. ICongress of the International Primatological Society in Denpasar, Bali, Indonesia. p310 1994. Nietsch, A & C. Niemitz. 1993. Diversity of Sulawesi tarsiers. Deutsche Gesellschaft fur Saugetierkunde 67:45-46. Nietsch, A & M. Kopp. 1998. Role of vocalizations in species differentiation of Sulawesi tarsiers. Folia Primatologica 69(suppl 1)371-378. Nietsch, A & J. Burton. 2002. Tarsier Species in Southwest and Southeast Sulawesi. Abstracts, The XIXth Congress of the International Primatological Society (IPS), 49 Aug. 2002, Beijing, China: 20-21. Pallas, PS. 1778. Novae species quad e glirium ordine cum illustrationibus variis complurium ex hoc ordine animalium. Erlangen: W. Walther. Paterson, HEH. 1985. The recognition concept of species. In Vrba ES. editor. Species and Speciation. Pretoria: Transvaal Museum. p2129 Patton JL, MNF. da Silva, & JR. Malcolm. 1994. Gene genealogy and differentiation among arboreal

50

spiny rats (rodentia: echimydae) of the Amazon basin: a test of the riverine barrier hypothesis. Peres, CA, JL. Patton, & MNF. da Silva. 1996. Riverine boundaries and gene flow in Amazonian saddle-back tamarins. Folia Primatologica 67:113-124. Rowe, N. 1996. Pictorial Guide to the Primates. New York: Pogonia Press. Shekelle, M, S. Mukti, LLS. Ichwan & Y. Masala Y. 1997. The natural history of the tarsiers of North and Central Sulawesi. Sulawesi Primate Project Newsletter. Sody, HJV. 1949. Notes on some Primates, Carnivora, and the babirusa from the Indo-Malayan and indo-Australian regions. Treubia 20:121-185. Tremble, M, Y. Muskita & J. Supriatna. 1993. Field observations of Tarsius dianae at Lore Lindu Nation Park, Central Sulawesi, Indonesia. Tropical Biodiversity I(2):67-76. Wallace, AR. 1876. The Geographical Distribution of Animals, Vol. 1. London: MacMillan. Whitten, A, M. Mustafa & G. Henderson. 1987. The Ecology of Sulawesi. Yogyakarta, Indonesia: University of Gajah Mada Press. Woollard, HH. 1925. The anatomy of Tarsius spectrum. Proceedings of the Zoological Society of London 70:1071-1184.

Primates of The O riental Night

DISTRIBUTION OF TARSIER HAPLOTYPES FOR SOME PARTS OF NORTHERN AND CENTRAL SULAWESI Myron Shekelle1, Juan Carlos Morales2, Carsten Niemitz3, Luluk Lely Soraya Ichwan1, Don Melnick2 1

Center for Biodiversity and Conservation Studies, Faculty of Mathematics and Natural Sciences University of Indonesia, Depok 16421, Indonesia, Email: [email protected] 2 Center for Environmental Research and Conservation, Columbia University, New York, NY 10027, USA Email: [email protected]; [email protected] 3 Institute of Human Biology, Free University Berlin 14195 Berlin, Germany ABSTRACT DNA sequence data was taken from hair samples of wild Eastern tarsiers that were trapped-and-released. Twelve tarsier populations were sampled at approximately 100 km intervals along a transect that encircles Tomini Bay, running from Sangihe Island in the north, to the Togian Islands in the southeast. This study reports mtDNA haplotype data for 43 Eastern Tarsiers, and compares those with published sequences from Philippine and Western Tarsiers, as well as non-tarsier outgroups. A broad scale analysis, which included 21 unique tarsier haplotypes from 28 individuals, non-tarsier primates, and other mammals, found tarsiers to be a robustly supported monophyletic clade. The Eastern-Western-Philippine tarsier trichotomy was not resolved. Tarsier populations from Sangihe Island (T. sangirensis), which lies between Sulawesi and the Philippine island of Mindanao, and from the Togian Islands in Tomini Bay, were robustly supported monophyletic clades. Robust support was also provided for the basal position of T. sangirensis with respect to other Eastern tarsiers in this data set. A fine scale analysis, using only tarsiers and which included 26 unique haplotypes from 27 Eastern tarsiers samples (but used only the 3’ end of the 12s gene), also found the Sangihe and Togian populations to be robustly supported monophyletic clades. The basal position of T. sangirensis had only weak support in the fine scale analysis, however. In both analyses, broad scale and fine scale, other populations of Eastern tarsiers in this data set were generally paraphyletic or polyphyletic. Hypothesis testing found the most-parsimonious tree to be significantly shorter than trees constrained by by the assumption of monophyletic clades in by the assumption of monophyletic clades in regions of macaque endemism, geological microplates that compose Sulawesi, but could not refute the null hypothesis of no difference in overall tree length when constrained by the assumption of monophyletic clades within tarsier acoustic groups. Keywords: Tarsius tarsier, T. spectrum, T. bancanus, T. syrichta, T. dianae, T. sangirensis, T. dentatus molecular phylogeny, 12s, mtDNA

INTRODUCTION While tarsiers are commonly represented in molecular phylogenetic studies of primates, rarely has more than a single taxon been represented. Dijan and Green (1991) sequenced the involucrin gene for both T. bancanus and T. syrichta, and Adkins and Honeycutt (1994) sequenced the cytochrome oxidase c subunit II mtDNA gene for the same two taxa. Meireles et al. (2003) analyzed nuclear DNA sequence for those same two taxa at the globin locus. Never before has DNA sequence data been published for any Eastern Tarsier. The lack of knowledge about DNA sequence variation among tarsiers raises questions about the appropriate analysis of Eastern tarsiers. What is the relationship among Eastern, Western, and Philippine tarsiers? are each of the three species groups monophyletic? What is the most appropriate outgroup for an analysis of Eastern tarsiers?

In order to analyze patterns of haplotype variation among Eastern tarsiers (= Hill’s, 1955, Tarsius spectrum), it was necessary to conduct a broad scale analysis to address the larger questions about tarsier phylogenetics mentioned above. Since tarsiers are such a deep branch in the primate evolutionary tree, and since the outgroup to tarsiers has not been definitively determined despite the widespread acceptance of a monophyletic Haplorhini (see Morales et al. 1999, Yoder 2003), a fairly slowly evolving section of DNA was required. The 12s ribosomal RNA gene of the mtDNA genome has been used for addressing phylogenetic questions regarding Primates and superordinal relationships of mammals, and there is a substantial amount of comparative sequence data available (e.g. Springer and Douzery 1996, McNiff and Allard 1998). The 12s gene also has some hypervariable regions in the stem-and-loop structure,

51

Shekelle, Morales, Niem itz, Ichwan & Melnick - Distribution of Tarsier Haplotypes

notably a long loop near the 3’ end of the gene, that are valuable for population level analyses. Secondary structure of the gene product facilitates alignment of this variable-length gene (Springer and Douzery 1996). The 12s gene, therefore, is a practical compromise for this study, because it contains conservative regions with which to address issues of broad scale tarsier phylogeny as well as hyper-variable regions for a population level analysis of Eastern tarsiers. Shekelle et al. (2001) reported results of a preliminary analysis of this data set. Eastern, Western, and Philippine tarsiers were found to be an unresolved trichotomy. Genetic distances among the three tarsier species were comparable to genetic distances among Hylobates and two other hominoid genera, Pan and Homo, measured at the same locus, indicating a relatively old split, conceivably dating to the middle Miocene. Tarsius sangirensis was found to be the outgroup of other Eastern tarsiers in the data set reported here: large areas of Sulawesi remain unsurveyed for tarsier genetic diversity, so it has not been verified that T. sangirensis is the outgroup of all Eastern tarsiers, nor even that Eastern Tarsiers, as a whole, are monophyletic. Togian tarsiers were an autapomorphic subset with a diagnostic 2 base pair deletion in the hyper-variable loop near the 3’ end of the 12s gene. The taxonomy of Eastern Tarsiers bears on the question of Sulawesi biogeography. There are two broad categories of hypotheses regarding this topic. One such category derives from empirical biological data, notably the Sulawesi macaques. MacKinnon and Mackinnon (1980) offered an implicit hypothesis that a unique taxon of tarsier would coinhabit the distribution of each of seven macaque taxa, thus creating zones of primate endemism. They further observed that regions that are biogeographically linked to Sulawesi, which possess native tarsier populations, but which lack native macaque populations, represent distinct biogeographic regions where one could expect to find endemic tarsier taxa, e.g. the offshore island groups of Selayar, Banggai, and Sangihe, A similar hypothesis was implied by Niemitz et al. (1991). A second category of biogeographic hypotheses for Sulawesi derives from empirical

52

geological data. Prior to its current form, Sulawesi was an archipelago formed of numerous microplates of Asian, Australian, and oceanic origin. Hall (1996, 2001) reconstructed the geological evolution of Sulawesi by identifying these microplates and charting their movement over the past 50 million years. Evans et al. (2003) used genetic surveys of two distantly related taxa, primates of the genus Macaca and toads of the genus Bufo, to address hypotheses of Sulawesi biogeography. They found concordant distributions of macaques and toads, which they interpreted to indicate a shared history of range fragmentation. The faunal boundaries in their study showed little correspondence with the microplates identified by Hall (1996). Shekelle and Leksono (2004) used the distributions of tarsier acoustic forms to address the same topic. Using classic tools of biogeography, they layered the map of Evans et al. onto the map of Hall. Then they plotted the distributions of tarsier acoustic forms on the composite map. They found a nearly one-to-one correspondence between the distributions of tarsier acoustic forms and the composite map that combined the biological data and geological data. They reasoned that macaques were relatively recent immigrants to Sulawesi with much of their evolution occurring during the Pleistocene, after tectonic activity had already formed Sulawesi in its modern state. Macaque biogeography was, therefore, likely to have been shaped by Pleistocene range fragmentation/vicariance events. Indeed, many of the faunal boundaries in Evans et al.’s study appear to be consistent with geographic boundaries that are influenced by ocean level (e.g. the isthmus of Gorontalo, the Tempe depression). Tarsiers, on the other hand, were an older radiation that probably immigrated to Sulawesi in the Miocene. Tarsier biogeography would, therefore, be influenced by the geologic history of the microplates to a much greater extent than would that of the macaques, which may have colonized Sulawesi after the microplates had already coalesced. The tips of the tarsier branches would, nonetheless, be reshaped by the Pleistocene events that shaped the biogeography of macaques, and thus, tarsier biogeography also has elements of

Primates of The O riental Night

macaque biogeography. Shekelle and Leksono called this the “hybrid biogeographic hypothesis” for Sulawesi, because it combined empirical data from biology and geology and made explicit the observation that the time of dispersal to Sulawesi was one critical component that would affect biogeography. METHODS DNA sequence data were collected from hair samples from 101 wild-caught Eastern tarsiers (Figure 1). Geographic representation included: Sangihe (n = 5), Tangkoko (n = 21), Basaan (=Ratatotok) (n = 5), Molibagu (n = 14), Suwawa (n = 7), Libuo (n = 15), Sejoli (n = 6), Tinombo (n = 8), Marantale (n = 6), Kamarora (n = 7), Malenge (n = 5), Batudaka (=Wakai) (n = 2). This was supplemented by published sequence for T. bancanus borneanus, T. syrichta syrichta, and other mammals (from gen bank). Incomplete sequence data required many specimens to be excluded from the analysis presented here. Extraction of total genomic DNA (tDNA) from samples was accomplished using Qiagen extraction kits (#29304 and #29306). Tissue samples from wildcaught tarsiers were from plucked hair. DNA

concentration varied, and was adjusted accordingly using an estimate of the gel visualization to produce template DNA of nearly the same concentration. The target DNA was amplified with the polymerase chain reaction (PCR). A 50 microliter (μL) PCR reaction consisted of the following volumes: 30 μL deionized water (dH20), 10 μL 5X buffer solution 8.5 pH, 5 μL dNTP, 1 μL primer one (20 pM/μL), 1 μL primer two (20 pM/μL), 1 μL Taq polymerase, 3.5 mM Mg++ (1 heat-released bead), 1 μL template DNA. The 12s gene is about 950 base pairs (bp) in length. The following primers were used in various combinations to amplify the 12s gene in 1, 2, or 3 segments: 651f, 891f, 1247f, 933r, 1259r, and 1559r (Table 1). The primer names correspond roughly to the nucleotide number of the human sequence of the primer’s 5’ end. The reaction was placed in a thermal cycle machine set to the following cycle parameters: “hot start” = 94° (15 sec.) (first cycle only); 35 cycles of denaturation = 94° (30 sec.), annealing = 58° (30 sec.), and extension = 72° (60 sec.); final extension = 72° (7 min.) (last cycle only); and hold = 4° (infinity). The annealing temperature was sometimes varied to improve amplification with various primer pair combinations. Numerous other permutations of

Figure 1: Sampling localities for this study.

53

Shekelle, Morales, Niem itz, Ichwan & Melnick - Distribution of Tarsier Haplotypes

Table 1a. PCR and sequencing primers

1 2

Primer Name

Length

Sequence

651f (= 378)

18-mer

AGG TTT GGT CCT AGC CTT

891f (tarsier)

21-mer

A GG GTT GGT CAA TTT CGT GCC

925r-T.spec

17-mer

GCT TTA CGC CGT GCT TT

930r-T.ban

18-mer

CGC TTT ACG CCG GAT ATT

933r

20-mer

ATC TAA AAC ACT CTT TAC GC

1169r (tarsier)

23-mer

GGG A TG TGA AGC ACC GCC AAG TC

1247f (tarsier)

24-mer

CCC GAT A AA CCT TAC CAC CCC TTG

1259r

21-mer

GGT TTG CTG AA G ATG GCG GTA

1559r (=550r)

24-mer

CCA GTA CAC TTA CCA TGT TAC GAC

CA in T. syrichta G in T. syrichta

3 4

1

2

3

4

T,C also present GG in DUPC 6343

Table 1b. Some primer products and annealing temperatures (from Shekelle 2003)

Primer 1

Primer 2

Length

Optimal Annealing Temperature

651f

925r-T.spec

294 bp

53.5º

“

930r-T.ban

299 bp

53.7º

“

933r

302 bp

51.6º

891f

1259r

409 bp

56.3º

1247f

1559r

359 bp

54.4º

1247f

1169r

16,538 bp

58.3º

thermal cycle parameters were experimented with, but the above setting was deemed to be the best. Purification of the PCR product was achieved with the Qiagen PCR purification kit (#28106). Products were gel visualized with “DNA quant ladder” to estimate DNA concentration, evaporated on a speed vac, and then resuspended in a quantity of dH20 sufficient to make the concentration of all of the purified samples approximately equal. Sequencing of the purified PCR product used the Big Dye kit. A 10 mL sequencing reaction used the following volumes: 4 mL dH20, 4 mL “Big Dye” mix, 1 mL primer (20 pM/mL concentration), 1 mL purified PCR product. The same primers that were used for PCR were also used for sequencing. The reaction was placed in a thermal cycle machine set to the following cycle parameters: 35 cycles of

54

denaturation = 96° (05 sec.), annealing = 55° (10 sec.), extension = 60° (4 min.). Sequencing reactions were cleaned of impurities using sephadex. Five grams of sephadex and 80 mL of dH20 were combined and stirred until thoroughly mixed. An amount equal to 750-800 mL of mixture was aliquoted into spin columns. The columns were spun for 1 min. at 3000 rpm to remove excess moisture. The reactions were added to the top of the spin columns and were spun for 2 min. at 3500 RPM. The purified reactions were captured in 1.5 mL Eppendorf tubes, dried in a speed vac, and resuspended in 3 mL of loading dye (formamide dye mixed with 70 mL of loading solution). The samples were electrophoresed on a polyacrylamide gel (29:1) and were scored by an ABI 377 PRISM automated DNA sequencer. Both

Primates of The O riental Night

complementary strands were sequenced in order to double-check the reliability of the sequence data. Raw data were processed and pieced together usi ng Autoassembl er (ABI, Perkin El mer). Alignment was made by eye using comparative data from GENBANK and assumptions about the secondary structure of the 12s ribosomal RNA (Springer and Douzery 1996). A) Broad Scale Analysis of Tarsier Phylogenetics A broad scale analysis of tarsier phylogenetics was performed using DNA sequence data from the 12s ribosomal RNA region of the mitochondrial DNA genome. A data matrix was constructed of 900+ b.p. for 36 haplotypes. The matrix included unique haplotypes of 21 tarsiers, 1 strepsirhine, 10 anthropoids, 1 tree shrew, 1 flying lemur, 1 megabat, and 1 microbat. Eastern tarsiers were represented by 28 individuals. Individuals with identical haplotypes were grouped. Geographic representation was as follows: Sangihe (three haplotypes: ET048, ET049, ET050), Tangkoko (two haplotypes: ET001-002-005014-082, ET083); Basaan(=Ratatotok) (two haplotypes: ET084, ET085), Molibagu (three haplotypes: ET018020-023-024-027-029, ET019, ET026), Suwawa (not represented in the broad scale analysis), Libuo (two haplotypes: ET038, ET041), Sejoli (two haplotypes: ET096, ET100), Tinombo (one haplotype: ET074), Marantale (not represented in the broad scale analysis), Kamarora (one haplotype: ET062), Togian (three haplotypes: ET052, ET056, ET057). Also represented in the data matrix were other sequences taken from the literature: 1 Philippine tarsier (T. syrichta syrichta), 1 Western tarsier (T. bancanus borneanus), 1 strepsirhine (Lemur catta), 10 anthropoids (Homo sapiens, Pan troglodytes— two individuals, Pan paniscus, Gorilla gorilla, Pongo pymaeus—two individuals, Hylobates lar— two individuals, Papio hamadryas, and primate outgroups, the flying lemur (Cynocephalus variegatus), a tree shrew (Tupaia glis), a megabat (Donsonia mollucensis)), and a microbat (Eptesicus fuscus).

Forty-five base pairs were trimmed from the 5’ end of the gene, and 43 b.p. were trimmed from the 3’ end of the gene to accommodate for the PCR primers and areas where missing data predominated. This left a data matrix with 935 characters including gaps. The data set was rooted with the four non-primate taxa (the tree shrew, flying lemur and the two bats) using the “assume outgroup to be paraphyletic” option in PAUP. A parsimony analysis using PAUP [version 4.0b10 for Macintosh (PPC)] was used to produce strict consensus and bootstrap trees using two separate sequential approximation analyses. A sequential approximation analysis uses successive heuristic analyses, with characters being reweighted based upon the rescaled consistency index after each heuristic search. Heuristic searches, followed by character reweightings, are performed until successive analyses produce identical results. The rationale for this method is to allow the data themselves to adjust the weighting of relatively consistent characters versus relatively inconsistent ones. This method is particularly applicable for the 12s gene, since it includes both highly conserved and highly variable sites and saturation of some sites reduces their utility for older phylogenetic questions (Springer and Douzery 1996). The first heuristic search used equal weights for each character. Of the 935 total characters, 477 characters were constant and 118 characters were parsimony uninformative. This left 340 parsimony informative characters. The stepwise addition option was used, with 10 random replicates. Other options employed included: gaps are treated as “missing”, multistate taxa interpreted as uncertainty, branchswapping algorithm = tree-bisection-reconnection (TBR), steepest descent option not in effect, initial ‘MaxTrees’ setting = 100 (will be auto-increased by 100), branches collapsed (creating polytomies) if maximum branch length is zero, ‘MulTrees’ option in effect, topological constraints not enforced, trees are unrooted. A total of 108 most-parsimonious trees were found, each with a tree length of 1337. Following the first heuristic search, characters were reweighted based upon the rescaled consistency index using a base score of 100.

55

Shekelle, Morales, Niem itz, Ichwan & Melnick - Distribution of Tarsier Haplotypes

Characters were thus assigned a weight between 0 and 100, the latter number indicating a character that was completely consistent across the tree. A second heuristic search was performed using the same settings as the first, the only change being the reweighted characters. Four most parsimonious trees were found, each with a tree length of 50438 (because the reweighted characters use a base score of 100, the second heuristic search had a tree length that was 1 to 2 orders of magnitude greater than the first search). Following the second heuristic search, characters were again reweighted based upon the rescaled consistency index, and a third heuristic search was performed. Results of the third heuristic search were identical to the results of the second heuristic search and the sequential approximation analysis was deemed to be complete. B) Fine Scale Analysis of Eastern Tarsier Phylogenetics A second sequential approximation analysis was performed on a data set that used only the 3’ half of the 12s gene (~500 b.p.). Twenty-six unique haplotypes from 27 individuals were included in the analysis. Although the data matrix was comprised of only about half as many characters as in the broad scale analysis, it include the complete hyper-variable loop near the 3’ end of the 12s gene, which is more applicable to fine scale analyses. The geographic representation of Eastern tarsiers was as follows: Sangihe (two haplotypes: ET048, ET049), Tangkoko (three haplotypes: ET001, ET003, ET010); Basaan (=Ratatotok) (two haplotypes: ET 084, ET085), Molibagu (three haplotypes: ET018, ET019, ET025), Suwawa (two haplotypes: ET089, ET090), Libuo (two haplotypes: ET 032, ET034), Sejoli (three haplotypes: ET096, ET097, ET100), Tinombo (two haplotypes: ET072, ET077), Marantale (two haplotypes: ET066, ET068), Kamarora (two haplotypes: ET062, ET063), Togian (three haplotypes: ET052, ET054, ET056-058). The Philippine and Western tarsiers were used to root the analysis. The “root as basal polytomy” option in PAUP was employed.

56

The same PAUP settings were used, as were used in the broad scale analysis. The fine scale analysis had 518 characters, of which, 413 were constant. Fifty-eight characters were parsimony uninformative, leaving 47 informative characters. A total of 131,967 most-parsimonious trees were found, each with a tree length of 150. After reweighting of characters based upon the rescaled consistency index, 222 most-parsimonious trees, each with a tree length of 10931 were found. In the third and fourth heuristic searches, following reweighting of the characters after each search, 221 most-parsimonious trees were found, each with a tree length of 10931. RESULTS A) Broad Scale Analysis of Tarsier Phylogenetics Major elements of the strict consensus tree (Figure 2) included the erroneous placement of the flying lemur within Primates, a monophyletic Prosimii, and a monophyletic Philippine-Western tarsier clade. Within Eastern tarsiers, T. sangirensis was a basal outgroup, and Togian tarsiers and Tangkoko tarsiers were autapomorphic subsets. Tinombo and Kamarora tarsiers were represented by a single specimen and necessarily monophyletic. Sejoli, Libuo, and Molibagu tarsiers were each paraphyletic assemblages. Basaan (=Ratatotok) tarsiers were polyphyletic, one haplotype being basal to the Tangkoko clade and the other haplotype nested within the Molibagu clade. A bootstrap analysis using 1000 replicates was performed using the character weights that resulted from the sequential approximation analysis (Figure 3). All other options were the same as in the heuristic parsimony analyses. Highly supported elements of the bootstrap tree included 100% bootstrap value support for the monophyly of the following clades: Tarsius, Eastern tarsiers, T. sangirensis, and Togian tarsiers. Other phylogenetic structure that appeared in the strict consensus tree either collapsed or was supported by lower bootstrap values, in the range of 50-83%.

Primates of The O riental Night

B) Fine Scale Analysis of Eastern Tarsier Phylogenetics Major elements of the strict consensus tree (Figure 4) included a monophyletic clade of Sangihe tarsiers (T. sangirensis) being basal to other Eastern tarsiers in the data set. The Togian tarsier formed a monophyletic clade that was basal to the remaining tarsiers (i.e. non-Sangihe Eastern tarsiers). The remaining tarsiers clustered in a series of clades where, generally, primitive-to-derived haplotypes followed a south-to-north pattern (i.e. the most derived haplotypes are on the extreme northern end of Sulawesi), but where haplotypes from a given locality were not monophyletic. Haplotypes from three localities, Marantale, Sejoli, and Ratatotok were polyphyletic. Haplotypes from Kamarora, Tinombo, Libuo, Suwawa, and Molibagu, and Tangkoko were paraphyletic. The data set for the fine scale analysis was computationally intensive, probably because there were too few characters relative to taxa, and the bootstrap analysis was stopped after 230 replicates (Figure 5). The bootstrap analysis of the fine scale data set provided robust support for the monophyly of Eastern tarsiers, T. sangirensis, the Togian Island tarsier population, and T. dentatus, represented by a clade that consisted of two haplotypes found at Kamarora and one haplotype found at Marantale. Each of these clades was supported by bootstrap values between 91-100%. There was one major change in topology, with the bootstrap analysis finding the Togian Island tarsier population to be basal to other Eastern tarsiers in the data set, but this result was supported by a very low bootstrap value of only 54%. Other phylogenetic structure that appeared in the strict consensus tree either collapsed or was supported by lower bootstrap values, in the range of 66-81%. C) Other Analyses A final sequential approximation analysis was attempted using T. sangirensis to root the remaining Eastern tarsiers in the data set (not figured). This analysis was abandoned when it was found that there were only 31 parsimony informative characters for 26 taxa, many of those characters preferentially located on the branch that defines T. sangirensis.

D) Hypothesis Testing The mtDNA phylogeny produced in the finescale analysis was used to directly address questions about tarsier biogeography. 1. Do tarsiers co-inhabit regions of primate endemism with Sulawesi macaques, such as was hypothesized by MacKinnon and MacKinnon (1980) and Niemitz et al. (1991), and which would be consistent with the biogeography of macaques and toads found by Evans et al. (2003)? 2. Do tarsiers inhabit regions of endemism identified by the microplates of Sulawesi, such as might be predicted by Sulawesi’s geologic history as an archipelago? (Hall 2001) 3. Are tarsier acoustic and genetic groups statistically congruent and do tarsier acoustic groups diagnose discrete taxa, such as might be predicted by the mate recognition species concept of Paterson (1985), and MacKinnon and MacKinnon (1980), Niemitz et al. (1991), several papers by Nietsch (e.g. Nietsch and Niemitz 1993, Nietsch and Kopp 1998, Nietsch 1999), and therefore consistent with the hybrid biogeographic hypothesis for Sulawesi by Shekelle and Leksono (2004)? These three hypotheses, abbreviated as (1) “macaque”, (2) “microplates”, and (3) “acoustic form”, were tested by constructing constraint trees in MacClade and loading them into PAUP. In each case the constraint tree assumed a trichotomy among tarsiers, and monophyletic tarsier clades that were arranged in a star phylogeny (Figures 6, 7, 8). Thus, the acoustic form hypothesis test enforced a constraint tree in which each acoustic form had a monophyletic clade of haplotypes, but no other constraints on the topology were enforced, either within or among acoustic forms. A parsimony analysis to produce the most-parsimonious constrained tree was conducted for each hypothesis using the same PAUP settings that were used in the fine scale analysis, the only difference being the topological constraints. The macaque test found 48 mostparsimonious trees each with a tree length of 11,516. The acoustic form test found 34 most-parsimonious trees each with a tree length of 11,330. The microplate

57

Shekelle, Morales, Niem itz, Ichwan & Melnick - Distribution of Tarsier Haplotypes

Figure 2. Broad scale phylogenetic analysis—strict consensus tree.

58

Primates of The O riental Night

Figure 3. Broad scale analysis—bootstrap tree.

59

Shekelle, Morales, Niem itz, Ichwan & Melnick - Distribution of Tarsier Haplotypes

Figure 4. Fine scale analysis—strict consensus tree.

60

Primates of The O riental Night

Figure 5. Fine scale analysis—bootstrap tree.

61

62

monophyletic tarsier clades from those localities that share an acoustic form.

Figure 6. Acoustic constraint tree. This tree enforces

Figure 7. Macaque constraint tree. This tree enforces monophyletic tarsier clades from those localities that lie within each region of endemism identified by macaque distributions on Sulawesi.

Figure 8. Microplate constraint tree. This tree enforces monophyletic tarsier clades from those localities that lie within each region of endemism identified by the microplates that form Sulawesi.

Shekelle, Morales, Niemitz, Ichwan & Melnick - Distribution of Tarsier Haplotypes

Primates of The O riental Night

test found 235 most-parsimonious trees each with a tree length of 11,411. Most-parsimonious constrained trees generated by each hypothesis were tested against the most-parsimonious unconstrained tree with the non-parametric test (=Templeton test) in PAUP, assuming a one-tailed test. In the macaque and microplate analyses, the null hypothesis of no significant difference in tree length between constrained tree and unconstrained tree was very confidently rejected (P<0.005). In other words, the unconstrained tree is significantly shorter than the constrained trees generated by the macaque and microplate hypotheses described above. The null hypothesis for the acoustic form hypothesis, however, could not be rejected at the 1% confidence interval (Table 2). There is no significant difference between the length of the most parsimonious tree, and the most parsimonious tree constrained by the hypothesis of monophyletic clades identified by tarsier acoustic form. DISCUSSION Phylogenetic Analysis The most robust portions of the tree topology supported the monophyly of the Eastern tarsiers in this data set, and the monophyly of two isolated island populations, Sangihe and Togian tarsiers. The principal weaknesses were that the broad scale analysis could not confidently resolve the Eastern-Western-Philippine tarsier trichotomy, and the fine scale analysis had poor resolution for haplotypes from within insular Sulawesi. A Western-Philippine tarsier clade was supported by a bootstrap value of 72%, but experience shows that bootstrap values in this range are subject to instability in subsequent analyses. Given the unexpectedly large degree of genetic variation among tarsiers in general, and Eastern tarsiers, in particular, it was not practical to build a DNA sequence database that was sufficient to definitively address both the broad scale and fine scale taxonomic questions. Indeed, one of the most notable

results of this study also highlights a key drawback, that is, neither Philippine nor Western tarsiers are particularly closely related to Eastern tarsiers. Nevertheless, Philippine and Western tarsiers were required to root the fine scale analysis of Eastern tarsiers. In hindsight, the problem was roughly analogous to assembling a data set to answer phylogenetic relationships within Hominoidea from scratch using a lemur as the outgroup, while simultaneously addressing phylogeographic questions within the Hylobates lar group using an orangutan as the outgroup. Current evidence indicates that Western and Philippine Tarsiers are relatively distant cousins of the Eastern Tarsiers that diverged from the latter between 5.6 mya and 17 mya. Meireles et al. (2003) estimated the divergence of Philippine and Western Tarsiers at about 5.6 mya using a molecular clock based upon nDNA. It is logical to assume that the divergence of Eastern Tarsiers is at least as old as the Western and Philippine Tarsiers based on, 1) the morphologic data in Musser and Dagosto (1987) and Groves (1998) that supports a Philippine-Western Tarsier clade, and 2) the genetic data in this study that finds an unresolved trichotomy with weak support for a Western-Philippine tarsier clade. Several lines of evidence suggest that the divergence of Eastern Tarsiers is likely to be at least 9.5 mya. Morley (1998) offered palynological evidence from ocean core samples that showed evidence of biotic exchange across the Makassar Straits at 17, 14, 9.5, 3.5 and 1 mya. The two most recent of those dates, 3.5 and 1 mya, do not appear to be consistent with the tarsier genetic data of Meireles et al. (2003) or from this study. Hall (2001) estimated the most likely time for faunal exchange between Asia and Sulawesi as being about 10 mya, not necessarily via the Makassar Straits, but possibly via Java. Preliminary evidence from this study (Shekelle et al. 2001) offered a very rough molecular clock estimate of the divergence of Eastern tarsiers at about 13 mya, given that, 1) the average genetic distance between the Eastern tarsiers versus Western and Philippine tarsiers was found to be about 93% as great as the average genetic distance between Hylobates versus

63

Shekelle, Morales, Niem itz, Ichwan & Melnick - Distribution of Tarsier Haplotypes

Homo and Pan, and 2) Goodman et al. (1998) estimated the lesser ape-great ape split to be at least 14 mya. Additionally, Mercer and Roth (2003) gave a molecular clock estimate of 11.5 mya ago for the origins of the Sulawesi squirrels. There are, therefore, several independent lines of evidence that indicate that some of Sulawesi’s older endemic taxa may have origins dating from 9.5-17 mya. To put this problem in perspective with another analogy, it would be as though Evans et al. (2003) had been obliged to root their analysis of Sulawesi macaques with a baboon and a mandrill, or possibly even a langur and a colobus monkey. Future analyses will certainly benefit from using a more appropriate outgroup, such as the Sangihe tarsier or perhaps an as yet unsequenced Eastern tarsier that proves to be basal to the whole radiation (e.g. possibly T. pumilus), but this will have to be determined by further experimentation. Two other notable shortcomings of this study are that, 1) this data set analyzes DNA sequence from only a single gene and it is well known that different genes have different tree topologies, and 2) the gene in question is in the mtDNA genome, a genome that is known to be affected by the dispersal pattern of the organism in question. Matrilocal taxa and short-distance dispersers show more phylogeographic structure in their mtDNA gene tree than do patrilocal taxa and long-distance dispersers (Melnick and Hoelzer 1992, 1993). There are no data for dispersal among tarsiers (see Sussman 1999), so we cannot predict how dispersal patterns will affect these results.

Hypothesis Tests and Sulawesi Biogeography The hypothesis for Sulawesi biogeography that derives from empirical biological data, such as Evans et al. (2003), and the one based upon empirical geologic data (e.g. Hall 2001), have numerous areas of incongruence. Given the quality of the evidence that supports each of the two hypotheses above, this seems puzzling at first glance, i.e. why do the empirical biological data not fit the model based upon the geological history of the island? Adding further to the mystery, the tarsier genetic data in this study seem incompatible with either hypothesis, as results of this study found that constrained trees based upon those two hypotheses were very significantly longer than the unconstrained most-parsimonious tree. The distribution of tarsier acoustic forms appears to offer a novel solution to the issue of Sulawesi biogeography (Shekelle 2003, Shekelle and Leksono 2004). The distributions of tarsier acoustic forms share some boundaries with macaques, but also share some similarities with Sulawesi’s microplates. For instance, tarsiers and macaques share faunal boundaries at the isthmus of Gorontalo and the isthmus of Palu. But between the isthmus of Gorontalo and the isthmus of Palu, in the region of Macaca hecki, there are three tarsier acoustic forms—an area that not coincidentally has three microplates (Figure 9). Given the a priori predictions that acoustic forms are distinct taxa, it is remarkable that a unique acoustic form is present in almost every biogeographic region predicted by the hybrid biogeographic hypothesis.

Table 2: Results of hypothesis tests. The null hypothesis of no significant difference between tree lengths was rejected in all three tests using the Templeton test. Two of the results, the macaque and microplate tests, were very highly significant. The acoustic test was not rejected at the 1% confidence interval.

tree

64

tree length

N

z

critical value 0.01

t

p

most parsimonious

10,931

-

-

-

-

-

acoustic

11,330

10

1.8869

9

5

>0.01

macaque

11,516

12

2.7477

4

10

<<0.005

microplate

11,411

9

2.6656

0

3

<<0.005

Primates of The O riental Night

In retrospect, the implicit hypothesis that tarsiers might share regions of endemism with macaques, e.g. MacKinnon and MacKinnon (1980), Niemitz et al. (1991) was influenced by the somewhat limited understanding of Sulawesi’s geologic history that existed prior to the geological reconstructions of Hall in the mid 1990’s. Also, the implicit assumption that the Eastern tarsier radiation took place at approximately at the same age as the Sulawesi macaque radiation had not been examined. This assumption, however, is almost certainly false. It can be inferred that the Sulawesi macaque radiation is likely to have occurred mostly or entirely during the Pleistocene and Holocene (Delson 1980, Goodman et al. 1998, Evans et al. 1999). Thus, much or all of macaque evolution on Sulawesi occurred after the coalescence of the microplates into the modern Sulawesi (Hall 2001) and vicariance/range fragmentation is expected to greatly outweigh geological history as the primary biogeographic factor affecting differentiation. Indeed, macaque

distributions appear to have faunal boundaries that are associated with Pleistocene geographic barriers, such as the isthmus of Gorontalo and the Tempe depression, neither of which are associated with a microplate boundary. Sulawesi tarsiers, on the other hand, are a much older radiation than Sulawesi macaques with roots in the Miocene, and their arrival to Sulawesi almost certainly predates the formation of Sulawesi in its present form. It is logical to predict, therefore, that Eastern tarsiers have a pattern of distribution that was first shaped by colonization of the proto-Sulawesi archipelago during the Miocene and Pliocene, which could have included sweepstakes dispersal and ancient vicariance events. Subsequent to the coalescence of the microplates into Sulawesi, tarsier distributions were then reshaped by Pleistocene vicariance events, with the tips of the tarsier branches bearing the effects of the forces that shaped the distributions of macaques and toads. This observation is the core of the hybrid biogeographic

Figure 9. (from Shekelle and Leksono 2004). Left: a biogeographic map of Sulawesi and surrounding islands based upon empirical data from the distribution of genetic variability in Macaca and Bufo (Evans et al. 2003), plus regions that lack endemic macaques and are presumably biogeographically distinct (MacKinnon and MacKinnon 1980). Center: a biogeographic map of the same area based upon the geological reconstructions of Hall (2001) concerning the tectonic activity of the microplates of the proto-Sulawesi archipelago. Right: a composite map of the distribution of tarsier acoustic forms layered on top of the right and center maps.

65

Shekelle, Morales, Niem itz, Ichwan & Melnick - Distribution of Tarsier Haplotypes

hypothesis for Sulawesi, which synthesizes empirical data from biology and geology, and explicitly acknowledges a time component that is a critical factor shaping biogeography in the region. Results from the hypothesis tests presented here show that the genetic data in this study cannot refute the hybrid biogeographic hypothesis and, indeed, the phylogenetic analyses are broadly consistent with it. The deepest evolutionary splits predicted by the hybrid biogeographic hypothesis receive strong support from the tarsier genetic data, while the more recent evolutionary splits are consistent with the genetic data and traces of phylogeographic structure are apparent in the consensus tree. The regions identified by MacKinnon and MacKinnon (1980) that have tarsiers, but which lack native macaque populations (i.e. the island chains of Banggai, Sangihe, Selayar, and Togian), can be logically argued to be more ancient than those areas that possess both taxa. Two such regions appear in this study, Sangihe and Togian, and in both cases, the tarsiers are identified as robustly supported monophyletic clades that are distantly related to other tarsiers of Sulawesi. The relatively recent evolutionary events, i.e. those within insular Sulawesi, are not perfectly supported by congruence of tarsier acoustic forms and monophyletic genetic clades, but that is not necessarily what would be predicted. Even so, it is interesting to speculate about phylogeographic structure in the consensus tree from the fine scale analysis. The Marantale population (T. dentatus), for instance, lies within the known hybridization zone between M. tonkeana and M. hecki (Bynum et al. 1997). The gene tree shows Marantale tarsier haplotypes to be polyphyletic, one haplotype clusters with Kamarora to the south, while the other clusters with Tinombo to the north, which are themselves nested within a northern clade. This pattern might possibly indicate that T. dentatus males occasionally hybridize with females from the Tinombo acoustic form. The two other populations that are polyphyletic in this tree, Sejoli and Ratatotok, also sit close to faunal boundaries identified in the hybrid biogeographic hypothesis.

66

In a study that was very similar to this one, Shaw (1993) examined biogeography and taxonomy in the Hawaiian cricket genus Laupala using acoustics and mtDNA. Shaw (2002) revisited conclusions in her 1993 study in light of nDNA and found that the mtDNA phylogenies had provided “extensively misleading” results, probably because of interspecific hybridization. Shaw (2002) cautioned against basing evolutionary interpretations among closely related species groups on mtDNA phylogenies, and found that nDNA provided results that were more consistent with other factors, such as acoustics, biogeography, and morphology. Indeed, for some time it has been known that lineage sorting and hybridization can produce data sets wherein taxa are not defined by monophyletic groups (e.g. Melnick and Hoelzer 1992, 1993). Several lines of evidence are broadly consistent with the hybrid biogeographic hypothesis including: the distributions of macaques and toads, the microplates that form the island of Sulawesi, the self-evident observation that the time of dispersal to Sulawesi is critical for biogeography, the tarsier acoustic data, and the tarsier genetic data. Some puzzling issues remain, however. For one, the arrival of macaques on Sulawesi is certainly recent compared to tarsiers, but there is less evidence that Bufo shares a similarly recent arrival to the region. Bufo is an ancient genus, and we can speculate that if, perhaps, Bufo arrived long before Macaca, then why should those two taxa share congruent distributions? Another puzzle is why faunal boundaries should remain congruent with microplate boundaries after millions of years. Is there, perhaps, some relationship between the underlying bedrock of the microplates and the ecology of the flora and fauna on the surface? Additionally it is worth mentioning that the pattern of dispersal and range fragmentation will vary among taxa, so the 15 biogeographic subregions predicted by tarsier acoustic forms should be considered a minimum estimate for Sulawesi. With additional taxa, the overall picture is likely to be more complex with more subregions.

Primates of The O riental Night

C) Summary Two phylogenetic analyses (i.e. broad scale analysis and fine scale analysis) used sequential approximation to wring as much phylogenetic information as possible from the available DNA sequence data. As expected, the first analysis offered decisive support for the monophyly of tarsiers. It also offered robust support for the monophyly of Eastern tarsiers in the data set. It could not convincingly resolve the Eastern-Western-Philippine tarsier trichotomy, however. This question may be resolvable, but will always suffer somewhat from lack of a suitable outgroup. Results of the broad scale analysis indicated that all non-tarsier taxa could be pruned from the data matrix for the fine scale analysis, but that both Western and Philippine tarsiers were required to root the analysis. Eastern tarsiers are not closely related to Western or Philippine tarsiers, and future fine scale analyses will benefit by identifying a more suitable outgroup from within the Eastern tarsiers. The fine scale analysis confirmed the monophyly of T. sangirensis and offered support for its basal position among the Eastern tarsiers in this data set. Subsequent pruning of the Philippine and Western tarsier and rooting the remaining data set with the T. sangirensis, however, left a data set with too few informative characters to be worthwhile. From this it was concluded that no further phylogenetic information could be wrung from the DNA sequence data, and conclusions about evolution among Eastern tarsiers were based on the analysis where Eastern tarsiers were rooted with Philippine and Western tarsiers. The most parsimonious tree from the fine scale analysis was used to reject biogeographic hypotheses based on the distributions of macaques and toads as well as the microplates that form the geological history of Sulawesi. A third hypothesis of congruence between tarsier acoustic groups and genetic groups could not be rejected. Differences between macaque distributions and tarsier distributions are not unexpected given what we now know about the relative ages of these two radiations.

The hybrid biogeographic hypothesis for Sulawesi (Shekelle and Leksono 2004), a comprehensive hypothesis which combines empirical biological and geologic data and explicitly considers the time of immigration to Sulawesi, was examined in light of tarsier acoustic data, which shows a remarkable fit with the aforementioned hypothesis. Tarsier genetic data in this study support predictions of that hypothesis. The implication is that each of the 15 tarsier acoustic forms thus surveyed is a distinct taxon, the validity of which can be examined more rigorously with additional DNA sequence data. More field surveys will very likely result in the discovery of more acoustic forms. ACKNOWLEDGEMENTS This material is based on work supported by the National Science Foundation under Grant No. INT 0107277 to MS. Additional financial support was provided to MS by the National Science Foundation Predoctoral Fellowship, Primate Conservation, Inc., Washington University Department of Anthropology, L.S.B. Leakey Foundation, Explorer’s Club, Wenner Gren Foundation, and the Margot Marsh Biodiversity Fund. MS wishes to acknowledge the support Jatna Supriatna and Noviar Andayani for sponsoring his work in Indonesia. Permits and other logistical support were supplied by The Indonesian Institute for Sciences (LIPI), The Department of Forestry, and Siti Prijono of the Museum Zoologicum Bogoriense. Preliminary genetic data that was not used in this study was collected in the labs of Allan Larson (Washington University), LIPI Pusat Bioteknologi (Center for Biotechnology, Indonesian Institute of Science, Cibinong), and Rajawali Hospital (Bandung). The genetics lab work was assisted by Tom Titus, Susan Jacobs, and Todd Jackman, and Luluk Lely Soraya Ichwan. Many people offered advice, comments, and/ or assistance on this project including Tab Rasmussen, Allan Larson, Bob Sussman, Jim Cheverud, Jane Phillips-Conroy, Erik Trinkaus, Mitchell Sommers, and Alan Templeton. Portions of this appeared in the dissertation of MS.

67

Shekelle, Morales, Niem itz, Ichwan & Melnick - Distribution of Tarsier Haplotypes

REFERENCES Adkins RM & RL. Honeycutt. 1994. Evolution of the primate cytochrome c oxidase subunit II gene. J. Mol. Evol. 38:215-231. Bynum, EL, DZ. Bynum & J. Supriatna. 1997. Confirmation and location of the hybrid zone between wild populations of Macaca tonkeana and Macaca hecki in Central Sulawesi, Indonesia. Amer. J. Primatol. 43:181-209. Delson, E. 1980. Fossil macaques, phyletic relationships and a scenario of deployment. In The Macaques: Studies in Ecology, Behavior and Evolution, Lindburg, DG. (ed) pp:10-30. New York: Van Nostrand Reinhold Company. Dijan, P & H, Green. 1991. Involucrin gene of tarsioids and other primates: alternatives in evolution of the segment of repeats. Proc. Natl. Acad. Sci. USA 88:5321-5325. Evans, BJ, JC. Morales, J. Supriatna & DJ. Melnick. 1999. Origin of the Sulawesi macaques (Cercopithecidae, Macaca) as inferred from a mitochondrial DNA phylogeny. Biol. J. Linn. Soc. 66:539-560. Evans, BJ, J. Supriatna, N. Andayani & DJ. Melnick. 2003. Monkeys and toads define areas of endemism on Sulawesi. Evol. 57(6):1436-1443. Goodman M, CA. Porter, Czelusniak J, Page SL, Schneider H, Shoshani J, Gunnell G, & Groves CP, 1998. Toward a phylogenetic classification of primates based on DNA evidence complemented by fossil evidence. Mol. Phylo. Evol. 9:585-598. Groves, CP. 1998. Systematics of tarsiers and lorises. Primates, 39:13-27. Hall, R. 1996. Reconstructing Cenozoic SE Asia. In. Tectonic Evolution of Southeast Asia. Hall R, Blundell D. (eds) Geol. Soc. Lond. Special Publication 106:153-184. Hall, R. 2001. Cenozoic reconstructions of SE Asia and the SW Pacific: changing patterns of land and sea. In Faunal and Floral Migrations and Evololution. SE Asia-Australia, Metcalf I, Smith J, Morwood M, Davidson I. (eds)

68

pp:35-56. Lisse: Swets and Zeitlinger Publishers. Hill, WCO. 1955. Primates: Comparative Anatomy and Taxonomy. II. Haplorhini: Tarsioidea. Edinburgh: Edinburgh University Press. MacKinnon, J & K. MacKinnon. 1980. The behavior of wild spectral tarsiers. Int.J. Primat. 1(4):361-379. McNiff, BE & MW. Allard. 1998. A test of Archonta monophyly and the phylogenetic utility of the mitochondrial gene 12s rRNA. Am. J. Phys. Anthropol. 107:225-241. Melnick, DJ & GA. Hoelzer. 1992. Differences in male and female macaque dispersal lead to contrasting distributions of nuclear and mitochondrial DNA variation. Int. J. Primatol. 13(4):379-393. Melnick, DJ & GA. Hoelzer. 1993. What is mtDNA good for in the study of primate evolution? Evol. Anthro 2(1):2-10. Meireles, CM, J. Czelusniak, SL. Page, DE. Wildman & M. Goodman. 2003. Phylogenetic position of tarsiers within the order Primates: evidence from g–globin DNA sequences. In Tarsiers: Past, Present, and Future. Wright PC, Simons EL, Gursky S. (eds) pp:145-160. New Brunswick: Rutgers UP. Mercer, JM, & VL. Roth 2003. The effects of Cenozoic global change on squirrel phylogeny. Science 299:1568-1572. Morales, JC, TR. Disotell & DJ. Melnick. 1999. Molecular phylogenetic studies of nunhuman primates. In The Nonhuman Primates. Dolhinow P, Fuentes A. (eds) pp:18-28. Mountain View CA: Mayfield Publishing Company. Morley, RJ. 1998. Palynological evidence for tertiary plant dispersals in the SE Asian region in relation to plate tectonics and climate. In Biogeography and Geological Evolution of SE Asia. Hall R, Holloway JD (eds) pp:211234. Leiden: Backhuys. Musser, GG & M. Dagosto. 1987 The identity of Tarsius pumilus, a pygmy species endemic

Primates of The O riental Night

to the montane mossy forests of Central Sulawesi. Amer. Mus. Novit. 2867:1-53. Niemitz, C, A. Nietsch , S. Warter & Y. Rumpler. 1991 Tarsius dianae: A new primate species from Central Sulawesi(Indonesia). Fol. Primatol. 56:105-116. Nietsch, A. 1999. Duet vocalizations among different populations of Sulawesi tarsiers. Int. J. Primatol. 20(4):567-583. Nietsch, A & C. Niemitz, C. 1993. Diversity of Sulawesi tarsiers. Deutsche Gesellschaft für Säugetierkunde 67:45-46. Nietsch, A & ML. Kopp . 1998. Role of vocalization in species differentiation of Sulawesi tarsiers. Fol. Primatol. 68(suppl.1):371-378. Paterson, HEH. 1985. The recognition concept of species. In Species and Speciation. Vrba ES. (ed) pp:21-29. Monograph No. 4. Pretoria: Transvaal Museum. Shaw, KL. 1993. The Evolution of Song Groups in the Hawaiian Cricket Genus Laupala. Doctoral Dissertation, Washington University, 1993. Shaw, KL. 2002. Conflict between nuclear and mitochondrial DNA phylogenies of a recent species radiation: what mtDNA reveals and conceals about modes of speciation in Hawaiian crickets. PNAS 99(25):16122-16127.

Shekelle, M. 2003. Taxonomy and Biogeography of Eastern Tarsiers. Doctoral thesis. Washington University, St. Louis. Shekelle, M, JC. Morales & DM. Melnick. 2001. Genetic and acoustic evolution among Eastern Tarsiers of Northern and Central Sulawesi. Presented at the International Society of Primatologists, 14th Congress, Adelaide, Australia. January 7-12, 2001. Shekelle, M & SM. Leksono. 2004. “Rencana Konservasi di Pulau Sulawesi: Dengan Menggunakan Tarsius Sebagai ‘Flagship Taxon’. Biota IX(1):1-10. Springer, MS & E. Douzery. 1996. Secondary structure and patterns of evolution among mammalian mitochondrial 12s rRNA molecules. J. Mol. Biol. 43:357-373. Sussman, RW. 1999. Primate Ecology and Social Structure, Volume 1: Lorises, Lemurs, and Tarsiers. Needham Heights (MA): Pearson Custom Publishing. Yoder, A. 2003. The phylogenetic position of genus Tarsius: whose side are you on? In Tarsiers: Past, Present, Future. Wright P, Simons E, Gursky S (eds) pp:161-175. New Brunswick, New Jersey: Rutgers UP.

69

Primates of The O riental Night

A METHOD FOR MULTIVARIATE ANALYSIS AND CLASSIFICATION OF TARSIER TAIL TUFTS Myron Shekelle1, Colin Groves2, Sharon Gursky3, Irene Neri-Arboleda, & Alexandra Nietsch4 1 Center for Biodiversity Studies and Conservation, Faculty of Mathematics and Science University of Indonesia, Depok 16421, Indonesia Email: [email protected] 2 School of Archaeology and Anthropology, Australian National University, Canberra, ACT 0200 Australia 3 Department of Anthropology, Texas A&M University, College Station TX 77843, USA 4 Freie Universität Berlin, Institut für Verhaltensbiologie, Haderslebener Str. 9, 12163 Berlin, Germany ABSTRACT Several independent studies of acoustics, DNA, and morphology support the hypothesis of numerous cryptic sibling taxa within Tarsius tarsier (=spectrum), but direct comparison among studies is often hindered by the relative incomparability of specimens (e.g. those that measure acoustic form versus those that measure crania). Tail tufts have been used for taxonomic identification both among and within tarsier species groups, but traditional univariate measures of tail tuft have yielded disappointing results. We used an ordinary ruler to assess 105 museum specimens and captive animals for overall tail length and five measurements of tail tuft fur length. We used a discriminant function analysis for 2 sets of analyses, one on all 105 tarsiers specimens grouped by tarsier species groups, and another on 37 Eastern tarsiers grouped by taxon. To accommodate for possible changes in tail length due to preservation method, separate analyses were conducted for both using: a) all variables, and b) all variable except tail length. In the analysis of all tarsiers grouped by species group with all variables, 96.2% of the original grouped cases were correctly classified, and 93.3% of cross-validated cases were correctly classified, while the respective values were 91.4% and 89.5% when the variable, tail length, was excluded. In the analysis of Eastern tarsiers grouped by taxon with all variables, 86.5% of the original cases were correctly classified, and 67.6% of the cross-validated cases correctly classified, while the respective values were 78.4% and 45.9% when the variable tail length was excluded. This method clearly has utility for diagnosing tarsier species groups, and it shows strong potential for assisting in the diagnosis of fine scale taxonomic variation within the Eastern tarsier species group, and perhaps within other tarsier species groups, as well. It is applicable for a great variety of specimen types, requires almost no cost and very little training, and takes very little time. We expect that this method will find utility for studies of both tarsier taxonomy as well as wildlife monitoring. We intend to extend our examinations of tarsier tail tufts. Keywords: Tarsius, Taxonomy, Wildlife Monitoring

INTRODUCTION There is a growing body of evidence that supports the hypothesis of numerous cryptic sibling species within Tarsius tarsier (=spectrum) as predicted by MacKinnon and MacKinnon (1980). In addition to MacKinnon and MacKinnon, results of several other field surveys show geographicallystructured variation in the duet call (Niemitz 1984, Niemitz et al. 1991, Nietsch and Niemitz 1993, Shekelle et al. 1997, Nietsch and Kopp 1998, Nietsch 1999, Nietsch and Babo 2001, Nietsch and Burton 2002, Shekelle 2003). Groves (1998, 2001, 2003) found that multivariate analysis of cranio-skeletal material in museum collections partially corroborates the classification based upon duet form. Shekelle (2003) found that DNA sequence data offered robust support

for the taxonomic separation of insular island populations from Sangihe and the Togian Islands of Malenge and Batudaka, but the bootstrap tree showed poor resolution among populations from Sulawesi proper. Shekelle et al. (this volume) found that the most parsimonious tree in Shekelle (2003) was not significantly shorter than a tree constrained by the hypothesis that forces monophyly based upon duet form and showed that the distribution of tarsier acoustic groups (=duet forms) formed a nearly oneto-one match with a ‘hybrid biogeographic hypothesis’ that layered prior hypotheses based upon empirical biological and geological data. Merker et al (2007) examined mtDNA and nDNA sequence information from a total of 139 specimens from a short transect connecting two putative tarsier species identified by acoustic data and found limited hybridization, with a

71

Shekelle, Groves, Gursky, Arboleda, & Nietsch - Classification of Tarsier Tail Tufts

pattern of genetic diversity consistent with the hypothesis that acoustic groups are separate taxa. Thus, several independent studies of morphology, genetics, and biogeography are all broadly consistent with the hypothesis that tarsier duet forms identify distinct taxa. One problem that has proven vexing for tarsier taxonomy and evolutionary biology is that specimens that can be scored for cranio-skeletal variation are not typically scorable for acoustic variation, and vice versa. A second set of problems is that current methods for surveying tarsier populations, principally analysis of DNA sequence data and acoustic surveys, are labor intensive, expensive, and require technological sophistication. For the purposes of conservation and monitoring, what is needed is an inexpensive, low-tech method that can be applied to living animals, both captive and wild, as well as deceased specimens. At a workshop in Jakarta in January 2003, several specialists in tarsier biology held a brainstorming session on this topic and the decision was made to try using a multivariate technique to glean taxonomic information from the tail tuft. Hill (1953, 1955) identified systematic differences in the tail tufts of tarsiers from the Philippines, Sundaland, and Sulawesi, and classified these animals as Tarsius syrichta, T. bancanus, and T. spectrum, respectively. Tarsier taxonomy is currently in a state of flux, and these species are referred to herein as Philippine, Western, and Eastern tarsiers, respectively (see Shekelle, this volume). More subtle differences have been noted in some tarsier populations, offering some hope that this feature can be used for fine-scale taxonomic identification. Meyer (1897) described a new species, Tarsius sangirensis, from Greater Sangihe Island and from Siau Island, both in the Sangihe Island chain between Sulawesi and Mindanao, based in part on reduced furriness of the tail tuft. Meyer (1897) further noted that a tarsier in the Dresden museum collected from Selayar Island, south of Sulawesi, also had a less extensive tail tuft. Groves (1998) found a second specimen of a Selayar tarsier to have the same condition and recommended taxonomic separation of the Selayar population. Shekelle et al. (1997) surveyed wild tarsier populations

72

on Greater Sangihe Island and also noted differences in the tail tuft, consistent with what Meyer had noted for T. sangirensis. Field surveys corroborated the observation by Meyer and Groves that the Selayar Island tarsier population has a less extensive tail tuft (M.S., personal observation), and three specimens from that expedition are included in this study. Thus, several authors have used tail tuft variation to show pronounced differences among species groups and subtle differences within the Eastern tarsiers using non-quantitative observations. Efforts to quantify these differences have had limited success using univariate statistics, such as “tail tuft length” (C.G. unpublished data). This is partly because, under close examination, very fine hairs are visible along the length of the tail, and determination of where the tail tuft begins is arbitrary and repeatability of this measurement is questionable. Additionally, univariate statistics fail to capture some of the subtle complexity of the tail tuft’s structure. Therefore, we elected to try a multivariate approach that was methodologically simple and inexpensive, but which would capture some aspects of tail tuft shape. METHODS The length of the tarsier tail was measured using a scaled ruler. One end of the ruler butted into the spot where the tail articulates with the body, and tail length was measured to the limit of the fleshy portion of the tail (i.e. the length of the fur on the end of the tail was not included in this measurement) (Figure 1). The shape of the tail tuft was estimated by measuring the length of fur at certain relative points along the tail. For consistency, the dorsal surface was arbitrarily chosen for measuring fur length. Quartiles were chosen for the original method, and the length of the fur was measured at 100%, 75%, 50%, and 25% of the length of the tail. For example, if a tarsier tail measured 200 mm, the length of the tail fur was measured at the extreme distal end, 150 mm from the base, 100 mm from the base, and 50 mm from the base. Preliminary results indicated that quartiles would not adequately capture the variation in those

Primates of The O riental Night

Helga Schulze © 2008

Figure 1: Diagram illustrating how to collect the measurements discussed in this paper.

populations with relatively small tail tufts, (i.e., T. bancanus and T. syrichta) and one additional measurement, 90%, was arbitrarily chosen. Thus, the measurements in this study include tail length (TAILLENG, or TL), and the length of tail fur at 100% (FULL), 90% (NINETY), 75% (SEVFIVE), 50% (FIFTY), and 25% (TWENFIVE) of total tail length. All measurements were estimated to the nearest millimeter. Fur that appeared to be less than about 0.5 mm was scored as 0.0 mm. Where fur length varied at a given point, efforts were made to measure the longest hairs (but not necessarily the longest hair, such as in cases when one, possibly aberrant, hair is much longer than other hairs around it). The search for the longest hairs at a given point was by no means exhaustive, and typically entailed several seconds of examination at most. All of the measurements on a given animal could be completed in no more than five or ten minutes, usually much less. The original data set was collected on tarsiers in the Museum Zoologicum Bogoriense (MZB), including several live specimens, by the entire group of authors. Subsequent data were collected by M.S. at the Field Museum of Natural History (FMNH), the Smithsonian (USNM), and the Raffles Museum of Biodiversity Research (ZRC). Data were analyzed using the discriminant function analysis in SPSS 11.0 for Macintosh. Prior probabilities were calculated using the ‘all groups

equal’ option. In each analysis, the ‘enter independents together’ option was used. Only adult specimens were included in the analyses reported here. These amounted to 105 specimens: 37 Western tarsiers, 31 Philippine tarsiers, and 37 Eastern tarsiers. We speculate that the tail length changes after an animal dies and is preserved. To examine the possibility of error introduced by this, separate analyses were conducted, both with and without using the variable tail length. In all analyses, the robustness of the results was tested using a jackknife approach with the ‘leave one out’ option in SPSS (referred to as “cross-validated” results). RESULTS The first set of two analyses consisted of 105 specimens grouped by species group, i.e. Philippine (n=31), Western (n=37), and Eastern (n=37), analyzed first with all variables (Analysis 1a) and secondly without the variable “tail length” (Analysis 1b). Graphical results of analyses 1a and 1b are presented in Figure 2, with function 1 plotted against function 2. The three species groups appear as two well-defined clusters, with Eastern tarsiers clearly separated from a second Philippine-Western cluster. Philippine and Western tarsiers are, themselves, distinct clusters with very slight overlap. In analysis 1a, 96.2% of the original grouped cases were correctly classified, and 93.3% of cross-validated cases were

73

Shekelle, Groves, Gursky, Arboleda, & Nietsch - Classification of Tarsier Tail Tufts

Figure 2: Results of `DF analysis of three tarsier species groups. Analyses 1a, with tail length, (above) Analyses1b, without tail length, (below)

correctly classified, while the respective values for analysis 1b were 91.4% and 89.5% (Table 1). A second set of two analyses was performed on the Eastern tarsiers in our data set. Cases were grouped by taxon as follows: those taxa of Eastern tarsiers recognized by Brandon-Jones et al. (2004), i.e. T. sangirensis (n=3), T. pelengensis (n=5), T. dentatus (=dianae) (n=6), T. pumilus (n=2). Other taxa included in the analysis included: T. sp1 (from

74

Selayar Island, see Groves 1998) (n=3), and T. lariang (i.e. the Palu acoustic form, Merker and Groves 2006) (n=3). Brandon-Jones et al. (2004) list numerous other populations that may be undescribed taxa that warrant further taxonomic research (see Shekelle and Leksono 2004), but these were all lumped into T. tarsier sensu lato (n=15), following Groves (2001). Graphical results of analyses 2a and 2b are presented in Figure 3, with function 1 plotted against function 2. In analysis 2a

Primates of The O riental Night

Table 1: Results of Analyses 1a (above) and 1b (below)

(including tail length), the seven taxa appear as four partially distinct clusters. One cluster on the left-hand side of the graph is formed by T. sangirensis and T. sp. (Selayar)—both insular forms isolated on small, remote islands. A second cluster, in the middle is T. pelengensis—another insular form, this one from Peleng Island. A third cluster, in the lower right, is composed of T. lariang, T. pumilus, and T. dentatus— all from the central core of Sulawesi. A fourth cluster,

T. tarsier sensu lato, spreads across the upper right corner—most of which come from the northern peninsula. In analysis 2b (excluding tail length), the results are similar except that the centroid of T. pumilus has moved away from the centroids of T. lariang, and T. dentatus, with commensurately clearer separation of T. pumilus. In analysis 2a, 86.5% of the original cases were correctly classified, and 67.6% of the crossvalidated cases correctly classified, while the

75

Shekelle, Groves, Gursky, Arboleda, & Nietsch - Classification of Tarsier Tail Tufts

respective values for analysis 2b were 78.4% and 45.9% (see Table 2). DISCUSSION The three tarsier species groups, Philippine, Western, and Eastern tarsiers, can easily be distinguished by this method. More than 95% of the cases were correctly classified (101 of 105). Indeed, two of the misclassified cases (on the upper left-hand side of the Eastern tarsier cluster, see figure 1a) were clearly well within the Eastern tarsier cluster, but happened to be slightly closer to the Philippine tarsier centroid. Furthermore, checking the provenance of those two misclassified cases, it turns out that they are themselves part of a three-case cluster of T. sangirensis (again, visible on the upper left-hand side of the Eastern tarsier cluster, see Figure 2, analysis 1a). Thus, those two cases were misclassified as a result of an over-simplified taxonomy that examines species groups, as opposed to species. If we ignore them, more than 98% of the cases are correctly classified (103 out of 105). This leaves two misclassified cases, one each of Western and Philippine tarsiers, which were misclassified as one another. Therefore, the amount of overlap between Eastern tarsiers and other tarsiers is effectively zero, while the overlap between Philippine and Western tarsiers is 2 out of 68. The cross-validated classification results found slightly less than 90% of the cases correctly classified (94 out of 105). In the cross-validated analysis, all three of the T. sangirensis cases occasionally cluster with Philippine tarsiers, but the Eastern tarsier cluster is, nevertheless, clear-cut (Figure 2, analysis 1b, the three T. sangirensis cases are visible in the lower left-hand side of the Eastern tarsier cluster). Ignoring these, 92.4% of cases were correctly classified, with 8 misclassified cases, five Western and 3 Philippine tarsiers, which were misclassified as each other. The cross-validated analysis is designed to estimate how likely this method will be to correctly classify other cases, not yet included in this data set, and the answer is that the amount of overlap between Eastern tarsiers and

76

other tarsiers is, again, zero, while the overlap between Philippine or Western tarsiers is 8 out of 68 (or 88.2% correctly classified). In other words, if an otherwise unskilled wildlife monitor were to confiscate a tarsier and use this method, we would expect that he or she would could correctly classify Western and Philippine tarsiers 88% of the time, while Eastern tarsiers could be correctly classified essentially 100% of the time— using multivariate analysis of tail tuft only. The applicability of this method for use within the Eastern tarsier species group is less clear cut. Two insular island populations, those from Sangihe and Selayar, both of which have reduced furriness of the tail tuft, are clearly separated from all other tarsiers, but overlap with each other. This is ironic because Sangihe Island is at the extreme northern end of the distribution of Eastern tarsiers, while Selayar Island is at the extreme southwestern end. Thus, the first working hypothesis is that the similarity is convergent, and does not indicate a close phylogenetic relationship. Future DNA studies will address that more definitively. A third insular island population, from Peleng, with a tail tuft that is intermediate in furriness between the mainland tarsiers and the Sangihe-Selayar cluster, is partially separated from all other tarsiers. Four out of five cases were correctly classified, while a fifth was misclassified as T. tarsier sensu lato. Interestingly, the T. pelengensis centroid appears to lie more-or-less halfway between the centroids of those populations from Sulawesi proper (e.g. T. lariang, T. pumilus, T. dentatus, and T. tarsier sensu lato) and the centroids of the other insular island populations (i.e. T. sp. (Selayar) and T. sangirensis). Once again, our first working hypothesis would not be to suspect a special phylogenetic relationship between the Peleng Island population, which lies at the extreme eastern end of the distribution of Eastern tarsiers, with the Sangihe (extreme north), Selayar (extreme southwest) populations. Our first best guess is that these populations are convergent, which in turn implies two hypotheses, a) chance or, b) directional evolution of tail tuft shape for insular island populations. An alternative hypothesis is that a lightly furred tail is a primitive retention, but this goes against the principle of global polarity, which predicts

Primates of The O riental Night

Notes: T. sp2 (Palu) = T. lariang, T. dianae is a junior synonym of T. dentatus, and T. tarsier sensu lato is a wastebasket of all specimens not classified as something other than T. tarsier and almost certainly contains cryptic taxonomic diversity.

Figure 3: Results of DF analysis within the Eastern tarsier species group. Analyses 1a, with tail length, (above), Analyses1b, without tail length, (below)

that the primitive condition is like other primates, i.e. relatively furry. We intend to investigate this curious phenomenon further. Fewer than 50% of the cases were correctly classified in the cross-validated analysis within the Eastern tarsier species group. This is unsatisfactory, but does not necessarily indicate that this method is

not practical for Eastern tarsier taxonomy. First, the sample sizes are low. and the number of putative taxa are many. The only group that exceeds the minimum sample size requirements of the statistical model is T. tarsier sensu lato, itself a wastebasket assemblage of everything that fit nowhere else. Even so, if we investigate the misclassified cases, patterns emerge.

77

Shekelle, Groves, Gursky, Arboleda, & Nietsch - Classification of Tarsier Tail Tufts

Table 2: Results of Analyses 2a (above) and 2b (below)

Notes: T. sp2 (Palu) = T. lariang, T. dianae is a junior synonym of T. dentatus, and T. tarsier sensu lato is a wastebasket of all specimens not classified as something other than T. tarsier and almost certainly contains cryptic taxonomic diversity.

Of the cluster composed of T. sp. (Selayar) and T. sangirensis, all six cases are correctly classified within the cluster. Thus, even with very low samples sizes, we would expect our hypothetical wildlife monitor to correctly classify 100% of confiscations of T. sp (Selayar) or T. sangirensis as having originated from one of those two islands. Likewise, with the cluster of populations from the central core of Sulawesi (T. lariang, T. pumilus, and T. dentatus), 10 out of 11

78

cases were correctly classified within the cluster, with a lone specimen of T. dentatus being classified together with the wastebasket T. tarsier sensu lato. Of the two remaining clusters, there is no clear pattern to the misclassified cases of T. tarsier sensu lato and T. pelengensis in the cross-validated analysis, and perhaps this should not be surprising. As mentioned, the former is a wastebasket taxon, while the latter appears to be morphologically midway between the

Primates of The O riental Night

very clearly isolated island populations of Sangihe and Selayar, on the one hand, and Sulawesi proper, on the other. In summary, we are encouraged by the results of this pilot study, the product of a brainstorming session by several experienced tarsier biologists. Our results show that this method clearly has utility for diagnosing tarsier species groups, and it shows strong potential for diagnosing finer scale taxonomic variation within the Eastern tarsier species group. It is applicable for a great variety of specimen types, including wild, captive, and deceased specimens. It requires almost no cost and very little training, and scoring an individual animal, whether alive or deceased, takes a matter of a few minutes at most. We expect that this method will find utility for studies of both tarsier taxonomy as well as wildlife monitoring. We intend to further our examinations. We plan to increase the data set by sampling other museum specimens, by adding this method to our field protocol, and by encouraging other tarsier biologists to take these measurements. We have begun a study to examine the utility of analyzing deciles versus quartiles, as well as identifying which measurements have the greatest influence on the classification. We intend to examine both intra and inter observer repeatability. The pattern of decreased furriness of the tail tuft among insular island populations opens a window for future studies of character evolution and biogeography that we intend to explore. ACKNOWLEDGEMENTS This material is based grants to MS from the National Science Foundation (No. INT 0107277), the Margot Marsh Biodiversity Foundation, and the Gibbon Foundation. For access to museum collections we thank Dr. Siti Prijono (MZB) Dr. Robert Martin and Michi Schulenberg (FMNH), Dr. Richard Thorington and Linda Gordon (USNM), Dr. Peter Ng and Kelvim Lim (ZRC), and Paula Jenkins in the NHM (London). We thank Willie Smits and the Schmutzer Primate Center, and Dr. SN. Prijono, Research Centre in Biology (LIPI) for sponsoring the workshop at which this work was conducted.

REFERENCES Brandon-Jones, D, AA. Eudey, T. Geissmann, CP. Groves, DJ. Melnick, JC. Morales, M. Shekelle & CB. Stewart. 2004. Asian Primate Classification. International Journal of Primatology. 25:97-164. Groves, CP. 1998. Systematics of tarsiers and lorises. Primates, 39:13-27. Groves, C. 2001. Primate Taxonomy. Washington D.C.: Smithsonian Institution Press. 350 p. Groves, C. 2003. The tarsiers of Sulawesi. In Tarsiers: Past, Present, and Future. Wright PC, Simons EL, Gursky S. (eds) pp:179-195. New Brunswick: Rutgers UP. Hill, WCO. 1953. Caudal cutaneous specializations in Tarsius. Proceedings of the Zoological Society of London 123:17-25. Hill, WCO. 1955. Primates: Comparative Anatomy and Taxonomy. II. Haplorhini: Tarsioidea. Edinburgh: Edinburgh University Press. MacKinnon, J. & K. MacKinnon K. 1980. The behavior of wild spectral tarsiers. International Journal of Primatology 1:361379. Merker, S, & CP. Groves. 2006. In Press. Tarsius lariang: A New Primate Species from Western Central Sulawesi. International Journal of Primatology 27: Merker, S, C. Driller, D. Perwitasari-Farajallah, & H. Zischler. 2007. Hybridisation in tarsiers. Prosimians 2007. Ithala Game Reserve, KwaZulu-Natal, 15-19 July 2007 Meyer, AB. 1897. Säugethiere vom Celebes- und Philippinen-Archipel, I. Abhandlungen und Berichte der Kaiserlich Zoologische und Anthropologische-Ethnologische Museum zu Dresden, 6:I-VIII, 1-36. Niemitz, C. 1984. Vocal communication of two tarsier species (Tarsius bancanus and Tarsius spectrum). In Biology of Tarsiers. Niemitz C. (ed) pp:129-142. New York: Gustav Fischer Verlag. Niemitz, C, A. Nietsch, S. Warter, & Y. Rumpler. 1991 Tarsius dianae: A new primate species from

79

Shekelle, Groves, Gursky, Arboleda, & Nietsch - Classification of Tarsier Tail Tufts

Central Sulawesi(Indonesia). Fol ia Primatologica 56:105-116. Nietsch, A. 1999. Duet vocalizations among different populations of Sulawesi tarsiers. Int. J. Primatol. 20:567-583. Nietsch, A, & C. Niemitz. 1993. Diversity of Sulawesi tarsiers. Deutsches Gesellschaft fur Saugetierkunde 67:45-46. Nietsch, A, & ML. Kopp. 1998. Role of vocalization in species differentiation of Sulawesi Tarsiers. Folia primatologica, 68(suppl.1):371-378. Nietsch, A, & N. Babo. 2001. The tarsiers of South Sulawesi. In Konservasi Satwa Primata. pp:114-119. Yogyakarta: Fakultas Kedokteran Hewan dan Fakultas Kehutanan Universitas Gajah Mada University - Yogyakarta. Nietsch, A, & J. Burton. 2002. Tarsier species in southwest and southeast Sulawesi. Abstracts, The XIXth Congress of the International Primatological Society (IPS), 49 Aug. 2002, Beijing, China: 20-21.

80

Shekelle, M. 2003. Taxonomy and Biogeography of Eastern Tarsiers. Doctoral thesis. Washington University, St. Louis. Shekelle , M, SM. Leksono, LLS. Ichwan, & Y. Masala. 1997. The natural history of the tarsiers of North and Central Sulawesi. Sulawesi Primate Newsletter, 4(2):4-11. Shekelle, M, & SM. Leksono. (2004) “Rencana Konservasi di Pulau Sulawesi: Dengan Menggunakan Tarsius Sebagai ‘Flagship Taxon’”. Biota IX(1):1-10. Shekelle, M, JC. Morales, C. Niemitz, LLS. Ichwan & DM. Melnick. (this volume). The distribution of tarsier mtDNA haplotypes for parts of north and central Sulawesi: a preliminary analysis. In Primates of the Oriental Night. Shekelle M, Maryanto I, Groves C,Schulze H, FitchSnyder H. (eds). (This volume). Indonesian Institute of Sciences, Bogor, Indonesia.

Primates of The O riental Night

Appendix 1: DF Statistics Table A-1: Results for Analysis 1a “All Variables, by Species Group”

81

Shekelle, Groves, Gursky, Arboleda, & Nietsch - Classification of Tarsier Tail Tufts

Table A-2: Results for Analysis 1b “All Variables except Tail Length, by Species Group”

82

Primates of The O riental Night

Table A-3: Results for Analysis 2a “Eastern Tarsiers Only: All Variables, by Taxon””

83

Shekelle, Groves, Gursky, Arboleda, & Nietsch - Classification of Tarsier Tail Tufts

Table A-4: Results for Analysis 2b “Eastern Tarsiers Only: All Variables except Tail Length, by Taxon””

84

Primates of The O riental Night

TARSIER LONGEVITY: DATA FROM A RECAPTURE IN THE WILD AND FROM CAPTIVE ANIMALS Myron Shekelle1 & Alexandra Nietsch2 1

Center for Biodiversity and Conservation Studies, Faculty of Mathematics and Natural Sciences, University of Indonesia, Gedung A Lantai 2 FMIPA, Kampus Baru, Depok 16424, Indonesia, Email: [email protected] 2 Freie Universität Berlin Institut für Verhaltensbiologie, Haderslebener Str. 9, 12163 Berlin Germany ABSTRACT

We report the longevity record for T. bancanus in captivity: born September 15, 1988 died April 25, 2006; current age 17 years, 7 months, 13 days, for T. syrichta in captivity (14 years, 80 days; minimum age ≈ 16 years), and for T. tarsier? in the wild (8 years and 3 months between captures; minimum age ≈ 10 years and 9 months). The first of these records is the longevity record for any tarsier in any condition. The last of these records is the longevity record for any wild tarsier. Ulmer (1960) predicted that tarsier longevity in the wild might exceed twenty years. Several tarsiers that survived 12-14 years in captivity exhibited behavior that was interpreted to indicate advanced age, but did not show clinical signs of old age. Keywords: Tarsius tarsier, T. spectrum, T. syrichta, T. bancanus, life-span, life history, longevity.

INTRODUCTION Ulmer (1960) reported that a captive T. syrichta died after 12 years 2 months in captivity— the maximum known longevity for a tarsier at the time— and speculated that tarsiers may live as long as 20 years in the wild based on the fact that post-mortem examination of the captive animal did not reveal evidence of advanced aging in the dentition, heart, kidneys, arteries, nor ovaries that would allow him to categorize the animal as being “old”. In recent years, several Philippine Tarsiers, Tarsius syrichta, have survived 10-14 years in captivity (Fitch-Snyder 2003). Such individuals are described as having the gestalt of old animals, although post-mortem examinations have not yet revealed conclusive evidence of advanced age in these animals (Mark Campbell, Cincinnati Zoo Veterinarian, personal communication). On the evening of June 30, 1999 a wild caught Philippine Tarsier died in the Cincinnati Zoo. The tarsier in question, a female named Tasaday (Philippine Tarsier Studbook record #1083), had been captured on the island of Leyte on April 11, 1985 (Fitch-Snyder 1994). This animal spent 14 years and 80 days in captivity. This surpassed the previous maximum known longevity for a tarsier. The animal was adult at the time of capture and estimated to have been born in 1983, making her about 16 years old when she died.

Surprisingly, even this is not the current longevity record for tarsiers in captivity. One T. bancanus borneanus in the Cleveland Zoo was born at the US National Zoo in Washington D.C. on September 15, 1988. She surpassed the longevity record in captivity on December 5th, 2002, and pershed on april 25, 2006, age 17 years, 7 months, 13 days. Niemitz (1979) trapped an adult male Tarsius bancanus borneanus in 1972 near Kuching (Sarawak, East Malaysia) that had previously been trapped by Fogden (1974) in 1965 and 1966. The animal was adult at the time of first capture, so Niemitz estimated the minimum age of this animal to be eight years. Prior to this report, that was greatest estimated minimum age for any wild tarsier, and indeed, the only published recapture of a wild tarsier of non-trivial age (i.e. greater than 1-2 years). METHODS Tarsiers were trapped in an isolated forest patch surrounded by alang-alang grassland near the village of Batuputih within the Tangkoko Dua Saudara Nature Reserve in the province of North Sulawesi, Indonesia between November 1987 and June 1988 (AN), and again between November 16, 1995 and June 1, 1996 (MS) (Nietsch 1993, Shekelle 2003). The site was resurveyed in July 1997, shortly after the

85

Shekelle & Nietsch - Tarsier Longevity: Data From a Recapture in The Wild and From Captive Animals

surrounding grasslands had burned. No tarsiers were found. Tarsiers were trapped in mist nests (AN, MS), or caught by hand during the daytime while sleeping (AN). Animals were given colored numbered leg bands and released. No sedatives were used nor warranted. Observations on trapped tarsiers include: body weight (Avinet Precision Spring Scale S300 for MS; Pesola Spring Scale for AN), skull length (Tajima Carbon Fiber Vernier Calipers for MS; Gneupel Vernier Calliper for AN), tail length (metal ruler in 1 mm graduates), characteristics of pelage and gross morphology, and hair samples for genetic analysis. Shekelle recorded the WGS84 geographic coordinates of the capture site with a Sony IPS-760 global positioning system. RESULTS On the evening of June 1, 1996, an adult female tarsier entered a mist net near Batuputih, on the edge of Tangkoko Nature Reserve, North Sulawesi (Indonesia) wearing an aluminum leg band numbered 0021. In her mouth she carried a suckling infant. Inquiries led to the discovery that this same tarsier had been trapped at 10 am on March 9, 1988. At the time of the initial capture, the female was trapped in association with an adult male and a subadult male. The female was determined to be sexually mature and the subadult male was assumed to be her offspring. A conservative estimate of her minimum age at the time of first capture was 2 years 6 months (17 months at first conception + 6 months gestation + 7 month old subadult offspring). Our best estimate of the animal’s minimum age upon recapture, therefore, is 10 years 9 months (2 years, 6 months upon initial capture + 8 years, 3 months between captures) (see below for discussion of estimating minimum age in wild tarsiers). When first captured, the orange tinge that indicates relatively youthful tarsiers was noted, while upon recapture her pelage did not offer any particular indications of age, having neither an orange tinge, nor gray around the face and head that might indicate relatively older animals (discussions of relative age based on pelage are found in Nietsch, 1993, and concur with unpublished data of Shekelle).

86

Her teeth did not show obvious wear. Dental wear in tarsiers appears to be minimal, with only one individual out of 101 wild caught Eastern tarsiers having dental wear that was obvious to the naked eye (unpublished data, Shekelle). Comparisons of the tarsier upon first capture and upon recapture are presented in Table 1. It is apparent that her weight did not change much—101 g to 104 g. Her tail may have grown slightly—231 mm to 242 mm—or there may be some interobserver measurement errors. Her head did not likely shrink, and the discrepancy—41 mm to 38 mm—is likely the result of interobserver measurement error. Shekelle found variation in this measurement to be low, particularly relative to i n tr a obser ver er r or, a n d a ba n don ed th i s measurement (2003, unpublished data). Upon first capture, the tarsier was in a simple family group— one adult male, one adult female, and one immature offspring. Upon recapture she was part of a more complex group that included one adult male, one adult female, two subadult females, and one infant male (probably born in late May 1996). Review of our notes has not yet determined the relationship between the trapping sites (i.e., was she trapped at the same site, or did she migrate?). Both trapping records show that the nest tree was a large strangler fig, but this is a common nest site for Eastern Tarsiers in primary forest. Upon first capture, all tarsiers in the group were found to be suffering from orange ectoparasites around the anogenital region and the base of the ears. This is a common condition among Eastern tarsiers, which upon closer examination has always been shown to be mites (see Merker 2003). Upon recapture, the tarsier was healthy and no parasitic infestations were noted. DISCUSSION Eight years and three months elapsed between captures of the wild tarsier that we report on, but we can use information about tarsier life history to better estimate the animal’s minimum age upon recapture. First, the animal was sexually reproductive

Primates of The O riental Night

Table 1. Recapture of an Adult Female Spectral Tarsier, Tarsius spectrum, from Tangkoko Nature Reserve.

1988 (Nietsch)

1996 (Shekelle)

09 March 1988

01 June 1996

104 g

101 g

tail

231 mm

242 mm

head

41 mm

38 mm

capture date weight

in social group with:

adult male (120 g) subadult male (100 g)

location

"Cathedral Tree"

condition upon capture

all had orange ectoparasites in the anogenital region and base of ears

other

orange tinge suggests young animal

at the time of initial capture. There has been speculation about the age that tarsiers reach sexual maturity. Roberts (1994) reported that one captiveborn female T. bancanus borneanus first gave birth at 922 days old, indicating she was sexually mature at 744 days (i.e. 2 years and 2 weeks), assuming a 178 day gestation period. Other estimates are based primarily upon observations of tarsiers in the wild, and may therefore be less rigorous. Niemitz (1977, 1979) reported sexual maturity for tarsiers from Tangkoko, T. tarsier? (=T. spectrum?) to be 518 days (i.e. a little more than 1 year 5 months), and for T. bancanus to be 11 months. Fogden (1974) stated that T. bancanus females may reproduce within one year. MacKinnon and MacKinnon (1980) stated that female tarsiers at Tangkoko with their parents until adulthood, and their observations showed that female offspring still lived with the parents in their second year. Thus, using Niemitz’s estimate of 17 months—the youngest report of sexual maturity for any tarsier in the T. tarsier species complex—we assume this animal’s minimum age at first conception was 17 months. We can add the age of the subadult male to her age, assuming he was her offspring. Estimating the age of a 100 gram subadult tarsier male is a difficult task because the animal’s weight is so close to the final adult weight—that is, the area where the growth

infant (37 g) adult male (117 g) subadult female (104 g) subadult female (110 g) (recaptured) Tangkoko 2

01 Jun 96 16 Nov 95 16 Nov 95 01 Jan 96 01 Jun 96 01°33.916' N 125°09.890' E

no problems no signs of advanced age

curve becomes nearly level—such that there must be very large error bars around the estimate. Nietsch (1993) argued that a 100 gram T. tarsiers (ie Tangkoko), male or female, will be at least one year old, which accords with MacKinnon and MacKinnon (1980) who reported that yearling tarsiers were still visibly smaller than their parents. On the other hand, Gursky (1997) provided formulas for tarsier body weight growth curves based upon data from captive T. bancanus (Roberts 1994), captive T. syrichta (Haring and Wright 1989), and her own data on 2 wild tarsier from Tangkoko (T. tarsier?) infants (1997)—curves that she stated do not differ statistically. The estimates based upon these formulas are 121 days, 206 days, 160 days, and 172 days, respectively. It is important to note, however, that none of these curves are designed (nor necessarily valid) for body weights as large as 100 grams. Furthermore, the curve for T. bancanus shows that by around 90 grams, the curve is clearly leveling off. Thus, these formulas are very likely to underestimate the age for a 100 g animal. Consequently, we conclude that 206 days (a little less than 7 months) is a safe estimate of the animal’s minimum age, and is even likely to be an underestimate of the subadult tarsier’s age, The gestation period in tarsiers is about 6 months. Izard et al. (1985) measured it as being 178

87

Shekelle & Nietsch - Tarsier Longevity: Data From a Recapture in The Wild and From Captive Animals

days for captive T. bancanus borneanus. Gurksy (1997) reported observations of wild female T. tarsier? from Tangkoko, North Sulawesi, that allowed her to estimate gestation length in two pregnancies as 195 days (+/- 3 days), and either 193 days or 182 days (+/ - 4 days) (uncertainty derives from two observed copulations separated by 9 days). Gursky estimated gestation to average 191 days in this species. Recently, the interbirth interval in T. tarsier, from Maros, South Sulawesi, was measured in captivity as a surprisingly short 172 days (June 29th to December 18th, 2002), which not only confirms the postpartum estrous noted by Roberts (1994) in captive T. bancanus borneanus, but also indicates a gestation period shorter than that found in other studies (MS, unpublished data). The animal in the present study

Figure 1: a photograph of the recaptured tarsier with her newborn infant.

88

was trapped and retrapped at Tangkoko and we use Gursky’s estimate of gestation length among Tangkoko tarsiers, 191 days on average, rounded down to six months to calculate minimum age. SUMMARY The longevity record for T. syrichta in captivity is 14 years and 80 days—a wild-caught animal with a minimum age of about 16 years old (FitchSnyder 2003). Several captive T. syrichta have survived beyond the previous longevity record of 12 years (Fitch-Snyder 2003, Ulmer 1960). The longevity record of T. bancanus in captivity is 17 years, 7 months and 13 days. Relatively few T. bancanus have been held in captivity compared to T. syrichta (Fitch-Snyder 2003). Far fewer Eastern tarsiers have ever been held in zoos or primate centers than either Western or Philippine tarsiers, and we know very little about that species’ longevity in captivity (Fitch Snyder 2003). Evidence from captive T. syrichta and T. bancanus shows that tarsiers behave like old animals by 14-16 years, although they do not show clinical signs of aging. There are too few data to know if longevity varies among tarsier species. We present data from a female tarsier that was re-trapped in the wild—first in March of 1988 and again in June of 1996. We estimate the minimum age of this animal as being 10 years and 9 months at the time of recapture. This is the current longevity record for any wild tarsier. Relative signs of advanced age, such as graying of the pelage around the face and head and noticeable dental wear, were not noticed in this animal (Figure 1). Thus, it may be that the tarsier in question was not particularly old, and that some tarsiers survive in the wild longer than this individual. Physical indicators of advanced age are very rare in both captive and wild tarsiers. While there is no direct evidence that any tarsier has lived longer than about 16 years, neither is there evidence to refute the speculation by Ulmer that a tarsier’s natural life span may exceed 20 years.

Primates of The O riental Night

ACKNOWLEDGEMENTS This material is based on work supported by the National Science Foundation under Grant No. INT 0107277 to MS. Additional financial support was provided to MS by: National Science Foundation (Protectoral Fellowship), Primate Conservation, Inc., Washington University Department of Anthropology, L.S.B. Leakey Foundation, Explorer’s Club, Wenner Gren Foundation, Garuda Indonesia Airlines. Alexandra Nietsch is grateful to the many colleagues and friends who provided help with the fieldwork, and to the Indonesian Institute of Sciences (LIPI) for scientific research permit. The study was supported by the German Research Foundation (DFG). Myron Shekelle wishes to thank Dr. Jatna Supriatna, Dr. Noviar Andayani, Suroso Mukti Leksono, and Yunus Masala. Permits and logistical support were arranged for by LIPI; Department of Forestry and KSDA in Manado. K. A. I. Nekaris read a draft of this manuscript and made constructive comments. REFERENCES Fitch-Snyder H. 1994. Asian Prosimian: North American Regional Studbook. Zoological Society of San Diego. San Diego. Fitch-Snyder H. 2003. History of Captive Tarsier Conservation. In Tarsiers: Past, Present, and Future. Wright PC, Simons EL, Gursky S. (eds) pp:277-295. New Brunswick: Rutgers UP. Fogden, MPL. 1974. A Preliminary Field-Study of the Western Tarsier, Tarsius bancanus Horsfield. In Prosimian Biology. Martin, RD, GA. Doyle, AC. Walker(eds) pp:151-165. Gursky, S. 1997. Modeling Maternal Time Budgets: The Impact of Lactation and Infant Transport on the Time Budget of the Spectral Tarsier,

Tarsius spectrum. PhD. Thesis. SUNY, Stony Brook, New York. Haring, D & P. Wright. 1989. Hand-Raising an Infant Tarsier, Tarsius syrichta.. Zoo Biology 8:265274. Izard, MK, PC. Wright & EL. Simons. 1985. Gestation length in Tarsius bancanus. Amer. J. Primatol. 9:327-331. MacKinnon J & K. MacKinnon. 1980. The behavior of wild spectral tarsiers. Int. J . Primatol 1:361-379. Merker, S. 2003. Vom Aussterben bedroht oder anpassungsfaehig? - Der Koboldmaki Tarsius dianae in den Regenwaeldern Sulawesis. PhD. Thesis. University of Goettingen, Germany. Niemitz, C. 1979. Outline of the behavior of Tarsius bancanus. In The Study of Prosimian Behavior. Doyle GA, Martin RD (eds) pp:631-660. New York. Nietsch, A. 1993. Beiträge zur Biologie von Tarsius spectrum in Sulawesi - Zur Morphometrie, Entwicklung sowie zum Verhalten unter halbfreien und unter Freilandbedingungen, PhD. Thesis, Freie Universität Berlin 1993. Roberts, M. 1994. Growth, Development, and Parental Care Patterns in the Western Tarsiers, Tarsius bancanus, in captivity: Evidence for a Slow Life History and Non-Monogamous Mating Sysytem. Int. J. Primatol 25(1):1-28. Niemitz, C. 1977. Zur Funktionsmorphologie und Biometrie der Gattung Tarsius Storr, 1780 (Mammalia, Primates, Tarsiidae). Cour Forsch Inst Senckenberg 25:1-160. Shekelle, M. 2003. Taxonomy and Biogeography of Eastern Tarsiers. PhD. Thesis. Washington University, St. Louis. Ulmer , F 1960. A Longevity record for the Mindanao Tarsier. J. Mamm. 1960:41:512.

89

Primates of The O riental Night

EASTERN TARSIERS IN CAPTIVITY, PART I: ENCLOSURE AND ENRICHMENT Keely Severn2 ,Donatus Dahang1& Myron Shekelle3 1

Oxford Brookes University, School of Social Sciences and Law, Department of Anthropology Oxford, OX3 0BP, email: [email protected] 2 Department of Biology, Faculty of Mathematics and Natural Sciences, University of Indonesia Gedung A Lantai 2 FMIPA, Kampus Baru, Depok 16424, Indonesia 3 Center for Biodiversity and Conservation Studies, Faculty of Mathematics and Natural Sciences, University of Indonesia, Gedung A Lantai 2 FMIPA, Kampus Baru, Depok 16424, Indonesia. email: [email protected] ABSTRACT Tarsiers have never formed successful breeding colonies in captivity, and the survival of tarsiers is presently dependent on in situ conservation. Five tarsiers were used to found a colony at a facility on the outskirts of Jakarta in October 2001. Four of these survive. One animal died after 14 months in captivity. The colony includes one mated pair that has produced 4 offspring, the most recent, born on July 25 th, 2004, survives. The others died after 32 days, 2 days, and seven months, respectively. These results indicate to us that our cage design is effective for the captive maintenance of adult animals. We are hopeful that we will soon have captive born animals that survive and reproduce. Tarsier populations are threatened by habitat loss, and there will be a persistent threat that this will lead to the extinction of some tarsier taxa, and we will be powerless to stop it until such time that ex situ tarsier conservation methods are developed. Keywords: Tarsius tarsier, Tarsius spectrum, captive environment, enclosure, enrichment, ex situ, conservation.

INTRODUCTION We use the term Eastern tarsiers for tarsiers from Sulawesi and surrounding islands. Hill (1955) classified these animals as Tarsius spectrum and accepted several subspecies. The current trend is to recognize that these animals are probably a cluster of related taxa with a confused taxonomic history (Groves 2001, Brandon-Jones et al. 2004). Groves et al (This volume) recognized T. tarsier as a senior subjective synonym of T. spectrum, both with a type locality of Makassar, a large city in southern Sulawesi from which there are no known existing museum specimens, no field surveys, and no known tarsier populations (see Shekelle 2003, Brandon Jones et al. 2004). Fitch-Snyder (2003) reviewed the history of tarsiers in captivity, an endeavour so fraught with failure that she commented that her work was a ‘documentation of the extinction of captive tarsier populations in North America and Europe since the first known import in 1850’. Thus, any published reports of tarsier husbandry need to be treated with the caveat that the effort failed. These failed efforts

to maintain captive populations of tarsiers have been disproportionately weighted toward Philippine and Western tarsiers. Fitch-Snyder reported only ten Eastern tarsiers having been kept in captivity outside of their country of origin. Six of the captive tarsiers listed by Fitch-Snyder date from the 1990’s at the Night Safari in Singapore. The other four are much older, and, without further evidence, it is not likely that they were Eastern tarsiers, as the use of T. spectrum in older reports is ambiguous and rarely refer to Eastern tarsiers (see Shekelle 2003; Groves et al, this volume). In consideration of the record of failure in North American and Europe, Fitch-Snyder (2003) recommended that ex situ conservation be targeted primarily at keeping colonies within habitat countries The opportunity to pursue Fitch-Snyder’s recommendation arose when reference material for new tarsier taxa were needed. Rather than collect and sacrifice these animals, one of us (M.S.) applied for permits to trap and cage the necessary animals. We are unaware of any published data on enclosure and enrichment for Eastern tarsiers. Establishing husbandry methods has been a process of research,

91

Dahang, Severn & Shekelle - Eastern Tarsiers in Captivity, Part II: a Preliminary Assessment

intuition, and experimentation. Basic parameters, such as minimum cage dimensions can only be guessed at, such as by comparisons with related species. The following recommendations were used as guides in the design of our colony: “Effective tarsier husbandry requires the resolution of a wide variety of problems. Cage environment must be optimum for both the animals and their prey. Cage furniture must be designed so that animals are able to effectively forage, rest, socialize, reproduce, sleep, and escape/hide from both caretakers and conspecifics. Observant and patient caretakers must be able to monitor each animal’s health individually and recognise and take action to correct potential problems. Tarsiers in captivity cannot survive without effective foraging skills, and weak or sick animals very quickly die. An understanding of the animal’s ecological niche is essential in designing adequate caging, particularly if behavioural research is to be carried out. The closer the captive habitat matches the animals’ natural habitat in terms of climatic conditions and provisions of choices for social interactions and foraging the more content the animals will be in their captive environment.” (Wright et al. 1989) Prior Efforts at Captive Tarsier Colonies Previous tarsier enclosures were of the following constructions. Wharton (1950) kept two Philippine Tarsiers for two months outside in a box measuring 0.61 by 1.6 by 0.46 m. Evans (1967) enclosed T. syrichta in mobile wooden cages with a mesh side measuring 1.3 by 0.8 by 0.8 m with a nest box measuring 0.45 by 0.1 by 0.1 m. Schreiber (1968) housed three T. syrichta in a steel box 0.58 by 0.86 by 0.68 m. Haring and Wright (1989) records the Duke University Primate Centre cage measuring 2 by 2 by 3 m containing a variety of bamboo and vine substrates of varying dimensions and angles. A 5-year study on a pair of T. bancanus was carried out in an enclosure measuring 5.1 by 3.6 by 4.5 m high. (Roberts et al. 1984; Roberts 1985; Roberts and Kohn 1993; Roberts 1994). The rooms had concrete floors, ceilings, and solid walls with a complex network of branches from floor to ceiling. Three types of nest boxes were provided to see if there were preferences. Wooden and fibreglass boxes 4 m high were never used by the tarsiers. Cardboard boxes with an open bottom with

92

vertical and horizontal bamboo inserted into the box were favoured. Current Captive Tarsier Populations Wright et al. (1987) observed that the ultimate measure of a healthy colony of nocturnal primates is a good rate of reproduction and low mortality (Wright et al. 1989). Unfortunately the three present captive populations of tarsiers described in ISIS (1998-2001) as follows have had no instances of breeding. Until recently, Cleveland Zoo held a single female Tarsius bancanus born in 1988. This animal was the lone tarsier in all of North America and Europe. Since 1992 she had been housed in a low light, understory tropical rainforest exhibit along with five greater mouse deer (Tragulus n. napu). The exhibit was approx. 260 square ft of floor space and 10 feet high (approximately 23.9 m2, 3 m high). Ground cover of fine bark mulch, artificial rock (shotcrete), live fine stemmed bamboo plantings and cut bamboo stalks not more than 1.5 inches (approximately 3.8 cm) diameter the full height of the exhibit (Don Kuenzer pers. comm.). This animal died on April 25, 2006 at more than 17 years of age, a longevity reccord for tarsiers (Shekelle & Nietsch this volume) Records from the Singapore Zoological Gardens Night Safari indicate that a male-female pair from North Sulawesi was donated by the Republic of Indonesia on 27 June 1996. Additionally, two confiscated males of unknown provenance that arrived on 2 July 1998 were both Eastern tarsiers. A lone male survives. The exhibit is an outdoor glass-fronted cage with wire mesh sides and ceiling measuring 10 by 5 by 8 feet high (approximately 3 x 1.5 x 2.4 m) . The furnishings are natural substrates, shrubs and leaf litter that provide a variety of criss-crossing vertical and angled supports, as well as leafy cover for camouflage (Sim Siang Huat pers. comm.). Multiple nest boxes are provided at about 2.5 m. There have been no records of captive birth, which is unusual, since records in Fitch-Snyder (2003) seem to indicate that births are fairly common, even though infant survival is low. One of us (MS) was asked to investigate the cause of this and, in 2003, the three

Primates of The O riental Night

surviving tarsiers at that time were all found to be male. Currently, a lone male survives. The only other known population of tarsiers outside of habitat countries is at the Ueno Zoological Gardens in Tokyo, Japan. ISIS (1998-2001) shows the presence of one male and two female Tarsius tarsier (=spectrum). We have no information on the enclosure or husbandry methods. The source of this colony is puzzling since zoo records indicate they are captive born, and yet there are no records of successful tarsier breeding colonies of this species in Indonesia or elsewhere, nor could we locate records at the Indonesian Department of Forestry export permits were issued for this species. All zoos and research institutions outside habitat countries are urged to rigorously verify the source of any tarsier available for import. METHODS Our tarsiers are housed at Biological Research Centre-Division of Zoology of the Indonesian Institute of Science at Cibinong, West Java, Republic of Indonesia; a colony founded in October 2001. They are housed in a building that contains twelve identical enclosures arranged in two rows of six with a large hallway for keepers between them. The enclosures are constructed of wire mesh on three sides with a cement floor and back wall. The ceiling is wire mesh, which in turn is covered by a permanent roof with a skylight. The entire enclosure is protected by rain, and natural sunlight falls only on the forward most portion of the cage. The colony in question was housed in outdoor enclosures with a natural light cycle. The enclosures measure 2 by 3.5 by 3 m high. Ventilation, lighting and humidity are all at natural levels for lowland West Java, which varies only slightly from the tarsiers’ capture localities in Sulawesi. There is a plywood nest box measuring 0.3 by 0.2 by 0.25 m mounted at 2 m on the back wall. Water and food are provided in plastic trays. The water tray measured 330 by 260 by 50 mm (see Dahang et al., this volume for information on feeding). Maintenance of the

enclosures included sweeping every 2 to 3 days to remove droppings and prey remains, and rinsing the enclosure with water. The following substrates were used: (i) 4 nutmeg trees planted in steel drums (0.38 m diameter, 0.5 m high); smallest tree 1.3 cm circumference, height 2.20 m; largest tree 2.1 cm, height 2.50 m, arranged in a square formation in the cage. (ii) Bamboo poles resting against the floor, walls, and steel drums angled from steel drums at various lengths and angles. The animals were housed in two groups. Each group had two enclosures (described above) connected by a small door that was left open. One group was trapped from the wild as a mated pair from Pattanuang, about 40 km northeast of Makassar, which we classified as T. tarsier. The second group was an unnatural association that began as a subadult male and an adult male from Selayar Island (T. sp, see Groves 1998, Nietsch and Babo 2001), together with a subadult male from Gimpu, Central Sulawesi (T. lariang, “Palu form” see MacKinnon and MacKinnon 1980, Niemitz 1984, Shekelle 2003, Merker and Groves 2006). By the conclusion of this study, the subadults had matured and the enclosure housed three adult males. Originally, the Gimpu animal was housed separately in a third enclosure, but this arrangement was abandoned after a few weeks when the colony manager sensed that the animal was stressed as a result of being caged alone. RESULTS Five tarsiers were delivered to this facility. Four infants were born. Four founders and one infant survive. On January 15, 2003, after about 14 months in captivity, the animal from Gimpu was discovered on the floor of the enclosure with traumatic wounds to the head and neck. The animal was treated intensively for three days, but was found dead on the morning of January 18, 2003. The cause of the injuries could not be determined from the necropsy, but intragroup aggression is a possibility.

93

Dahang, Severn & Shekelle - Eastern Tarsiers in Captivity, Part II: a Preliminary Assessment

The first infant was born on June 29, 2002 and died on July 30, 2002. Necropsy revealed massive trauma that could have come from a number of factors such as a fall, or from an attack by the adult male. About one week before the infant died, a major construction project began near the enclosure that seemed to cause the tarsiers stress. During the construction, the tarsiers were moved to a temporary facility. At the temporary facility the tarsiers were spared the stress of the construction project, but the enclosures were unsuitable for longterm housing. Nevertheless, the tarsiers were required to remain there until December 20, 2002. The second infant was born on December 18, 2002, and died the next day from injuries suffered during a fall. When the tarsiers returned from their temporary enclosures they were put in permanent enclosures that were nearly identical to their original permanent enclosures, except that the new enclosures were designed for parrots and the wire mesh was replaced with steel bars. A third infant, was born on 25 June 2003 in the tarsiers’ new permanent enclosure. Our prior experience prompted us to modify the enclosure. The cement floor was covered with dried leaves and grass to act as padding. The nest box was lowered to a height of one meter. The adult male was moved to an adjoining enclosure approximately two weeks before birth. The nest box was modified so that a hole in the back allowed the male to see the mother and infant in the nest box. Although Eastern Tarsiers live in family groups and males are known to provide some parenting (Gursky 1997), the precaution was taken to improve the chances of infant survival in the event that stress in a captive environment might lead the male to commit aggression where he would not in the wild. After slightly more than 7 months, the third infant was discovered dead with one leg missing. We suspect that a predator from outside the enclosure grabbed the tarsier while it was using the external wire mesh as a support. The tarsiers were moved again to a new facility, and double mesh was added to defend against predators. A fourth offspring was born on July 25th 2004, and this animal survives.

94

DISCUSSION Four founder tarsiers arrived at our facility in October, and a fifth in November of 2001. The animal that arrived in November died of massive trauma after about 14 months in captivity, possibly of intragroup aggression or by a predator that entered the enclosure. The four tarsiers that arrived in October have survived nearly 4 years in captivity. A mated pair have produced four offspring. The first three died after 32 days, 1 day, and 7 months, respectively. The fourth animal, born on July 25th, 2004, survives. Cage design was improved after each infant death to increase the chance of survival. Principal improvements were padding the floor to reduce injuries from falling, protecting the enclosure with a second mesh screen to prevent predation. These results indicate to us that our cage design is effective for the captive maintanance of adult animals. With steady improvement in the survival of offspring, we are hopeful that we will soon have captive born animals that survive and reproduce. Tarsiers have never formed successful breeding colonies in captivity (Fitch-Snyder 2003), and the survival of tarsiers is presently dependent on in situ conservation. This fact is troubling given habitat loss across the extent of occurrence of tarsiers, coupled taxonomic subdivision within the tarsier species groups. Habitat loss is proceeding virtually unchecked throughout much of tarsiers’ distribution, particularly in Indonesia where illegal logging is a massive problem. Thus, while total numbers of tarsiers may be very large for species groups (see MacKinnon 1986), individual populations within each species group could be under a high threat of local extinction. Given that the primary taxonomy of tarsiers is poorly understood and thought to be vastly underrepresented (Brandon Jones et al. 2004), it is likely that the extinction of local populations will, in some cases, lead to the extinction of tarsiers taxa. Unless a solution is found for the problem of habitat loss, or a successful ex situ tarsier conservation program can be developed, nothing can stop the processes that are currently operating that will very likely lead to the extinction of one or more tarsier taxa.

Primates of The O riental Night

ACKNOWLEDGEMENTS This material is based on work supported by the National Science Foundation under Grant No. INT 0107277 to MS and grants from the Margot Marsh Biodiversity Foundation and the Gibbon Foundation to MS. Sponsorship for MS in Indonesia was provided by the Center for Biodiversity and Conservation Studies, University of Indonesia (CBCSUI) and by the Indonesian Institute for Science (LIPI). Facilities for the tarsiers were provided by the Indonesian Institute for Science, Center for Biological Research—Division of Zoology (host institution of the Museum Zoologicum Bogoriense, or MZB). Permits for conducting research in conservation areas, for trapping tarsiers, transferring live tarsiers among provinces, and maintaining tarsiers in captivity were provided by the Indonesian Department of Forestry. Specieal thanks are due to Dr. SN. Prijono for initial support and encoragement in building this colony REFFERENCES Brandon-Jones D, AA. Eudey, T. Geissmann, CP Groves, DJ. Melnick, JC. Morales, M. Shekelle & CB. Stewart. 2004. Asian Primate Classification. International Journal of Primatology. 25(1):97-164. Evans, CS. 1967. Maintenance of the Philippine Tarsier (Tarsius syrichta) in a research colony. International Zoo Yearbook. 7: 201-202. Fitch-Snyder, H. 2003. History of Captive Tarsier Conservation. In Tarsiers: Past, Present, and Future. Wright PC, Simons EL, Gursky S. (eds) pp:277-295. New Brunswick: Rutgers UP. Groves, C. 2001. Primate Taxonomy. Washington D.C.: Smithsonian Institution Press. 350 p. Groves, CP. 1998. Systematics of tarsiers and lorises. Primates. 39:13-27. Gursky, S. 1997. Modeling maternal time budget; the impact of lactation and infant transport on the time budget of the Spectral tarsier, Tarsius spectrum. [dissertation]. Stony Brook (NY): State University of New York, Stony Brook.

Haring, DM, & PC. Wright. 1989. Hand-raising an infant tarsier, T. syrichta. Zoo Biology. 8(3): 265-274. Hill WCO. 1955. Primates: Comparative Anatomy and Taxonomy. II. Haplorhini Tarsioidea. Edinburgh: Edinburgh University Press. ISIS. SPARKS: Single Population Analysis and Record Keeping System. 1998-2001. Apple Valley, NM. MacKinnon, J. & K, MacKinnon. 1980. The b e h a viour of wild Spectral Tarsiers. International Journal of Primatology, 1:361-379 MacKinnon K. 1986. The conservation status of nonhuman primates in Indonesia. In Benirschke K (ed), Primates: The Road to Self-Sustaining Populations. SpringerVerlag, New York. Merker, S.& C. P. Groves. 2006. Tarsius lariang: A new primate species from western central Sulawesi. Int. J. Primatol. 27: 465–485. Niemitz, C. 1984. Vocal communication of two tarsier species (Tarsius bancanus and Tarsius spectrum). In C.Niemitz (ed.), Biology of Tarsiers, 129-141. Gustav Fischer Verlag, Stuttgart & New York. Nietsch, A & N. Babo. 2001. The tarsiers of South Sulawesi. In Konservasi Satwa Primata. pp:114-119. Yogyakarta: Fakultas Kedokteran Hewan danFakultas Kehutanan Universitas Gajah Mada University- Yogyakarta. Roberts M. 1985. The management and husbandry of the western tarsier (Tarsius bancanus) at the national zoological park. AAZPA 1985 Annual Proceedings 466-475. Roberts, M. 1994. Growth, development, and parental care in the western tarsier (Tarsius bancanus) in captivity – evidence for a slow life history and nonmonogamous mating system. International Journal of primatology. 15(1): 1-28. Roberts M & F. Kohn. 1993. Habitat use, foraging behaviour and activity patterns in reproducing western tarsiers T. bancanus in captivity: A management synthesis. Zoo Biology. 12(2): 217-232.

95

Dahang, Severn & Shekelle - Eastern Tarsiers in Captivity, Part II: a Preliminary Assessment

Roberts M, F. Kohn, A. Keppel, E. Malimak, & M. Deal. 1984. Management and husbandry of the western tarsier (Tarsius bancanus) at the ational zoological park. AAZPA 1984Annual Proceedings 588-600. Schreiber, GR. 1968. A note on keeping and breeding the Philippine tarsier (Tarsius syrichta) at Brookfield Zoo, Chicago. International Zoo Yearbook. 8: 114-115. Shekelle, M. 2003. Taxonomy and Biogeography of Eastern Tarsiers. Doctoral thesis. Washington University, St. Louis.

96

Wharton, CH. 1950. The tarsier in captivity. Journal of Mammology. 31: 260-268. Wright, PC, D. Haring, MK. Izard, & EL. Simons. 1989. Psychological well-being of nocturnal primates in captivity. In Segal E, editor, Housing care and psychological well-being of captive and laboratory primates, 61-74. Park Ridge, NJ: Noyes Publications. Wright PC, D. Haring D, Simons, & P. Andau. 1987. Tarsiers: a conservation perspective. Primate Conservation. 8: 51-54.

Primates of The O riental Night

EASTERN TARSIERS IN CAPTIVITY, PART II: A PRELIMINARY ASSESSMENT OF DIET Donatus Dahang1, Keely Severn2 & Myron Shekelle3 1

Department of Biology, Faculty of Mathematics and Natural Sciences, University of Indonesia Gedung A Lantai 2 FMIPA, Kampus Baru, Depok 16424, Republic of Indonesia 2 Oxford Brookes University, School of Social Sciences and Law, Department of Anthropology Oxford, OX3 0BP, Email: [email protected] 3 Center for Biodiversity and Conservation Studies, Faculty of Mathematics and Natural Sciences, University of Indonesia, Gedung A Lantai 2 FMIPA, Kampus Baru, Depok 16424, Republic of Indonesia [email protected] ABSTRACT Five Eastern tarsiers were kept in cages, three males in one cage and a mated pair (a pregnant female) in the other cage. We classify these animals as Tarsius tarsier, T. sp (“Selayar Form”), and T. lariang (“Palu Form”). The tarsiers were offered and consumed a variety of 22 species of commercially available and wild-caught insects and lizards. On average, 15.4 g of food were consumed per tarsier per day. Four species accounted for over 95% of the diet: commercial crickets, commercial mealworms, the large-bodied nocturnal grasshopper Caedicia major, and house geckos. The tarsiers’ favored food items in order of preference were: crickets, house geckos, grasshoppers, and mealworms. They appeared to have a negative preference for mealworms and the large-bodied diurnal grasshopper, Austricris guttulosa. The tarsiers’ preference for specific items may be inversely proportional to that item’s scarcity in the tarsiers’ diet. Keywords: Tarsius tarsier, T. spectrum, T. lariang, Diet, Body Weight.

INTRODUCTION We use the term Eastern tarsier for tarsiers from Sulawesi and surrounding islands. Until recently, these tarsiers were classified as two taxa, Tarsius spectrum and T. pumilus (Niemitz 1984a, Musser and Dagosto 1987). Groves et al (this volume) argued that T. tarsier is a senior synonym of T. spectrum, and has a type locality of Makassar (=Ujung Pandang), South Sulawesi (see Shekelle 2003, Brandon Jones et al. 2004). Several authors have offered evidence that Eastern Tarsiers are actually a constellation of related taxa (e.g. MacKinnon and MacKinnon 1980, Niemitz et al. 1991) including: T. sangirensis from Sangihe Island (Feiler 1990, Shekelle et al. 1997, Groves 1998), T. pelengensis from Peleng Island (Groves 2001), and T. dianae from areas of Central Sulawesi that flank the southern rim of Tomini Bay (Niemitz et al. 1991, Shekelle 2003, Brandon Jones et al. 2004). Our colony, which was founded in October 2001, has a mated pair of tarsiers from Pattanuang, near Maros, South Sulawesi, about 40 km northeast of Makassar, that are classified as T. tarsier. The

colony also has two tarsiers (one adult male, one subadult male at time of capture) from Selayar Island southeast of Makassar that are classified as T. sp1 “Selayar form” (see Groves 1998, Nietsch and Babo 2001). There is also a single subadult male from Gimpu (Central Sulawesi) that is classified as T. lariang, see MacKinnon and MacKinnon 1980, Niemitz 1984b, Shekelle 2003, Merker and Groves 2006), this animal having arrived in November 2001. The mated pair is housed in one enclosure, and the remaining tarsiers are housed in a second enclosure (Severn et al. this volume for more on enclosure design) Wild Eastern tarsiers are obligate faunivores, as are all other known tarsiers (Sussman 1999, Gursky 2002). The diet consists principally of insects, but is supplemented with virtually any wild prey item that the tarsier can catch and eat, including snakes, birds, bats, and others. Food preference has not been studied systematically in the wild. Fitch-Snyder (2003) records 10 Eastern tarsiers having been kept in captivity outside of Indonesia, six of which date from the 1990’s at the Night Safari in Singapore. The other four are

97

Dahang, Severn & Shekelle - Eastern Tarsiers in Captivity, Part II: a Preliminary Assessment

insufficiently documented, and it is uncertain whether they were in fact Eastern tarsiers. Prior to Hill (1955), T. spectrum was most often used for tarsiers other than Eastern tarsiers, and Eastern tarsiers were usually referred to something other than T. spectrum (e.g. T. fuscus). Thus, any reference to T. spectrum prior to 1955, for example, Clark (1924) and presumably Woollard (1925), typically referred to something other than an Eastern tarsier. In any event, we are unaware of any published data from any of the 10 animals mentioned by Fitch-Snyder. Both Philippine and Western tarsiers have been kept in captivity in North America and Europe (Fitch-Snyder 2003), but it is unclear whether dietary data from Philippine and Western Tarsiers would be for Eastern tarsiers. Both Musser and Dagosto (1987) and Groves (1998) view Philippine and Western tarsiers as forming a clade relative to Eastern Tarsiers. Groves (1998) went so far as to suggest that Eastern Tarsiers should be generically separated from the Philippine and Western tarsiers. To our knowledge there are no publications regarding the diet of captive Eastern tarsiers. We report our results regarding the diet of captive Eastern tarsiers after more than eight months in captivity. METHODS The tarsiers are housed at the Center for Biological Research—Division of Zoology of the Indonesian Institute of Science at Cibinong, West Java, Republic of Indonesia (the same administration that runs the Museum Zoologicum Bogoriense). Systematic data on diet were collected from March 11, 2002 until May 31, 2002. Systematic data on body weight were collected from the time of capture until control of the colony passed from that of MS to a LIPI staff scientist, Wirdateti in August 2004. For body weight we used an electronic scale with a one kilogram capacity measured in 1 gram increments (“Thinner” Measurement Specialties Electronic Chrome Kitchen Scale, Model MS-6845). Each enclosure is equipped with two feeding dishes. A converted ashtray used specifically for mealworms (10 cm X 10 cm X 4.5 cm) and a larger tray (45 cm X 23 cm X

98

8 cm) used both for feeding as well as weighing the tarsiers. Food items were placed in the tray. During weekly weighing, the tray was placed on top of the scale and zeroed after food items had been added. Cellophane tape was used on the inner walls of the tray to create a slick surface that inhibited insects from climbing out. Nevertheless, many food items were able to escape, but tarsiers are skilled hunters and the large majority of these were caught and eaten by the tarsiers. Observations suggested that the tarsiers first caught and ate those food items that could most easily escape, leaving the other food items in the tray to be eaten later. A grasshopper enclosure was used to store wild-caught grasshoppers for up to 24 hours (dimensions 17 cm x 13.5 cm x 18 cm). Following capture, grasshoppers were placed in the enclosure, identified and separated by species, and weighed before being offered to the tarsiers. The following food items were offered to the tarsiers: commercial crickets (Gyllus sp.), commercial mealworms (Tenebrio molitor), a commercially available bamboo worm (possibly Synochycha grandis), wildcaught house geckos (Gekko hemidactylus frenatus, Gekko hemidactylus turcicus), and many types of wildcaught grasshoppers including Tettigoniidae (Caedicia major, Zaprochilus australis, Conocephalus sp., Yorkiella picta, Phasmodes renatriformis), Acrididae (Austracris guttulosa, Bermiella acuta, Coryphistes ruricola, Acrida coneca), Gryllacrididae (Hadrogrylacres magnifica), Eumastacidae (Biroela sp., Keyacris scurra, Waramunga desertorum), Pyrgomorphidae (Atractomorpha crenaticeps, Desmoptera truncatipennis, Psednura sp.), and Mantidae (Creoboter spp.). No other nutritional supplements were provided. The total weight of all food items was measured in the feeding dish before it was re-zeroed. Unconsumed items, if any, were then weighed again to estimate the amount of food that was actually consumed. The overall weight of food items offered per day was adjusted ad hoc to minimize leftovers. Food was offered twice per day. The first feeding was at about 18:30 and the second feeding was at about 23:00.

Primates of The O riental Night

RESULTS In the 82 day measurement period, the tarsiers consumed 6329 g of food. This yielded an average intake of 77.2 g per day for five tarsiers, or 15.4 g of food per tarsier per day. Of the total food consumed, the

breakdown by food type by weight was: 48.4% crickets, 23.9% meal worms, 19% grasshoppers (Caedicia major), 4.2% house geckos, 2% grasshoppers (Austracris guttulosa), 0.4% grasshoppers (Bermeiella acuta), and 2.1% other grasshoppers.

Table 1. Weekly change in body weight prior to and during the feeding study. Bold italic = capture weight. Italic = first weighing after arrival at MZB. Asterisk (*) = not weighed. ET-105 and ET-106 lost appreciable body weight (26 g and 22 g, respectively) in the time between capture and transport.

ET-105

ET-106

ET-108

ET-109

ET-113

female

male

male

male

male

adult

adult

subadult

adult

subadult

18-Sep

113 g

133 g

*

*

*

28-Sep

*

*

75 g

103 g

*

2-Oct

87 g

111 g

*

*

*

3-Nov

116 g

128 g

98 g

108 g

*

10-Nov

116 g

128 g

98 g

108 g

*

12-Nov

*

*

*

*

85 g

19-Nov

116 g

129 g

98 g

110 g

102 g

25-Nov

118 g

129 g

100 g

110 g

102 g

1-Dec

121 g

131 g

105 g

113 g

*

8-Dec

122 g

132 g

107 g

115 g

100 g

15-Dec

124 g

134 g

107 g

117 g

103 g

22-Dec

125 g

134 g

107 g

*

104 g

29-Dec

127 g

135 g

108 g

119 g

106 g

5-Jan

127 g

134 g

109 g

120 g

106 g

12-Jan

128 g

135 g

111 g

121 g

108 g

19-Jan

131 g

136 g

113 g

122 g

111 g

26-Jan

131 g

136 g

113 g

121 g

112 g

9-Mar

134 g

137 g

*

122 g

*

17-Mar

136 g

138 g

*

119 g

*

24-Mar

138 g

138 g

*

124 g

*

30-Mar

141 g

139 g

*

126 g

114 g

8-Apr

140 g

139 g

*

124 g

*

15-Apr

143 g

138 g

112 g

123 g

117 g

23-Apr

142 g

138 g

113 g

128 g

121 g

29-Apr

142 g

139 g

*

128 g

*

8-May

143 g

139 g

117 g

131 g

*

21-May

146 g

136 g

114 g

*

117 g

30-May

148 g

136 g

*

*

*

99

Dahang, Severn & Shekelle - Eastern Tarsiers in Captivity, Part II: a Preliminary Assessment

Of the tarsiers in this study, ET-105 was a pregnant female that gave birth on June 29th. With a gestation period of approximately six months, ET-105 predictably gained weight throughout the study, starting at 134 g shortly before the study began and increasing to 148 g at the time the study concluded (Table 1). ET-106 was the adult male that was mated with ET-105. His weight stayed fairly constant throughout the study, fluctuating between 136 and 139 g. ET-108 was a subadult at the time of capture. This was the most difficult animal to weigh and often would not come down to the feeding dish. Prior to the study, ET-108 had been weighed at 113 g. Near the end of the conclusion of this study, he weighed 114 g. ET-109, an adult male, increased in body weight throughout the study, from 122 g to 131 g. ET-113 was a subadult male at the time of capture. This animal was also difficult to weigh. Body weight for this animal increased slightly from 112 g prior to the study to 117 g. DISCUSSION Nearly three-quarters of the tarsiers’ diet (72.3%) consisted of commercial crickets and mealworms. The rest was composed almost entirely of wild-caught food items. Of these, Caedicia major and house geckos accounted for 23.2% of the total diet. Therefore, these four food items accounted for over 95% of the diet. Sixteen different species of wildcaught food items plus commercially available insects known locally as ulat bambu (=bamboo worm) accounted for the remaining 4.5%. The data on diet cannot be broken down to amounts consumed by individual tarsiers because they were caged together. Body weight data from the tarsiers in this study are consistent with what we would expect from healthy, wild Eastern tarsiers (e.g. see Shekelle 2003 for a table of over 100 wild-caught Eastern tarsier body weights). Indeed, after nearly three years in captivity, four of the tarsiers in this study are still healthy, and the mated pair have an offspring born July 25th, 2004. Male ET-113 died on January 18th 2003, three days after showing signs of having been in a violent fight, either with another

100

tarsier in his cage, or with a predator that had entered his cage. Total daily consumption was probably affected by the fact that the adult female was pregnant during the study period. She gave birth on the 29th of June, 2002. The gestation period of all tarsiers for which these data are known is similar, around six months (reviewed in Sussman 1999). Gursky (1997) estimated the gestation period of Eastern tarsiers from Tangkoko (North Sulawesi) to be 190 days. This means that the tarsiers possibly conceived on December 21st, 2001, and the female tarsier was about 80 days pregnant when the study began and about 161 days pregnant when the study concluded. Roberts and Kohn (1993) reported that T. bancanus borneanus in their study self-selected a diet that was composed almost entirely of crickets. They reported the following average numbers of crickets consumed per day: adult male = 33.6, non-lactating females = 25.3, lactating females = 34, and juveniles (measured as <110 g in their study) = 33.2. Using their value for average cricket weight equals 0.368 g, we calculate the following average intakes in grams: adult male = 12.4, non-lactating females = 9.3, lactating females = 12.5, and juveniles (measured as <110 g in their study) = 12.2. Therefore, we calculate that had their colony had the same composition as ours, i.e. four adult males and one non-lactating female (by this study’s start date, March 11, 2002, all of our male tarsiers would have been considered as adult in the Roberts and Kohn study), they would have consumed 58.9 g of food per day [(12.4 x 4) + 9.3 = 58.9)], or 11.8 g per individual per day, on average. Compared with the value we found, 15.4 g, the tarsiers in Roberts and Kohn consumed only about 76.6% by weight the amount of food as the tarsiers in our study. This difference could be due to a several factors, e.g., metabolic differences among species. Western tarsiers are described as much less active than either Philippine or Eastern tarsiers, when they are caged side-by-side (Wright et al. 1989, MS personal observation). Izard et al. (1985) found that food intake, in grams, did not differ between non-pregnant animals and tarsiers during the first two trimesters of pregnancy. In the third trimester, food intake doubled.

Primates of The O riental Night

They cited Niemitz (1979) as estimating a daily intake of 10-12 g, a value they felt was consistent with the tarsiers they observed. In fact, Niemitz cited both Fulton (1939) and Ulmer (1963) who both estimated that tarsiers consume about 10% of their body weigh each day. What Niemitz wrote was, “in the present study it was not possible to determine exactly how much T. bancanus eat in grams per night, but the figure is probably approximately correct for this species in captivity as well” (Niemitz 1979). Wright et al. (1987) stated that each tarsier pair in their colony was offered approximately 12 oz. (~340 g) per day. This amount must include substantial excess and waste, since it exceeds 100% of the tarsiers’ body weight per day. The data on percentage of diet by food item cannot necessarily be construed as food preference, as diet composition was based largely upon availability. We can make some inferences about preference, however, based upon food items that were not consumed and the rank order of food item consumption. Crickets and mealworms made up most of the diet because they were commercially available and could be offered to the tarsiers in large quantities. Anecdotal observations indicate preference in this order: crickets, house geckos, grasshoppers, mealworms. Of the wild-caught grasshoppers, Caedicia major formed a much larger percentage of the tarsiers’ diet than the others, probably due to their much larger body size than most of the others and not because tarsiers necessarily prefer Caedicia major. Another large-bodied grasshopper, however, Austricris guttulosa, was frequently left unconsumed, and we infer a negative preference for this food item. It may be coincidence, but our observations indicate that A. guttulosa has a diurnal activity pattern, whereas tarsiers and most other food items offered to them were nocturnal. It was rare for any of the other wildcaught species to be left unconsumed by the tarsiers. Finally, it appeared to us that if any food items were given to the tarsiers routinely, to the exclusion of the other food items, their interest in the common item went down and interest in the other items went up. For this reason, we infer that tarsiers in our study may prefer to vary their diet. That is, although they

typically prefer crickets, if fed crickets exclusively for long enough, and then exposed to a different food, such as grasshoppers, in addition to crickets, their preference is for the new food item (i.e. in this example, grasshoppers). Therefore, it may be that among the items they commonly received, preference varied in inverse proportion to availability. Wright et al. (1987, 1989) fed T. bancanus borneanus and T. syrichta a similar diet of crickets and Anolis lizards (similar to the house geckos in our study) and remarked that the lizards were “relished primarily by T. syrichta, although occasionally T. bancanus will eat them” (Wright et al. 1987). They also offered mealworms and found that “only a few individuals” ate them. As in our study, their tarsiers’ diet was supplemented by opportunistically captured wild insects. For Wright et al.’s tarsiers, these supplements consisted of grasshoppers, dragonflies, cicadas, praying mantis, and scarabid beetles (Wright et al. 1987), as well as katydids (Wright et al. 1989). Similar to our study, Wright et al. (1989) commented that tarsiers were more enthusiastic about hunting and eating these rare items, which had only a “limited role” in the tarsiers’ diet (Wright et al. 1987). Roberts and Kohn (1993) reported that the T. bancanus borneanus in their study were fed a variety of live and prepared food, but their diet consisted almost entirely of crickets. Live Anolis lizards were seldom eaten. The same was true for Haitian cockroaches. Infant mice, both live and dead were completely ignored, as were mealworms. This observation may imply that the tarsiers in their study showed the reverse of the pattern seen in this study and that of Wright et al. (i.e. tarsiers studied by Roberts and Kohn preferred crickets exclusively and did not have a preference for rare food items). The lack of enthusiasm for lizards does not seem to agree with Izard et al. (1985), who studied gestation length in T. bancanus borneanus from the same location as the animals in Roberts and Kohn, and who described that large weight gain seen in the third trimester of pregnant females was due mainly to an increased consumption of lizards. Wright et al. (1987) and Roberts and Kohn (1993) both reported on the problems of a diet based

101

Dahang, Severn & Shekelle - Eastern Tarsiers in Captivity, Part II: a Preliminary Assessment

too heavily on crickets because of the low levels of some minerals, notably calcium, in crickets. To overcome this deficiency, Wright et al. (1987) purchased commercial crickets and then fed them a diet of “apples, crushed high protein monkey chow, Zeigler Brothers’ cricket diet and fresh water” before offering them to the tarsiers. Roberts and Kohn (1993) reported feeding their crickets “Cricket Diet 53-900000” from Zeigler Brothers, a granulated alfalfa pellet enriched with calcium, phosphorus, trace minerals, and Vitamin D, which was placed in the crickets’ holding bin as well as in the tarsier enclosure. Crickets raised on the mineral-supplemented diet as well as “presupplemented” crickets were placed in the tarsiers’ enclosures at the rate of about 6000 crickets per week. Wright et al. (1987) and Roberts and Kohn (1993) also commented on the problem that crickets tended to congregate on the floor, but that the tarsiers did not hunt animals on the floor, except for T. syrichta and then ‘only to retrieve a highly desirable prey item like a praying mantis” (Wright et al. 1987). Both reports described techniques to get the tarsiers’ prey items off of the floor and onto the various supports that filled the enclosures. In the wild, tarsiers commonly hunt prey on the forest floor, ambushing them from vertical perches about 0.5-1.0 m high. Our tarsiers have shown no such tendency to avoid the floor. There are potential sources of error in our study for example , some food items might have escaped the enclosure , but these would have been recorded as consumed. Similarly, since the enclosures are outdoors, the tarsiers might supplement their diet with wild prey that wander into the enclosure, and such supplements not have been recorded. Neither event was commonly witnessed, but each remains a potential source of error. Given that food intake was averaged over an 82 day period, we suspect the effects of these potential sources of error to be small. In summary, mean intake of food was 15.4 g per tarsier per day. Body weight data and observations of health status indicate that this diet was healthy and sufficient. Four food items accounted for over 95% of dietary intake: commercial crickets (48.4%), commercial mealworms (23.9%), the largebodied nocturnal grasshopper Caedicia major

102

(19.0%), and house geckos (4.2%). Unconsumed food items were rare with the exception of commercially available mealworms and wild-caught grasshoppers of the species Austricris guttulosa. In general, these Eastern Tarsiers prefer food items in the following order: crickets, house geckos, grasshoppers, mealworms. However, preference for a food item tended to increase when availability of that item was reduced. ACKNOWLEDGEMENTS This material is based on work supported by the National Science Foundation under Grant No. INT 0107277 to MS and grants from the Margot Marsh Biodiversity Foundation and the Gibbon Foundation to MS. Sponsorship for MS in Indonesia was provided by the Center for Biodiversity and Conservation Studies, University of Indonesia (CBCSUI) and by the Indonesian Institute for Science (LIPI). Facilities for the tarsiers were provided by the Indonesian Institute for Science, Center for Biological Research—Division of Zoology (host institution of the Museum Zoologicum Bogoriense, or MZB). special thanks Dr. SN. Prijono for permits for conducting research in conservation areas, for trapping tarsiers, transferring live tarsiers among provinces, and maintaining tarsiers in captivity were provided by the Indonesian Department of Forestry. Donatus Dahang’s participation was courtesy of a collaboration with Universitas Satria, Makassar. Helpful criticism came from Jim Sackett and Helena Fitch-Snyder who reviewed this manuscript. REFERENCES Brandon-Jones D, AA. Eudey, T. Geissmann, CP. Groves, DJ. Melnick , JC. Morales, M. Shekelle & CB. Stewart. 2004. Asian Primate Classification. Int. J. Primatol.. 25(1):97-164. Clark WEL. 1924. Notes on the living tarsier (Tarsius spectrum). Proc. Zool. Soc. London 217-223. Feiler, A. 1990. Ueber die Säugetiere der Sangiheund Talaud-Inseln der Beitrag A.B.Meyers für ihre Erforschung (Mammalia).

Primates of The O riental Night

Zoologische Abhandlungen der Staatlisches Museum für Tierkunde in Dresden, 46:75-94. Fitch-Snyder, H. 2003. History of Captive Tarsier Conservation. In Tarsiers: Past, Present, and Future. Wright PC, Simons EL, Gursky S. (eds) pp:277-295. New Brunswick: Rutgers UP. Fulton, JF. 1939. A trip to Bohol in quest of Tarsius. Yale J. Biol. Med. 11:561-573. Groves, C. 1998. Systematics of tarsiers and lorises. Primates 39:13-27. Groves, C. 2001. Primate Taxonomy. Washington D.C.: Smithsonian Institution Press. 350 p. Gursky, S. 1997. Modeling maternal time budget; the impact of lactation and infant transport on the time budget of the Spectral tarsier, Tarsius spectrum. [dissertation]. Stony Brook (NY): State University of New York, Stony Brook. Gursky, S. 2002. The Behavioral Ecology of the Spectral Tarsier. Evol. Anth. 11:226-234. Hill, WCO. 1955. Primates: Comparative Anatomy and Taxonomy. II. Haplorhini: Tarsioidea. Edinburgh: Edinburgh University Press. Izard, MK, PC. Wright & EL. Simons. 1985. Gestation length in Tarsius bancanus. American Journal of Primatology. 9: 327-331. MacKinnon, J & K. MacKinnon. 1980. The behavior of wild Spectral Tarsiers. Int.J. Primatol. 1:361-379. Merker, S. & CP. Groves. 2006. Tarsius lariang: A new primate species from western central Sulawesi. Int. J. Primatol. 27: 465–485. Musser, GG & M. Dagosto. 1987. The identity of Tarsius pumilus, a pygmy species endemic to the montane mossy forests of Central Sulawesi. American Museum Novitates, 2867:1-53. Niemitz, C. 1984a. Taxonomy and distribution of the genus Tarsius Storr, 1780. Pp 1-16 in C. Niemitz, ed., Biology of Tarsiers. Stuttgart: Gustav Fischer Verlag. Niemitz, C. 1984b. Vocal communication of two tarsier species (Tarsius bancanus and Tarsius spectrum). In C.Niemitz (ed.), Biology of

Tarsiers, 129-141. Gustav Fischer Verlag, Stuttgart & New York. Niemitz, C. 1979. Results of a field study on the Western tarsier (Tarsius bancanus borneanus Horsfeld, 1821) in Sarawak.” Sarawak Mus. J. 27:171-228. Niemitz, C, A. Nietsch , S. Warter & Y. Rumpler. 1991. Tarsius dianae: a new primate species from Central Sulawesi (Indonesia). Fol. Primatol. 56:105-116. Nietsch, A & N. Babo N. 2001. The tarsiers of South Sulawesi. In Konservasi Satwa Primata. pp:114-119. Yogyakarta: Fakultas Kedokteran Hewan dan Fakultas Kehutanan Universitas Gajah Mada University - Yogyakarta. Roberts M & F. Kohn. 1993. Habitat use, foraging behavior and activity patternsin reproducing Western tarsiers, Tarsius bancanus in captivity. Zoo. Biol. 12: 217-232. Shekelle, M. 2003. Taxonomy and Biogeography of Eastern Tarsiers. Doctoral thesis. Washington University, St. Louis. Shekelle, M, SM. Leksono, LLS. Ichwan & Y. Masala. 1997. The natural history of the tarsiers of North and Central Sulawesi. Sulawesi Primate Newsletter, 4, 2:4-11. Sussman, RW. 1999. Primate Ecology and Social Structure, Volume 1: Lorises, Lemurs, and Tarsiers. Needham Heights (MA): Pearson Custom Publishing. 283p. Ulmer, FA. 1963. Observations on the tarsier in captivity. Deutscher Zool Garten. 27: 106-121. Woollard, HH. 1925. The anatomy of Tarsius spectrum. Proc. Zool Soc. London. 70:10711184. Wright PC, D. Haring, MK. Izard & EL. Simons. 1989. Psychological well-being of nocturnal primates in captivity. In Segal E, editor, Housing care and psychological wellbeing of captive and laboratory primates, 61-74. Park Ridge, NJ: Noyes Publications. Wright PC, D. Haring, EL. Simons & P. Andau. 1987. Tarsiers: a conservation perspective. Primate Cons. 8: 51-54.

103

Primates of The O riental Night

THE CONSERVATION STATUS OF INDONESIA’S TARSIERS Sharon Gursky1, Myron Shekelle2, & Alexandra Nietsch3 11 Texas A&M University, Department of Anthropology, TAMU 4352,College Station TX 77843-4352 2 Center for Biodiversity and Conservation Studies (CBCS-UI), Faculty of Mathematics and Science (F-MIPA), University of Indonesia, Depok 16424, Republic of Indonesia. Email: [email protected] 3 Freie Universität Berlin, Institut für Verhaltensbiologie, Haderslebener Str. 9, 12163 Berlin, Germany ABSTRACT We present a method for making best guess estimates for the conservation status of 20 Indonesian tarsier taxa and populations listed in the taxonomy of Brandon-Jones et al. (2004). Published distribution maps were used to make rough estimates of the extent of occurrence. The accuracy of our estimates is sufficient that nearly all taxa and populations can be confidently assigned to one of the four size-based categories (i.e. 1 - 100 km², 100 - 5000 km², 5000 - 20,000 km², >20,000 km²) in the IUCN Red List guidelines. We used data and reports concerning habitat loss throughout Indonesia, and inferred commensurate range fragmentation, and declines in habitat quality and overall population numbers throughout the range of tarsiers in Indonesia. These inferences were supplemented with other information and personal observations where available. Based on our calculations, we make the following recommendations for the conservation status of Indonesia’s tarsiers. One taxon and one population warrant Critically Endangered (CR) status: Tarsius bancanus natunensis and T. sangirensis Siau population. Three taxa and four populations warrant Endangered (EN) status: Tarsius pelengensis, T. pumilus, T. sangirensis, and four acoustic forms of T. tarsier (i.e., the Tinombo form, the Togian form, the Selayar form and the Kabaena form). Indonesian tarsier species that we recommend be listed as Vulnerable (VU) include two taxa and three populations: Tarsius bancanus saltator, Tarsius dentatus (=dianae), and three acoustic forms of T. tarsier (=spectrum) (i.e., the Manado form, the Gorontalo form, and the Buton form). Indonesian tarsier species that we recommend be listed as Lower Risk include two taxa: Tarsius bancanus bancanus, and Tarsius bancanus borneaus. Indonesian tarsier species which we regard as data deficient include one taxon and three populations: Tarsius tarsier (i.e. T. tarsier, the Makassar form), and three other acoustic forms of T. tarsier (the Sejoli form, the Palu form, and the Kendari form). The Palu form has since been described as T. lariang. Several of these taxa / populations are distributed in regions that completely lack conservation areas of any kind. These include the two Critically Endangered forms (T. b. natunensis, T. sangirensis Siau population), four of the Endangered forms (Tarsius pelengensis, Tarsius sangirensis, Togian form, and Selayar form), and one Vulnerable form (Tarsius bancanus saltator). Keywords: Tarsius bancanus, T. dentatus, T. dianae, T. lariang, T. pelengensis, T. pumilus T. sangirensis, T. spectrum, T. syrichta, T. tarsier

INTRODUCTION Indonesia is one of the richest nations in the world, at least in terms of its biodiversity. Although it covers only 1.3% of the globe, the Indonesian Archipelago accounts for nearly 10% of the world’s remaining tropical forest, making it second only to Brazil in the amount of biodiversity it harbors (Cowlishaw & Dunbar 2000; Oates 1999). Unfortunately however, despite the country’s extensive system of protected areas, Indonesia’s forest cover has declined dramatically in the past decade (Jepson et al. 2001; Whitten et al. 2000), and is projected to decline still further (see Supriatna et al. 2001). Holmes (2002) reported that 20 million ha of Indonesia’s forests have been lost since 1989, at an

average annual deforestation rate of 1.7 million ha. Although 57 million ha of forest still remain on the three main islands of Sumatra, Borneo, and Sulawesi, less than 15% of this is lowland forest, which supports the highest biodiversity (MacKinnon 1997; Whitten et al. 2000, 2002). The dramatic loss of Indonesia’s forest cover is attributed to a variety of factors, including logging (legal and illegal), development of estate crops (primarily oil palm and pulpwood plantations), conversion to agriculture (by opportunistic settlers, refugees from ethnic conflict, and those arriving through Indonesia’s official transmigration program), and fires (natural and manmade) (Sunderlin 1999; Barber & Schweithelm 2000; Whitten et al. 2000; Robertson & van Schaik 2001; Holmes 2002). A new threat has emerged quite

105

Gursky, Shekelle & Nietsch - The Conservation Status of Indonesia’s Tarsiers

recently—tree felling to facilitate reconstruction efforts following the devastating tsunami of December 2004. The amount of forest loss attributable to each of these actions is highly contested and quite variable from island to island. One of the major goals of the Indonesian Prosimian Workshop (sponsored by Pusat Primata Schmutzer and the Museum Zoologicum Bogoriense, and organized by Myron Shekelle, Colin Groves, Helena Fitch-Snyder and Helga Shulze) was to revisit the results of the Indonesian Primate Conservation Assessment and Management Program (Indonesian Primate CAMP) (Supriatna et al. 2001) as regards tarsiers, and to re-evaluate the conservation status of Indonesia’s tarsier species in light of their newly revised taxonomy (Brandon-Jones et al. 2004). METHODS According to the IUCN Red List, a species conservation status can be determined based on a variety of different criteria (Cowlishaw & Dunbar 2000; Seal et al. 1994; Mace 1995). These include: population reduction, extent of occurrence, area of occupancy, limited population size and or quantitative analysis showing high probability of extinction (PVA and PVHA). Using various levels of these characteristics, a species can be classified as Critically Endangered (CR), Endangered (EN), Vulnerable (VU) or Lower Risk (LR). In evaluating the conservation status of Indonesia’s tarsiers, our efforts focused on estimating the extent of occurrence for a given taxon or population. Tarsiers are one of the more poorly known primate species, and we argue that the estimated extent of occurrence is the most easily quantifiable characteristic with which to begin assessing tarsier conservation status. Threats within a given region were assessed based upon observation and / or inference. These represent our best guess estimates, and include the following assumptions, given the well documented and extensive habitat loss in Indonesia over the past 10 years. First, observation and inference indicate that tarsier populations within Indonesia have become fragmented. Second, we infer

106

steep declines in the area of occupancy, habitat quality, and overall population numbers. Third, given that the social forces that are causing habitat loss appear to be uncontrollable at the moment, we infer continued declines in these variables. Thus, we argue that tarsiers throughout Indonesia are facing conservation threats severe enough to warrant concern, and that— lacking additional quantitative data—best guess conservation status estimates for tarsier species can be based on the estimated extent of occurrence supplemented where possible with additional evidence. Therefore, following IUCN Red List guidelines, and given that all other necessary conditions are met, populations whose extent of occurrence is between 1 – 100 km2 we recommend be given a conservation rating of Critically Endangered. Populations whose extent of occurrence is between 100 - 5,000 km2 we recommend be given a conservation rating of Endangered. Populations whose extent of occurrence is between 5,000 - 20,000 km2 we recommend be given a conservation rating of vulnerable. Populations whose extent of occurrence is greater than 20,000 km2 we recommend be given a conservation rating of low risk. We estimated extent of occurrence for each taxon and population in Brandon-Jones et al. (2004) using a map of Indonesia (scale: 30 cm = 200 km), distribution maps of tarsiers (Hill 1955; Niemitz 1984a; Musser & Dagosto 1987) and a map of tarsier acoustic form distributions complied from several sources (see Shekelle & Leksono 2004). We used a ruler and basic arithmetic to estimate length, breadth, and area for each distribution. The accuracy of our estimated extent of occurrence, although not high, was nevertheless usually sufficient to confidently classify each distribution within one of the four categories listed above (i.e. 1 - 100 km2, 100 - 5000 km2, 5000 20,000 km2, and >20,000 km2). RESULTS Tarsius bancanus bancanus was estimated to occur in an area almost equal to 100,000km2 (85,000 km2 on Sumatra and 13,400 km2 on Bangka). We

Primates of The O riental Night

recommend changing the conservation status from LRlc (Least Concern) to LRcd (Conservation Dependent). We maintain that given the massive habitat destruction throughout Indonesia, particularly in Sumatra, this species’ conservation status is clearly dependent upon future conservation of the remaining habitat. Tarsius bancanus borneanus was estimated to occur over the entire island of Borneo. It has a huge extent of occurrence. Thus, we recommend maintaining its assessment of LRlc (Least Concern). However, it should be kept under consideration for LRcd (Conservation Dependent) for the following reasons: One, Borneo is the site of some of the worst habitat destruction in Indonesia, including massive fires, land clearing, and logging throughout the Indonesian portions of the island. Second, as with many nocturnal taxa, there is the possibility that the primary taxonomy of the tarsiers on Borneo has been underrepresented. Western tarsiers have not been the focus of recent taxonomic investigations, as have the Eastern tarsiers of Sulawesi, and one might wonder if a landmass as large as Borneo might not hold more diversity than a single subspecies of a small mammal, like a tarsier. Indeed, the peculiar looking tarsier from the montane forests of Borneo described by Gorog & Sinaga (this volume), as well as the substantial behavioral differences between tarsiers studied Niemitz (1984b,c) in Sarawak and those studied by Crompton & Andau (1986, 1987) in Sabah could be interpreted as clues of unrecognized taxonomic variation. If this were the case, the estimated extent of occurrence would be a massive overestimate if T. b. borneanus were reclassified into multiple taxa, such has occurred in Sulawesi. Tarsius bancanus natunensis is listed by the IUCN as DD (Data Deficient). We estimate the extent of occurrence on Serasan to be a mere 90 km2. Thus, we recommend the conservation status for this taxon be changed Critically Endangered (CR B12c.). We note the threat of a continued decline in the extent of occurrence, area of occupancy,

and quality of habitat, particularly with regard to the development of the Natuna gas fields. There are reports of tarsiers on nearby Subi Island, with an estimated size of 180 km2. If confirmed, this would give T. b. natunensis a total extent of occurrence of 270 km 2, in which case the conservation status should be reduced to Endangered (ER B12c). Tarsius bancanus saltator is restricted to the island of Belitung. We estimated its extent of occurrence to be 5,625 km2. On this basis, we recommend changing its conservation status from Data Deficient (DD) to Vulnerable (VU B12bc). Note that our estimated extent of occurrence is near the 5000 km2 threshold, and this taxon might warrant consideration for Endangered status. Independent surveys by two experienced tarsier field biologists produced similar results, that is, potential tarsier habitat is clearly fragmented, and that tarsiers were only located in a small region on the center of the island (Suroso Mukti Leksono & Indra Yustian, personal communication). Thus, the inference from their surveys is the area of occupancy may be far smaller than the extent of occurrence. We project a continuing decline in the area of occupancy and quality of habitat. Tarsius dentatus has a very large estimated extent of occurrence (Brandon-Jones et al. 2004) that is greater than 20,000 km2. On this basis, it would be classified as Low Risk-Conservation Dependent or Near Threatened. However, we recommend that this species be listed as Vulnerable (VU B12ac). Each of us has extensive, first-hand experience in central Sulawesi, and it is difficult to imagine tarsiers there as being LR given the massive habitat destruction, ethnic conflict, and political upheaval throughout the province. Thus, although Tarsius dianae occurs in two large protected areas, Morowali and Lore Lindu National Parks, these are unlikely to provide effective refuge conservation that would shelter this animal from the threat of extinction. For example, Morowali is home to the Wana, traditional blowgun hunters who practice slashand-burn agriculture. This practice turns large

107

Gursky, Shekelle & Nietsch - The Conservation Status of Indonesia’s Tarsiers

swaths of the park into grass (alang-alang) which is primarily unused by tarsiers. Similarly, parts of Lore Lindu National Park were overrun by refugees, ostensibly fleeing the ethnic conflict in Poso although anecdotal reports indicate they were simply opportunistic illegal loggers, and entire villages were constructed within the park. Tarsius pelengensis has an estimated extent of occurrence of approximately 1925 km2. On this basis we recommend the Peleng tarsier be given a conservation status of Endangered (EN B1ab). Evidence indicates that the area of occupancy is likely to be substantially less than the extent of occurrence. Mochamad Indrawan (personal communication) conducted a survey on Peleng and estimated that less than 10% of the island is suitable for tarsiers. Thus, it may be necessary to raise the conservation status of this species to Critically Endangered. Tarsius pumilus has an estimated extent of occurrence of 4,112 km2, estimating montane forests within the region of the three known trapping localities of this species. We recommend a conservation status of Endangered (EN B12c). This conservation status is dependent on the projected disturbance to montane forest throughout central regions of Sulawesi. It is very clear that we need a study of the mountaintops throughout central and southern Sulawesi to confirm that pumilus has a distribution anywhere near as large as we have estimated here. Given that this species is known from only three specimens trapped between 1917 and 2000, an area of concern is the lingering question whether the mountain tarsier will be found on the mountaintops of all mountains in this area, or only a limited subset. Low population density is characteristic of many high altitude small mammal populations, and future studies could reveal that pattern will influence this species’ conservation status. Tarsius sangirensis is estimated to have an extent of occurrence of 576 km2 on Sangihe and 72 km2 on Siau. On this basis, we characterize the conservation status of this species as

108

Endangered (EN B12bc). One of us (MS) has first-hand experience on both of these islands, and the habitat that is appropriate for tarsiers is clearly fragmented. Bathymetric maps indicate that deep ocean separates Sangihe and Siau Islands, with no apparent means of gene exchange between the two populations. BrandonJones et al. (2004) listed the Siau population as a priority for future taxonomic research. If it were taxonomically separated from T. sangirensis, the Siau tarsiers would be candidates for the rating Critically Endangered (CR B12bc). In addition, we project a continuing decline in the area of occupancy and the area, extent and or quality of habitat, as well as the threat from volcanism. Tarsius tarsier (=T. spectrum), which has been called Tarsius spectrum since Hill (1955), is argued to be a subjective junior synonym of T. tarsier, both of which have the type locality of Makassar (see Shekelle 2003; Brandon-Jones et al. 2004). Tarsius tarsier itself is subdivided into numerous allopatric and parapatric populations that are identifiable by their vocalizations. There are no known museum specimens from Makassar, and Makassar today is a large city with no known tarsier populations. The acoustic form that is believed to be representative of this species comes from two closely located conservation areas, Bantimurung and Pattanuang, that were recently joined together into a single park, Bantimurung National Park that lies within 40 km NE of Makassar. From our limited distributional information, we cannot estimate the extent of occurrence of this population. The sympatric macaque, Macaca maura, however, is Endangered, so we will not be surprised if this population warrants EN Endangered status, as well. The “Buton form” (Nietsch & Burton 2002) is estimated to have an extent of occurrence of 5,776 km2, inferring that it includes the entire island. On this basis, we recommend that this form be classified with a conservation rating of Vulnerable (VU B12bc). The estimated extent of occurrence is near the 5000 km² threshold, and this population

Primates of The O riental Night

might warrant endangered status. Press reports identify Buton Island as a center of large-scale illegal logging that is clearing forests on the island. The “Gorontalo form” has been surveyed from Gorontalo, Panua, Libuo, and Tanjung Panjang (MacKinnon & MacKinnon 1980; Shekelle 2003). We infer that it occupies both coasts of the peninsula from Gorontalo to Tanjung Panjang, and has an estimated extent of occurrence of approximately 8100 km 2. On this basis we recommend that it receive a conservation status of Vulnerable (VU B12bc). As with other regions of Indonesia, this range of this species is experiencing tremendous habitat loss and fragmentation. The “Kabaena form” has an estimated extent of occurrence of 1350 km2, inferring that it includes the entire island. On this basis we recommend that it receive a conservation status of Endangered (EN B12bc). Little is known of this population, other than that tarsiers occur on this island and that their vocalizations are quite distinct (Nietsch & Burton 2002) We are unable to estimate the extent of occurrence of the “Kendari form” since the few sites from where it has been surveyed are in a relatively straight line (Nietsch & Burton 2002). We recommend that the Kendari form receive a conservation status of Data Deficient (DD). The “Manado form” has been surveyed from Tangkoko to Gorontalo (MacKinnon & MacKinnon 1980; Shekelle 2003). We infer that it occupies both coasts of the peninsula from Gorontalo to the northeastern tip of Sulawesi. We estimate the extent of occurrence to be approximately 10,951 km2. We recommend a conservation rating of Vulnerable (VU B12bc). The “Palu form” (MacKinnon & MacKinnon 1980) recontly described T. lariang (Merker and Groves 2006) is estimated to have a large distribution (Brandon-Jones et al. 2004). It is expected to share a parapatric distribution with T.dianae to the east. However, the known sites of the Palu form are all in a straight line so we cannot estimate

the extent of occurrence. Thus, we recommend that this form be given a conservation rating of Data Deficient (DD). The “Sejoli form” was surveyed by Shekelle in 1996 (Shekelle et al. 1997; Shekelle 2003). It is known from only one site, so it’s not feasible to estimate the extent of occurrence. Thus, we recommend this population be listed as Data Deficient (DD). The “Selayar form” (see Groves 1998; Nietsch & Babo 2001) has an estimated extent of occurrence of approximately 820 km2, inferring that it includes the entire island. On this basis we recommend conservation status of Endangered (EN B12bc). Selayar Island is the proposed site for a large oil refinery, and presently has no protected areas, so the conservation status of this tarsier population could be a candidate for Critically Endangered in the near future. The “Tinombo form” was surveyed by Shekelle in 1996 (Shekelle et al. 1997; Shekelle 2003) and again by Merker in 2001 (see Brandon-Jones et al. 2004). We infer that it occupies both coasts of the peninsula from Tinombo to the Ampibabo. From these surveys we estimate the extent of occurrence as approximately 3150 km2. On this basis we recommend conservation status of Endangered (EN B12bc). The “Togian form” tarsiers have been surveyed on both Batudaka and Malenge Islands (Nietsch & Niemitz 1993; Shekelle 2003). Assuming that they are distributed on all the islands in the chain except for the geologically unrelated volcanic island, Una, we estimate an extent of occurrence of 1980 km 2. On this basis we recommend conservation status of Endangered (EN B12bc). DISCUSSION In her recent discussion of future conservation action for tarsiers, Wright (2003) argued that “the first step in tarsier conservation is to change their ‘Data Deficient’ status. Within the next few years we should prioritize a survey for all tarsiers.” This is an admirable goal that we agree with, in principle, but one of us (MS) remarked ironically that the number of tarsier

109

Gursky, Shekelle & Nietsch - The Conservation Status of Indonesia’s Tarsiers

taxa is greater than the number of active tarsier field biologists (Shekelle 2005), such that the goal is unrealistic within the time frame of ‘a few years.’ There are three major types of data that we believe need to be obtained in order to more accurately determine the conservation status of the Indonesian tarsier populations (Table 1). First and most importantly, we believe that satellite maps capable of detailing the different types of habitat should be compared for 2003, 1998 and then 1993. This will provide quantifiable data on the change in available habitat over 10 years that will add rigor to our

assumption of habitat loss. Such maps could also provide more accurate information on the estimated area of occupancy, which we expect will probably be substantially less than the extent of occurrence. The use of satellite maps will also provide a quantifiable rate of habitat destruction over time. We will not be surprised, for example, if the rate of habitat destruction observed in satellite images over the last ten years turns out to be greater than the rates reported by the Indonesian Primate CAMP based on “visual estimates”: 1.87% estimated to be occurring in Sulawesi, 2.47% in Kalimantan and 2.5% in Sumatra.

Table 1. Conservation status recommendations for 20 Indonesian tarsier taxa and populations in Brandon-Jones et al. (2004) Taxon / Population

Tarsius bancanus bancanus T. b. borneanus T. b. natunensis T. b. saltator Tarsius dianae Tarsius pelengensis Tarsius pumilus

Tarsius sangirensis Siau population T. tarsier Buton Form Gorontolo Form

Location

Sumatra/Bangka Borneo Serasan Is. (Subi Is.) Belitung Is. Sulawesi: E. central Peleng Is. Sulawesi: central montane (1800-2200 m) Sangihe & Siau Is. (Sangihe only) Siau Is. Sulawesi: near Makassar Buton Island Sulawesi: N. peninsula, Gorontalo to Tanjung Panjang

Est. Extent of Occurrence (km2) 98,400 >100,000 90 (180) 5,625 >20,000 1,925 4,112

LR/lc DD DD

Conservation Status Indonesian Primate CAMP (Supriatna et al., 2001) LRlc LRlc VU (C2aii)*

DD LR/cd DD DD

EN (B2ab)* LRlc EN (B1ab) DD

VU (B12bc) VU (B12ac) EN (B12ab) EN (B12c)

Current (IUCN, 2004)

These recommendations LR (cd) LR CR (B12c)

648 (576) 72 ?

DD

EN (B1abc)

EN (B12bc)

not listed LR/nt

NL VU (B1ab)

CR (B12bc) DD

5,776 8,100

not listed not listed

VU (B1ab) LRnt

VU (B12bc) VU (B12bc)

1,350

not listed

not listed

EN (B12bc)

?

not listed

EN (B1ab)

DD

Kabaena Form

Kabaena Is.

Kendari Form

Sulawesi: near Kendari

Manado Form

Sulawesi: N. peninsula, Gorontalo to northeastern tip

10,951

not listed

LRlc

VU (B12bc)

Palu Form

Sulawesi: W. central

?

not listed

LRnt

DD

Sejoli Form

Sulawesi: N. peninsula, near Sejoli

?

not listed

LRnt

DD

Selayar Form

Selayar Is.

820

not listed

EN (B1ab)

EN (B12bc)

Tinombo Form

Sulawesi: N. peninsula, Tinombo to Ampibabo

3150

not listed

LRnt

EN (B12bc)

Togian Form

Togian Islands

1,980

not listed

EN (B1ab)

EN (B12bc)

110

Primates of The O riental Night

Second, we believe it is imperative that we obtain accurate and detailed data on population density and distribution for all tarsier species and subspecies. Although estimating the extent of occurrence (and assuming massive declines in habitat, habitat quality, and overall population throughout Indonesia) is a quantifiable method for making a best guess estimate of conservation status, it should be considered as nothing more than a handy shortcut compared with actual demographic data. For example, Gursky (1998) found a substantial decline in the density of tarsiers within Tangkoko Nature Reserve (Tarsius tarsier) from the time of the MacKinnons’ study in 1980 in comparison to her own study, 15 years later. Thus, even though the extent of occurrence may have remained the same, the population density for this species declined. At present we only have population density data for three populations of tarsier within Indonesia T. dentatus (Gursky 1998; Merker & Muhlenberg 2000), the “Manado form” (MacKinnon & MacKinnon 1980; Gursky 1998) and T. bancanus borneaus (Niemitz 1984b; Crompton & Andau 1986). It should be self-evident that the more criteria that are used for determining a species’ conservation status, the more accurate the estimate. Thus, in addition to using criteria B (extent of occurrence), using criteria A (quantified reduction in population) should be employed. Additionally, it is important that not only should we obtain data on density, but also distribution. Specific tarsier species / populations for which distribution data is lacking (and we were therefore unable to determine the extent of occurrence) include: the Kendari form, the Palu form, the Sejoli form, the Tinombo form, and Tarsius tarsier (the Makassar form). Thus, more field surveys are required. Additionally, GIS data could provide more accurate estimates of extent of occurrence than we used in this study. Third, we should perform PHVA (Population Habitat Viability Analyses) for each tarsier species. Although there are many assumptions made by PVHA (Cowlishaw & Dunbar 2000), the results of such studies will provide the conservation community with a quantifiable measure of the threat of extinction for each tarsier species. This analysis has already been

conducted for the Philippine tarsier (Neri-Arbodela 2001), but not yet for any Indonesian tarsier population. Additional Concerns While determining the conservation status of Indonesia’s nocturnal prosimians, such as tarsiers, is clearly important, actually conserving these primates is, of course, far more important. Specifically, it is critical that we identify the major threats to each of the tarsier species and identify ways to minimize these threats. Our discussions have highlighted that the major threat to all tarsier species is habitat destruction. For example, when Tangkoko Nature Reserve was set up in 1980 it contained approximately 8000 ha with an equivalent sized buffer zone surrounding the reserve (MacKinnon & MacKinnon 1980). By 1990, the buffer zone was completely destroyed and was replaced by coconut plantations (SG, MS, pers. obs.). The lushness of the coconut plantations fools many tourists concerning the actual threat of deforestation. In 1995, Yopie Muskita (personal communication) while working for WWF measured the boundaries of Tangkoko. He found that the boundaries of the reserve had been encroached upon substantially by neighboring communities, such that the reserve was less than 6000 ha. A few years later the Indonesian government downgraded the status of half of Tangkoko from kawasan konservasi (conservation area) to hutan wisata (recreation forest), a move that allowed greater legal exploitation of the reserve. The useful area of the reserve was reduced still further by forest fires, which the local people were unwilling to fight owing to a political struggle in which they saw their right to profit from ecotourism within the park being encroached upon by outside interests. (Unpblished, interviews with locals). It needs to be emphasized that, inspite of this , Tangkoko of this the more successful protection areas at limiting explorations and habitat destruction, when such activitiesare feasible. Unfortunately, there is no end in sight to this destruction. Additional lobbying of the government regarding its policies toward forest concessions and conservation needs to be conducted.

111

Gursky, Shekelle & Nietsch - The Conservation Status of Indonesia’s Tarsiers

Also, it is imperative that nature reserves be set up in areas where tarsiers exist, but there are no protected areas. Tarsier species / populations that presently lack any protected areas include: Tarsius bancanus natunensis, T. b. saltator, T. pelengensis, T. sangirensis, Togian form, and Selayar form. Of these six species / populations that do not occur within a protected area: one is Critically Endangered (Tarsius bancanus natunensis,), four are Endangered (Tarsius pelengensis, Tarsius sangirensis, Togian form, and Selayar form) and one is Vulnerable (Tarsius bancanus saltator), but near the threshold for Endangered. While the development of nature reserves is important, more efforts need to be expended toward enforcing these protected areas. At present we do not know of a single protected area (nature reserve or national park) that is not being exploited by the community—local regional, and / or national—in ways that are incompatible with the legislation for protected areas. These include: mines in Dumoga Bone (that use very primitive techniques with toxic runoff into the watershed), 5000 traditional slash and burn horticulturalists in Morowali, rattan harvesters and squatters in Lore Lindu and Tangkoko being downgraded to recreation forest from a protected area following years of extensive harvesting. It is our opinion that, in the short run, the health and vitality of the tarsier populations will be dependent upon actual enforcement of the rules governing protected areas in Indonesia, although the outlook for this is not good at present. At some point, we hope that public compliance with conservation laws will be the norm and actual enforcement will no longer be as critical. In the meantime, we recognize that these protected areas do not have the resources for patrolling and maintaining these areas. Following John Oates’ (1999) publication, “Myth and Reality in the Rain Forest,” we would like to suggest that a trust fund be set up for each nature reserve / national park within Indonesia. The dividends earned from these funds could then be used to pay guard’s salary and materials / equipment needed to protect the boundaries and contents of the reserve. Imagine if one million dollars were put into a trust fund for

112

Tangkoko. The dividends invested conservatively accruing a meager 5% annual would bring in $50,000 per year. That is more than sufficient to provide the resources to protect it. Efforts need to be made to minimize the pet trade of tarsiers. During this workshop, some participants went to the animal markets. Several tarsiers were for sale. They found that the cost for a single tarsier was approximately one million rupiah, or slightly less than $100. However, our own experience is that local people may try to sell tarsiers for as little as 5,000 rupiahs ($0.50). The mortality rate between the trapper and the market in Jakarta is thought to exceed 75%. Efforts to encourage confiscation of these animals at the markets needs to be made by the Indonesian police, not just rescue centers. Also, additional penalties to discourage the pet trade need to be developed. If all efforts to conserve tarsiers within Indonesia come up wanting in the short and medium term, then the only hope for some of the more threatened species is ex situ conservation. Although all previous efforts have failed (Fitch-Snyder 2003), reports by Severn at al. (this volume) and Dahang et al. (this volume) offer some promise that ex situ conservation of tarsiers will be feasible in the near future. Critical for this to happen will be either a sincere desire by Indonesian authrorities to work in cooperation with foreign centers of expertise and funding, or a concerted and determined effort to develop such a project on their own. In summary, tarsier species / populations vary greatly in their respective threat of extinction. Some small, insular populations, such as those on Serasan and Siau Islands are probably at imminent risk of extinction. What is most troubling is that, while we are able to estimate with some degree of precision those populations at greatest risk, we know of no solution in operation that is effectively reducing that risk. AKNOWLEDGEMENTS This material is based on work supported by the National Science Foundation under Grant No. INT 0107277, a grant from the Margot Marsh Biodiversity

Primates of The O riental Night

Foundation, and another by the Gibbon Foundation, all of which were awarded to MS. We thank Willie Smits and the Schmutzer Primate Center, and Dr. SN. Prijono, Research Centre in Biology (LIPI) for sponsoring the workshop at which this work was conducted. Instrumental in our project design was participation in the Indonesian Primate CAMP by MS. Matt Richardson reviewed this manuscript and offered valuable criticism. REFERENCES Barber CV, & J. Schweithelm. 2000. Trial by Fire. Washington DC: World Resources Institute. Brandon-Jones D, AA. Eudey, T. Geissmann, CP Groves, DJ. Melnick, JC. Morales, M. Shekelle & CB. Stewart. 2004. Asian primate classification. International Journal of Primatology. 25(1):97-164. Cowlishaw G & R. Dunbar. 2000. Primate Conservation Biology. Chicago: University of Chicago Press. Crompton R & P. Andau. 1986. Locomotion and habitat utilization in free ranging Tarsius bancanus: A preliminary report. Primates 27:337-355. Crompton R & P. Andau. 1987. Ranging, activity rhythms, and sociality in free-ranging Tarsius bancanus: A preliminary report. International Journal of Primatology 8: 43-71. Gorog AJ & MH. Sinaga. 2008. A tarsier capture in montane forest on Borneo. In Primates of the Oriental Night (ed) Shekelle, M, I. Maryanto, CP. Groves, H. Schulze, & H. Fitch-Snyder. (this volume). Indonesian Institute of Sciences & Indonesian biological Society Groves, CP. 1998. Systematics of tarsiers and lorises. Primates 39:13-27. Gursky, SL. 1998. The Conservation Status of Two Sulawesian Tarsier Species: Tarsius spectrum and Tarsius dianae. Primate Conservation 18:88-91.

Gursky, SL. 1998.The Conservation Status of the Spectral Tarsier, Tarsius spectrum, in Sulawesi Indonesia. Folia Primatologica 69:191-203. Hill WCO. 1955. Primates: Comparative Anatomy and Taxonomy. II. Haplorhini: Tarsioidea. Edinburgh: Edinburgh University Press. Holmes, DA. 2002. The predicted extinction of lowland forests in Indonesia. In Terrestrial ecoregions of the Indo-Pacific: a conservation assessment, Wickramanayake E, Dinerstein E, Loucks CJ, Olson DM , Morrison DM, Lamoreux J, McKnight M, Hedao P. (eds) pp:7-13. . Washington DC: Island Press. IUCN. 2004. 2004 IUCN Red List of Threatened Species. <www.redlist.org>. Downloaded on 20 August 2005. Jepson, P, JK. Jarvie, K. Mackinnen, & KA. Mon. 2001. The end for Indonesia’s lowland forests? Science 16:859 860. Mace, G. 1995. An investigation in methods for categorizing the conservation status of species. In Large Scale Ecology and Conservation Biology, Edwards PJ, May RM, Webb NR. (eds) pp:293-312. Oxford: Blackwell. MacKinnon, J. 1997. Protected area systems review of the Indo-Malayan realm. The Asian Bureau for Conservation Limited, Canterbury, United Kingdom. MacKinnon J & K. MacKinnon. 1980. The behavior of wild spectral tarsiers. International Journal of Primatology 1: 361-379. Merker, S & M. Muhlenberg. 2000. Traditional land use and Tarsiers - Human influences on population densities of Tarsius dianae. Folia Primatologica 71 (6): 426-428. Merker, S & CP. Groves. 2006. Tarsius lariang: A new primate species from western central Sulawesi. Int. J. Primatol. 27: 465–485. Musser GG & M. Dagosto. 1987 The identity of Tarsius pumilus, a pygmy species endemic to the montane mossy forests of Central Sulawesi. American Museum Novitates 2867:1-53.

113

Gursky, Shekelle & Nietsch - The Conservation Status of Indonesia’s Tarsiers

Neri-Arboleda I. 2001. Ecology and behavior of Tarsius syrichta in Bohol, Philippines: Implications for conservation. MSc Thesis Department of Applies and Molecular Ecology, Unviersity of Adelaide. Niemitz, C. 1984a. Taxonomy and distribution of the genus Tarsius Storr, 1780. In The Biology of Tarsiers. Niemitz C. (ed) pp:1-16. New York: Gustav Fischer Verlag. Niemitz, C. 1984b. An investigation and review of the territorial behaviour and social organization of the Genus Tarsius. In The Biology of Tarsiers. Niemitz C. (ed) pp:1-16. New York: Gustav Fischer Verlag. Niemitz, C. 1984c. Locomotion and posture of Tarsius bancanus. In The Biology of Tarsiers. Niemitz C. (ed) pp:1-16. New York: Gustav Fischer Verlag. Nietsch, A & C. Niemitz. 1993. Diversity of Sulawesi tarsiers. Deutsche Gesellschaft fur Saugetierkunde 67:45-46. Nietsch, A & N. Babo. 2001. The tarsiers of South Sulawesi. In Konservasi Satwa Primata. pp:114-119. Yogyakarta: Fakultas Kedokteran Hewan dan Fakultas Kehutanan Universitas Gajah Mada University - Yogyakarta. Nietsch, A & J. Burton. 2002. Tarsier species in southwest and southeast Sulawesi. Abstracts, The XIXth Congress of the International Primatological Society (IPS), 49 Aug. 2002, Beijing, China: 20-21. Oates, J. 1999. Myth and reality in the rain forest: How conservation strategies are failing in West Africa. California: University of California Press. Robertson, J & C. van Schaik. 2001. Causal factors underlying the decline of the orang-utan. Oryx 35(1): 26-38.

114

Seal US, TJ. Foose, & S. Ellis. 1994. Conservation Assessment and Management Plans (CAMPs) and Global Captive Action Plans (GCAPs). In Creative conservation, Olney PJ, Mace GM, Feistner A. (eds) pp312-325. London: Chapman and Hall. Shekelle, M. 2003. Taxonomy and Biogeography of Eastern Tarsiers. Ph.D. Thesis. Washington University. Shekelle, M. 2008. Distribution and biogeography of tarsiers. In Primates of the Oriental Night Shekelle M, Maryanto I, Groves C, Schulze H, Fitch-Snyder H. (eds). (This volume). Lembaga Ilmu Pengetahuan Indonesia Shekelle, M, S. Leksono, LLS, Ischwan, & Y. Masala Y. 1997. The natural history of the tarsiers of north and central Sulawesi. Sulawesi Primate Newsletter 4(2):4-11. Shekelle M & SM. Leksono. (2004) “Rencana Konservasi di Pulau Sulawesi: Dengan Menggunakan Tarsius Sebagai ‘Flagship Taxon’”. Biota 9 (1):1-10. Sunderlin, WD. 1999. Between danger and opportunity: Indonesia and forests in an era of economic crisis and political change. Society and Natural Resources 12(6):559-570. Supriatna J, J. Manansang, L. Tumbelaka, N. Andayani, M. Indrawan, L. Darmawan, SM. Leksono, Djuwantoko, U. Seal & O. Byers, (eds) (2001) Conservation Assessment and Management Plan for the Primates of Indonesia: Final Report. Conservation Breeding Specialist Group (IUCN/SSC), Apple Valley, MN, USA. Whitten AJ, SJ. Damanik, J. Anwar & N. Hisyam. 2000. The Ecology of Sumatra. 2nd edition. Tuttle, Boston. Whitten T, M. Mustafa, & G. Henderson. 2002. The Ecology of Sulawesi. 2nd edition. Yogyakarta: Gadjah Mada University Press.

Primates of The O riental Night

CRANIOMETRY OF SLOW LORISES (GENUS Nycticebus) OF INSULAR SOUTHEAST ASIA Colin Groves1 and Ibnu Maryanto2 School of Archaeology and Anthropology, Australian National University Canberra, ACT 0200, Australia. Email: [email protected] 2 Bogor Zoological Museum-LIPI, Jl Raya Cibinong KM 46 Cibinong Bogor, Indonesia, email: [email protected] 1)

ABSTRACT We measured skulls of slow lorises (Nycticebus) from all over Sundaland, and compared them by multivariate and univariate analysis. There are slight differences of shape within Java, Borneo and Sumatra, and perhaps between those of the Riau Archipelago and Sumatra. Skulls from the Malay Peninsula average slightly larger than those from Sumatra, but otherwise are very similar. A skull from Bangka falls well within the range of variation of those from Borneo; one from P.Bunguran and one from P.Tioman (Malaysia) fall within the Sumatra/Malay range. Measurements from the literature of skulls from Tawitawi (Philippines) show that they do not differ from those from Borneo. The Sumatra/Malay/Riau/Bunguran/Tioman sample differs greatly on average from the Borneo/Bangka /Ta wita wi sample, a nd they form two strongly distinct su bspecies, Ny cticeb us cou can g c ouc ang and N.c.menagensis, respectively. The Java sample differs rather more from the others, and this, taken together with its apparently consistent external differences, induces us to recognize it as a full species, N.javanicus.

Key words: Slow loris, Nyctecebus, Souteast Asia

INTRODUCTION Biological taxonomy typically goes through two phases: a long period during which specimens are collected from ever more localities and new species and/or subspecies are described one by one, followed by a period of consolidation when the described taxa are compared in detail to each other and overall revisions are proposed. In this paper, we will briefly review the history of the taxonomy of Indonesian Slow Lorises (genus Nycticebus), then reconsider the position on the basis of our own cranial data. During the collection/description phase the following taxa were proposed (Figure8): coucang Boddaert, 1785. “Bengal”; but kukang is the word for loris in some Melayu dialects.” javanicus E. Geoffroy St. Hilaire, 1812. Java. malaiana Anderson, 1881. Melaka. menagensis Lydekker, 1893. Referring to Nachtrieb’s “A new lemur (Menagensis)”; Lydekker, in compiling the Zoological Record for 1893, listed Nachtrieb’s paper under the name Lemur menagensis, so this is available (see Timm & Birney, 1992). Tawitawi, Sulu Is.

hilleri Stone & Rehn, 1902. Tanah Datar, Padang Highlands. natunae Stone & Rehn, 1902. P. Bunguran. borneanus Lyon, 1906. Sakaiam R., Sangau district, W. Kalimantan. bancanus Lyon, 1906. Bangka. philippinus Cabrera, 1908. Supposedly from Catagan, Mindanao (but see Fooden, 1991). insularis Robinson, 1917. P. Tioman. buku Robinson, 1917 (ex Martin, 1838, actually a langur). Sumatra. ornatus Thomas, 1921. Batavia (=Jakarta). brachycephalus Sody, 1949. Supposedly from P. Tebingtinggi. The period of consolidation can be said to have begun with Osman Hill’s (1953) monograph, in which all slow lorises were assigned to a single species, Nycticebus coucang, but with numerous subspecies. Groves (1971) revised these taxa and, after splitting off Indochinese N. pygmaeus a distinct species, divided N. coucang into four subspecies: N. c. javanicus (Java), N. c. menagensis (Tawitawi, Borneo and Bangka), N. c. coucang (Sumatra, Riau Archipelago, Malay peninsula, P.Bunguran and P.Tioman) and N.c.bengalensis (mainland Southeast Asia, from the Isthmus of Kra north to Assam and

115

Groves & Maryanto - Craniometry of Slow Lorises (Genus Nycticebus) of Insular Southeast Asia.

southernmost China). Noteworthy in this arrangement was the allocation of the Bangka lorises to the Bornean subspecies, not to that of nearby Sumatra; and of those of P. Bunguran, in the North Natuna Is., to the Malay/Sumatran subspecies rather than to that of Borneo. Groves (1971) explained these curiosities by reference to the pattern of drowned rivers on the Sunda Shelf. His conclusions were confirmed and extended by Ravosa (1998). Later, Groves (1998) returned to the problem, and reconfirmed this arrangement on the basis of multivariate morphometrics, but without detailed analysis. The major novelty was the proposal to separate N. bengalensis as a distinct species.

6

11? 5

1,3 13

Most recently, Supriatna & Hendras (2000) separated the Javan Slow Loris as a full species, N. javanicus, leaving just those from Sumatra, Borneo, Peninsular Malaya, and offshore islands in N. coucang. MATERIAL AND METHODS One of us (IM) measured the crania in the Museum Zoologicum Bogoriense, Research Centre for Biology-LIPI collection at Cibinong, Jawa Barat. CPG compared this dataset with that from the European and American collections, as described by Groves (1971), and incorporated the measurements of Tawitawi

4,?9

10

7

8

2,12

Figure 1: Type localities of described taxa of Sundaland slow lorises

116

Primates of The O riental Night

specimens given by Timm & Birney (1992). Both univariate and multivariate comparisons were made using SPSS version 11 for Windows. Because not all measurements were available for every specimen, a series of Discriminant Analyses and one Principal Components Analysis were run, each time using the measurements that were available for the largest number of specimens in each sample compared.

(1) Geographic variation within the Greater Sunda Islands There is no evidence of variation in size, as measured by skull length, from west to east in Java (Figure 2). On a Principal Components Analysis, the first Component (accounting for 51.4% of the total variance) is strongly dependant on absolute size, but the second (which accounts for 32.4%) is a shape component, representing predominantly a contrast between wide posterior palate and narrow snout; values of PC2 tend to decline (very slightly) from west to east, meaning that the posterior palate becomes relatively narrower, the snout relatively broader (Figure 2b). Figure2b. This in fact depends entirely on two

specimens from the east, and does not counter our impression, gained from pelage characters, of an essential homogeneity within Java, and there’s a single specimen from far west. It would be interesting to test the possibility of variation with altitude within Java, but our material does not span sufficient altitudinal range to enable us to test this. As Figure 3a shows, there may be an average difference between lorises from southeastern Kalimantan and Sabah on Discriminant Function 1, which accounts for 71% of the variance and contrasts large skull length with relative narrowness especially of the palate, but this is probably an effect of small sample size, because skulls from western Borneo span the entire range. As far as absolute size is concerned (Figure3b), Sabah skulls average somewhat larger than other samples, but there is almost total overlap. All Bornean skulls, like those from Java, will therefore be treated as a single sample. Figure 4 compares skulls from northern and southern Sumatra and the Riau Archipelago (Kepulauan Riau), also the type of brachycephalus Sody, 1949 (a zoo specimen, reputedly from P. Tebingtinggi). In an analysis using only three variables, and so

62

2.0

RESULTS

1.5 1.0

Principal C om ponent 2

60

SKULL LENGTH

58

56

54 106

107

LONGITUDE

108

109

110

111

112

113

.5 0.0 -.5 -1.0 -1.5 -2.0 106

107

108

109

110

111

112

113

LONGITUDE

Figure 2: Variation by longitude in Java a: skull length, b: second principal component

117

Groves & Maryanto - Craniometry of Slow Lorises (Genus Nycticebus) of Insular Southeast Asia.

b.

a. 3

62

Region 60

2

58

1

Ungrouped Cases 56

0

Riau Arch. P.Tebingtinggi

-2

N.Sumatra

-3

S.Sumatra -5

-4

-3

-2

-1

0

1

2

3

Skull length

F unctio n 2

54

-1

52 50 48 N=

4

Bangka

3

8

E Kalimantan

Function 1

7

W Kalimantan

12

Tawitawi Sabah

REGION

Figure 3: Geographic variation within Borneo a: Canonical Discriminant Functions Borneo (Gtl, Zyg, Palbr, Bican, Ramus), b: Skull length.

maximizing the number of specimens that can be analysed (Figure 4a), DF1 (accounting for 68.1% of the variance) contrasts wide zygomata with narrow snout; DF2 (29.7% of the variance) contrasts broad snout with short skull. In the analysis using six variables, and with a consequently smaller dataset (Figure4b), DF1 (52.1% of the variance) contrasts large size and, especially, flaring zygomata with narrow biorbital breadth and palate; DF2 (42.5%) emphasizes in particular a relatively long basicranium (staphylion to basion distance). In neither analysis is there any difference between skulls from northern and southern Sumatra. Specimens from the Riau Archipelago appear at first sight to differ, especially in Figure 4a, but the series labeled “ungrouped” are specimens labeled “Sumatra” without more definite locality; as these cover the range of those from Kep. Riau, it would therefore seem, on the face of it, that no separation can be maintained. Using a larger set of variables, much smaller samples are available. All Sumatran and Riau Archipelago specimens will therefore be treated as a single sample. 2. Geographic variation between Borneo and Sumatra Using only four variables enables us to include the Tawitawi sample, using the measurements given in Timm & Birney (1992). This is necessary

118

because Tawitawi is the type area of menagensis Lydekker, 1893, the earliest available name for a Slow Loris in island Southeast Asia. The results are shown in Figure 4a. The first Discriminant Function, accounting for 88.9% of the variance, contrasts long skull and broad palate with narrow zygomata; the second, accounting for only 10.4%, contrasts skull breadth with skull length. There is an average difference between Borneo and Sumatra; 12 of the 20 Sumatran skulls are correctly classified. The 10 Tawitawi skulls and the single Bangka skull (type of bancanus) fall within the range of Borneo; none of them is misclassified as Sumatra, and only 2 of the 23 Bornean skulls are misallocated to Sumatra. Using 9 variables (Figure 4b) means that no Tawitawi skulls are now able to be included. DF1 (96.9% of the variance) contrasts palate breadth and length and biorbital breadth with skull length, ramus height and snout length; DF2 (only 3.1%) contrasts predominantly zygomatic breadth with skull length. Borneo and Sumatra now separate much better, though there is still an overlap; all but 3 of the 17 Sumatran skulls are correctly classified, and all but 2 of the 19 Bornean skulls. The Bangka skull again assorts with Borneo, even though it was entered as a separate group, which would give it a more-than-even chance to sort separately.

Primates of The O riental Night

We conclude that (1) Borneo and Tawitawi skulls cannot be distinguished, so borneanus Lyon, 1906 is a synonym of menagensis Lydekker, 1893; (2) the Borneo-Tawitawi taxon can be distinguished, on average but not absolutely, from that from Sumatra; and (3) the single available Bangka skull allocates to Borneo/Tawitawi, not to Sumatra, so bancanus Lyon, 1906 is a synonym of menagensis. 3.The affinities of Malay peninsular lorises Consideration of the taxonomic affinities of the lorises of the Malay Peninsula are necessary because the type locality of Tardigradus coucang Boddaert, 1785, was considered to be “probably Malacca” by Chasen (1939), and this was implicitly accepted as a valid fixation by Groves (1971). It is also convenient to consider the affinities of the lorises of two small islands in this context: P. Tioman, east of the Malay Peninsula, and P. Bunguran, in the North Natuna group, southeast of the peninsula and north of Borneo; these will be considered further below. A single analysis was run (Figure 5), using 4 variables. DF1 (83.2% of the variance) largely contrasts long skull with narrow bizygomatic breadth; DF2 (16.8%) contrasts broad palate and snout with narrow bizygomatic breadth. The Malay and Sumatran samples differ weakly on average only. Consequently Sumatran lorises can be included in nominotypical

a.

coucang. Tioman appears to fall at the edge of the Malay sample, but the two Bunguran skulls fall within it. 4. Comparing Borneo with Malay/Sumatran and insular lorises Using 9 variables (Figure 7a), DF1 (62.8% of the variance) contrasts biorbital breadth to skull length, and DF2 (20%) contrasts palate breadth, basicranial length and mandible length to skull length and biorbital breadth. The substantial samples from the Malay Peninsula and Sumatra still overlap widely, though with somewhat different “centres of gravity”. The Borneo sample is largely separate from the Malay/ Sumatran samples, though with overlaps; Bangka again falls within Borneo. Two from P. Bunguran (including the type of natunae) and one from P. Tioman (type of insularis) fall within the Malay/Sumatran dispersion, even though they were entered as separate groups. This confirms and extends the results of Figure 5. As far as absolute size is concerned (Figure 7b), the Sumatran and Riau specimens average smaller than those from the Malay Peninsula, Bunguran and Tioman, but there is extensive overlap. 5. Comparing Java with other Sundaland lorises Combining Borneo, Bangka and Tawitawi as Nycticebus coucang menagensis, and Malay peninsula,

b. 3

4

Region

Region

2

2

1

Ungrouped Cases

Ungrouped Cases

0

Riau Arch.

Riau Arch.

-1

P.Tebingtinggi N.Sumatra

-2 -3 -5

S.Sumatra -4

-3

Function 1

-2

-1

0

1

2

3

Function 2

Function 2

0

-2 P.Tebingtinggi -4

N.Sumatra

-6

S.Sumatra -8

-6

-4

-2

0

2

4

6

Function 1

Figure 4: Geographic variation on Sumatra and its offshore islands. Canonical a:Canonical diskriminant functions (Reduce variable set: Gtl, Zyg, Bican), b: canonical discriminant funcitions (Reduce variable set: Gtl, Zyg, Biorb, Stbas, Stpros, Bican).

119

Groves & Maryanto - Craniometry of Slow Lorises (Genus Nycticebus) of Insular Southeast Asia.

Sumatra, Kep.Riau, P.Tioman and P.Bunguran as N.c.coucang, achieves samples large enough to make it worthwhile to perform a stepwise comparison with N.c.javanicus. A Stepwise Discriminant Function Analysis enters the variables separately, one by one, at each stage withdrawing those which add nothing to the discrimination (according to the criteria of the Mahalanobis method, partial F to enter >3.84, partial F to remove <2.71), until only a subset remains. In this case, after 16 enter-removal steps, four variables remained: Mandible Length, Posterior Palate Breadth, Biorbital Breadth, and Palate Length. DF 1, accounting for 85.9% of the total variance, contrasts low values for biorbital breadth with high values for the other variables; DF2 (14.1%) contrasts low values for Mandibular Length with high values for the rest. Figure8a is the result. The three taxa are all separated but overlap, though javanicus overlaps less with they other two than they do with each other. The percentage of each taxon that is correctly classified is as follows: javanicus 84.6% (n=13), menagensis 95% (n=20), coucang 65.8% (n=38). Figure 8b shows the absolute size of the three Sundaland taxa. Of the three, javanicus averages largest, followed by coucang, with menagensis much the smallest. The means, standard deviations and sample sizes are as follows: javanicus 60.2 ± 2.22 (25) menagensis 54.9 ± 2.25 (40) coucang58.6 ± 2.36 (59) a.

DISCUSSION AND CONCLUSION This study has shown that, while there are slight variations in loris craniometrics within each of the three Greater Sunda islands, it is between them that the major differentiation occurs. This corroborates their separation as three distinct taxa. In addition, multivariate analysis supports the proposition that the Bangka loris is consubspecific with that of Borneo, as are those of Tawitawi in the Philippines, whereas those of Kepulauan Riau, P.Tebingtinggi, P.Bunguran, and the Malay Peninsula and P.Tioman in Malaysia, are consubspecific with the Sumatran loris. How should these be classified? Our craniometric analyses indicate that, while all three taxa overlap, the Bornean and Sumatran forms overlap with each other more than either does with Java. Groves (1998) retained them as three subspecies of a single species, Nycticebus coucang, whereas Supriatna and Hendras (2000) separated Javan lorises as a full species Nycticebus javanicus. We are inclined to accept their revision, given that it is, in our experience, absolutely different in its colour pattern (see also Groves 2001). We note that Ravosa (1998, see especially Figure8b) likewise found that javanicus is strongly differentiated from other slow lorises. Ravosa (personal communication) has suggested to CPG that there might be a case for giving species rank to menagensis in addition; its high frequency of upper I2 absence (see b.

3

3

2

2

REGION

1

REGION

1 Sumatra

0 Bangka -1

Tawitawi Borneo

-2 -3

-2

Function 1

-1

0

1

2

3

4

5

Function 2

Function 2

Sumatra 0

Bangka -1 -2

Borneo -3

-2

-1

0

1

2

3

Function 1

Figure 5: Comparisons between Borneo and Sumatra, and the position of Bangka a: canonical discriminant functions (reduced variable set: Gtl, Zyg, Palbr, Ramus), b: canonical discriminant functions (reduced variable set: 9 craniometric variables)

120

Primates of The O riental Night

3 2 1

Malay Pen.

Function 2

0 -1

Bunguran

-2

Tioman

-3

Sumatra -4

-3

-2

-1

0

1

2

3

Function 1

Figure 6: Comparison of Malay and Sumatran lorises by canonical discrimanant functions (Reduced variable set: Gtl, Zyg, Palbr, Bican, Ramus)

a.

b.

REGION

66

3 Malay Pen.

64

2 P.Tioman 1 0

60

Sumatra

-1

58

Bangka Skull length

Function 2

62

P.Bunguran

-2 -3

Borneo -4

-3

-2

-1

0

1

2

3

56 54 52 N=

3

15

P.Bunguran

2

Malay Pen

Function 1

5

P.Penang

4

21

Riau Arch

Pen Thai

1

2

P.Tebin gtinggi

Sumatra

P.T ioman

REGION

Figure 7: Comparison of Borneo with Malaya and Sumatra and island lorises a: canonical discriminant fintions (9 craniometric variables), b: skull length

a.

b.

3

70

2 1

60

TAXON

-1

coucang

-2

menagrensis javanicus

-3 -4

-2

Function 1

0

2

4

6

SKULL LENGTH

Function 2

0 50

40 N=

25

40

59

javanicus

menagensis

coucang

TAXON

Figure 8: Comparison of Java with other Sundaland lorises a: canonical discriminant fintions (8 variables enterd 4 selected), b: skull length

121

Groves & Maryanto - Craniometry of Slow Lorises (Genus Nycticebus) of Insular Southeast Asia.

below) differentiates it strongly from coucang and brings it close to javanicus. While it is true that menagensis is strongly differentiated (see Figure 8b), in ways unrelated to allometry as was shown by Ravosa (1998), the available evidence does not indicate that they are 100% different. We await future DNA studies, which may alter our opinion. Of other characters found to be useful in loris taxonomy, the only other that is applicable to Sundaland forms is the number of upper incisors. In our sample, all 13 skulls of N. javanicus in which the character can be confidently detected have a single pair of upper incisors, as do all 22 skulls from Borneo and Bangka. Ravosa (1998), however, found that this character is not completely fixed: on the basis of larger samples than available to us, he found a single upper incisor pair present in 84% from Borneo, and 95% from Java (see Ravosa, 1998, Table II). Of 42 skulls of the Sumatra/ Malay taxon, however, 34 have two pairs of incisors on each side, 6 have a single pair, and two skulls have one pair on one side, two on the other. Out of 84 sides, therefore, 17% have a single pair, 83% have two pairs. This strong difference greatly adds to the case for subspecific differentiation between N.c.coucang (Sumatra/Malay) and N.c.menagensis (Borneo/ Bangka). The taxonomy of Sundaland lorises therefore is as follows: 1) Nycticebus javanicus E.Geoffroy St.Hilaire, 1812. Synonym ornatus Thomas, 1921. Java. 2) Nycticebus coucang (Boddaert, 1785). a) Nycticebus coucang coucang (Boddaert, 1785). Synonyms malaiana Anderson, 1881; hilleri Stone & Rehn, 1902; natunae Stone & Rehn, 1902; insularis Robinson, 1917; buku Robinson, 1917; brachycephalus Sody, 1949. Sumatra, Riau Archipelago, P.Tebingtinggi, P.Bunguran, Malay Peninsula, P.Tioman. b) Nycticebus coucang menagensis (Lydekker, 1893). Synonyms borneanus Lyon, 1906; bancanus Lyon, 1906; philippinus Cabrera, 1908. Borneo, Bangka, Tawitawi (Philippines).

122

ACKNOWLEDGEMENTS CPG once again acknowledges with pleasure the curators of the collections in Washington, New York, Philadelphia, London, Leiden, Amsterdam and Singapore for their kindness and hospitality. Grateful thanks are due to Matt Ravosa and Myron Shekelle for their helpful comments on the manuscript. We both acknowledge the help and friendship of Bpk.Boeadi, the former curator of the mammal collection of Museum Zoologicum Bogoriense. REFERENCES Chasen, F.N. 1939. A handlist of Malaysian mammals. Bull.Raff. Mus.Singapore, 15:i-xx, 1-209. Fooden, J. 1991. Eastern limit of distribution of the slow loris, Nycticebus coucang. Int.J.Primatol. 12:287-290. Groves, CP. 1971. Systematics of the genus Nycticebus. Proc.3rd Int.Congr.Primatol., Zurich 1970, 1:44-53. Groves, C. 1998. Systematics of tarsiers and lorises. Primates, 39:13-27. Grovcs, C. 2001. Primate Taxonomy. Washington: Smithsonian Institution Press. Hill, WCO. 1953. Primates: Comparative Anatomy and Taxonomy. Strepsirhini. Edinburgh: Edinburgh University Press. Ravosa, MJ. 1998. Cranial allometry and geographic variation in Slow Lorises (Nycticebus). Amer.J.Primatol.45:225-243. Supriatna, J. & EH.Wahyono. 2000. Panduan Lapangan Primata Indonesia. Jakarta: Yaysan Obor Indonesia. Timm, R.M. & E.C.Birney. 1992. Systematic notes on the Philippine Slow Loris, Nycticebus coucang menagensis (Lydekker, 1893) (Primates: Lorisidae). Int.J.Primatol.13:679-686.

Primates of The O riental Night

ENCLOSURE DESIGN FOR CAPTIVE SLOW AND PYGMY LORISES Helena Fitch-Snyder1 Helga Schulze2 and Ulrike Streicher3 Loris Conservation International, 5624 Jockey Way, Bonita, California 91902, USA, Email: [email protected] 2 Ruhr-University, MA 6/161 b, 44780 Bochum, Germany 3 Endangered Primate Rescue Center, Cuc Phuong National Park, Nho Quan District, Ninh Bin, Vietnam 1

ABSTRACT While large numbers of slow and pygmy lorises are commonly kept in local zoos and rescue centers, information about enclosure design and minimal housing requirements is often lacking. We present recommendations for designing indoor and outdoor loris enclosures for exhibits, rescue centers, and sanctuaries. We discuss the advantages and disadvantages of each enclosure type and address construction specifications, furnishings, environmental requirements, social considerations, and keeper monitoring. Essential requirements for loris release into naturalistic outdoor enclosures are presented along with questions for future study. Keywords: Loris, Nycticebus, primate husbandry, prosimian housing, cage design, free-ranging enclosures.

INTRODUCTION Slow and pygmy lorises are nocturnal, arboreal prosimian primates. Slow lorises (Nycticebus coucang and N. bengalensis) occur throughout most of Southeast Asia and parts of India (Schulze, et al., 2003), whereas, pygmy lorises (N. pygmaeus) are limited to Vietnam, Laos, and Southern China (Duckworth, 1994, Zhang et al. 1995). Their diets consist of fruits, insects, small fauna, tree sap, and other plant food such as floral nectar (Wiens, 2002; Fitch-Snyder and Thanh, 2002). (See also: http:// www.loris-conservation.org/database/captive_care/ nutrition.html.) Lorises use slow, deliberate movements to locomote from branch to branch; and they do not leap, as do most other primates. Urine scent markings have a strong characteristic odor and are an important means of intra-specific communication (Fisher et al., 2003). Female slow lorises are fertile throughout the year and have monthly estrous cycles (Izard and Weisenseel, 1989). Pygmy lorises are seasonally fertile during the months of July and August (Fitch-Snyder and Jurke, 2003). Infants are well developed when they are born after a gestation period of six months, and mothers often park their infants on branches while they leave to forage (Fitch-Snyder and Ehrlich, 2003). Slow lorises

normally have singletons, while litter size for pygmy lorises is usually two or more (Fitch-Snyder, 1997, Weisenseel and Izard, 1998). This paper provides recommendations for maintaining slow and pygmy lorises in captivity, especially for Southeast Asian zoos and sanctuaries. These facilities are often required to house large numbers of lorises, many of which are acquired through government confiscations or donations (personal observation). Some facilities can provide only minimal, off-exhibit housing, while others are able to build more complex environments that might also include public viewing. Basic information about enclosure design and minimal housing requirements is often lacking. Therefore, recommendations made here are based on conditions that have been successful in other loris facilities. Wire Cages In some facilities such as primate rescue centers, wire cages may be the best option available. An outdoor cage measuring 2.00 m x 2.50 m x 1.80 m can successfully house 1-3 slow lorises if the furnishings are sufficient. (See climbing structures and nest box sections.) Wire should always be free of rust or sharp edges. Poly vinyl coated wire is ideal

123

Snyder , Schulze & Streicher- Enclosure Design For Captive Slow and Pigmy Lorises

because it resists corrosion from moisture and loris urine marking. Wire gage of 2 cm x 2 cm is comfortable for lorises to grasp, and it will keep rodents and potential predators outside. Outdoor enclosures must also have a solid roof to protect lorises from sun and rain. Maximum flexibility can be achieved by building several smaller cages (minimum size of 1.70 m x 1.00 m x .70 m per slow loris), which are connected with removable wire tunnels. Depending on whether the tunnel gates are open or closed, lorises can be kept alone or given access to other enclosures. If cages share common walls, double wire mesh or solid walls must be used to prevent lorises from biting their neighbor’s fingers. Keeper doors should be large enough for a person to walk inside the enclosure or easily reach any area inside the cage. Doorframes must be made of a solid material that will not bend. Otherwise, lorises may be able to escape by squeezing their bodies through the small gaps between door openings. Cages should be elevated at least 15 cm above the ground to so that excreta and other waste will fall below. Indoor cages can easily be moved for cleaning if wheels are attached to the bottoms. Food dishes and nest boxes can be placed on wire shelves, which are also useful for loris resting places. INDOOR ENCLOSURES Advantages The primary advantage of indoor loris enclosures is that the environmental temperature and light conditions can be controlled. By reversing the lorises’ day-night cycle through manipulation of the photoperiod, these nocturnal primates will change their activity patterns to be active during daylight hours. This management technique is especially beneficial to keepers and the public because the lorises’ activities can be observed during normal working hours. Additional benefits of indoor enclosures are the reduced possibilities of encounters with wild predators or disease-carrying wildlife.

124

Disadvantages Slow lorises are especially sensitive to light cycles (Frederick and Fernandes, 1994), and they might not become active during the day if artificial lighting conditions are not adequate. Abnormal photoperiods can modify a lorises’ reproductive cycle. For example, pygmy lorises are normally fertile in late summer, but they reproduce throughout the year when they are exposed to artificial photoperiods (Fitch-Snyder and Jurke, 2003). Seasonally changing environmental conditions has been found to also influence lorises’ physiology (Streicher, 2002) the results of lack of these stimuli are not known. Changes in activity cycles can be due to insufficient contrast between day and night cycles or use of the wrong spectrum of light to simulate daylight. Artificial lighting, environmental temperatures, ventilation, and humidity are all dependant on electricity to function properly in indoor enclosures. These environmental conditions could be seriously compromised during a power outage. RECOMMENDATIONS Climbing structures Loris enclosures that feature large concrete climbing structures are not optimal. Branches that are small enough in diameter for them to grasp with their hands and feet are preferred. Some branches should be horizontal, with varying circumferences for comfortably sitting, climbing, and hanging. Small horizontal branches (1-2 cm in diameter) are especially important for breeding purposes, because copulation usually takes place in a suspended position with hands and feet clinging to horizontal branches (personal observation). Additionally, most behavioral postures are exhibited preferentially on horizontal branches (Glassman and Wells, 1984). Branches should be nontoxic natural wood that is flexible and allows speciesspecific natural activities. In the wild, lorises use their toothcomb to obtain gum (sap) from tree branches (Tan and Drake, 2001). This chewing behavior is also important for them to maintain dental health. Fresh

Primates of The O riental Night

branches should be provided regularly to encourage dental scraping. Branches must be arranged close enough to each other to form a continuous pathway in all directions. (See Figure 1 for maximum gap widths.) Because lorises can not jump across gaps (Schulze, 1998), they reach for one branch while firmly gripping another. The most desirable climbing structures should be situated so that the animals will spend their time in areas of their enclosure that are most visible to the public (Figure 1). It is also important to furnish dense, leafy branches to provide cover for lorises (Bottcher-Law, 2001). Leafy branches are used for sleeping sites and they help to reduce stress caused by prolonged exposure to people or other lorises. While dense vegetation and perching is beneficial to lorises, keeper access and cleaning procedures are also important to consider when designing the perching network.

Poorly positioned branches can be obstructive to the keeper as well as stressful to the lorises during prolonged chases to catch an animal. In open enclosures, it is important for lorises to have suitable climbing structures that do not also provide potential escape routes. Cement barrier walls need to be completely smooth so the lorises are unable to climb up and escape. Branches should not hang over the outside of the enclosure or be close to outside escape paths. (See Table 1. to determine possible loris reach capabilities) Food and water Food and water must be accessible from several places to avoid competition. Feeding sources should be in elevated areas so that the animals are not forced to come to the ground. A shelf above the food dishes will prevent contamination by debris from above.

Figure 1. Examples of proper placement of branches and perches in an indoor enclosure. These figures are only for illustration of different ways to use substrates. Lorises must not be housed together in large numbers as suggested by the illustration on the left. (Drawing by H. Schulze.)

125

Snyder , Schulze & Streicher- Enclosure Design For Captive Slow and Pigmy Lorises

Fresh food should be offered close to the time when lorises normally wake up. Food deteriorates and can attract vermin if it is not eaten soon after it is provided Whenever possible, lorises should be offered food items in ways that create a more enriched environment. For example, pieces of fruit can be attached to tips of branches to encourage lorises to forage naturally. See Bottcher-Law (2001) or http:// www.loris-conservation.org/database/captive_care/ manual/html/Habitat.html#Enrichment for additional ideas regarding enrichment for captive lorises. Nest boxes At least one next box measuring approximately 20 square cm and no smaller than 30 x10 x 16 cm must be provided for each loris to enable it to have a safe retreat. Nest boxes should be made out of wood, preferably, but an acceptable alternative is a PVC tube (minimum of 12 cm in diameter). Small cardboard boxes can be used as temporary nest boxes if they are discarded when they become soiled. Nest boxes can be disguised with moss or other natural items so they have the natural look of the exhibit. A removable nest box situated in an easily accessible area is advantageous for handling or capture. (Figure 2.) The whole box can be removed while the loris is contained inside. With a hinged lid, a loris can be accessed with ease. The nest box should have a smooth surface so the loris can easily be removed without being able to grasp the box (Bottcher-Law, 2001). A Plexiglas panel on one side of

the box enables the public and keepers alike to view the animal closely. Wooden nest boxes must be regularly replaced when they become urine-soaked or excessively worn from lorises chewing on the wood. Lighting The loris enclosure must provide a nocturnal and diurnal period in order to have a reversed light cycle. For a reversed light cycle, lights should be turned on for 12 hours during the night to encourage the lorises to sleep. A minimum of 75ft candles is required (Keeling, 1974). Simulated night time will then occur during the day. A dimming feature to simulate dawn and dusk is preferable to abruptly turning the “daylight” on and off. Neutral density acetate filters will create the most ideal simulation of natural moonlight (Pariente, 1980). This lighting scheme realistically depicts the forest at night and does not distort the animals’ true fur and eye color. If neutral density filters are not available, green lights can be substituted (L. Bottcher-Law, personal communication). Red or blue light can also be used if necessary, but see Fredrick and Fernandez (1994). Any light during the loris’ inactive period of darkness may disrupt their circadian rhythms and can result in extended inactivity during the lorises ‘ active period (Fitch-Snyder, unpublished data). In a nocturnal house, access to the outside must have a “double door” system so the outside door is completely closed before the inner door is opened. With this modification, the enclosure will not be lit inadvertently.

Figure 2. Wooden nest box for lorises. The lid should be attached with a hinge so it can be opened to view or remove the loris. (Drawing by H.Schulze.)

126

Primates of The O riental Night

Plants Leafy foliage is important for stress reduction because it provides visual barriers for lorises. Living plants will enhance the experience of both visitor and animal. Even with a reversed photoperiod, some plants can be grown indoors. Sensitive plants should be kept in moveable containers that can be periodically rotated outdoors for temporary recovery periods. Another option is to attach narrow water containers to climbing structures to hold fresh cuttings from leafy plants. Cuttings will stay fresh for several weeks if new water is added regularly. Artificial plants can also provide leafy foliage if no other options are available. Ventilation Air circulation is very important for indoor enclosures, especially for animals that scent-mark and communicate through olfaction. Ten to 15 air changes per hour are recommended (Anonymous, 1993). Recirculation of air is discouraged because it will result in a buildup of odors and ammonia levels. Lorises must be able to escape from potential drafts of cold air caused by air conditioning systems. Temperature Lorises should be kept within the normal temperature range of their native environment. With proper shelter, they can comfortably live in outdoor temperatures between 18° to 30° C. Thus lorises can be kept permanently outdoors in many facilities in Southeast Asia. In northern Vietnam and southern China, lorises in the wild endure cold temperatures of 5° C and below (Streicher, personal observation). However, captive indoor environments may not provide all the necessary furnishings for lorises to adjust comfortably to extreme ends of this range, and there may be forms from warmer areas that are not well adapted to low temperatures. Therefore, it is best to keep ambient temperature close to the middle of these limits and avoid rapid and extreme fluctuations. Lorises (especially N. pygmaeus) do not always choose warm sleeping places and have been found in a state of hypothermia after sleeping on cold substrates (FitchSnyder and Schweigert, pers obs). Torpor might be a physiological adaptation of pygmy lorises in a

seasonal climate (Streicher, in preparation) but the effects of captive hypothermic stress can result in illness (Schulze, pers obs). It is therefore advisable to avoid possible hypothermia and insulate any sleeping substrates made from concrete or metal. Humidity Lorises normally live in humid natural environments, and it is important that indoor enclosures maintain a similar humid atmosphere. While too much moisture can encourage mold growth, regular spraying of vegetation counteracts the drying effect that often occurs when artificial heaters are used. Cleaning It is important that enclosures are regularly cleaned and sanitized to dilute the accumulation of urine and scent marks. If chemical agents are used, the animals must first be removed, and all traces of chemical agents must be washed away before lorises are returned. While hygiene is important, this should not be at the expense of environmental complexity and stimulation. Over cleaning will remove the animals’ chemical signals and is likely to create an unfamiliar and possibly more stressful environment. This situation can be minimized by cleaning sections of the enclosure at intervals. Climbing structures, branches, nest boxes, and natural ground covers must be regularly replaced after several months as these items become especially worn or saturated with loris excreta. Noise Lorises and other animals are sensitive to loud noises from visitors and other sources such as ventilation systems. Indoor enclosures should be designed to absorb as much sound as possible to avoid an “echo effect”. Natural substrates on the ground and insulated walls can help absorb unwanted noise. Lorises do not normally live near loud, fastmoving water sources. Therefore, continuous noise from waterfalls, for example, could be disturbing and should thus be avoided around loris exhibit areas.

127

Snyder , Schulze & Streicher- Enclosure Design For Captive Slow and Pigmy Lorises

Ground cover A natural substrate on the floor of the enclosure will absorb excreta and muffle noise. Natural substrates could be leaves, bark, woodchip, or shredded coconut hulls. Sawdust should be avoided because it may cause respiratory problems (Fitch-Snyder, pers obs). Provision of plenty of well-insulated structures such as raised branches and next boxes will discourage the loris from sleeping on the floor or bare concrete surfaces. A thick layer of the suggested substrates will insulate a concrete floor. Some lorises prefer to sleep on the ground if they can hide under natural substrates, so care must be taken by keepers to not accidentally step on them when working in their enclosures. Monitoring area Lorises should be examined daily to monitor health and reproductive condition. Keepers must be able to clearly view the underside as well as the dorsum. A good way to do this is to install wire mesh windows between the enclosure and keeper areas. The wire gage should be approximately 1-2 cm wide and have a smooth surface (with no rust) so that lorises can climb on it comfortably. Using mealworms, grapes, or other preferred treats; lorises can be trained to climb on these wire panels. Keepers can then clearly view the lorises’ ventral areas to check for injuries, signs of reproductive changes, and pregnancy. Mixed species Under rescue center conditions, mixing of species is generally not recommended. In zoos, mixed species exhibits can be useful tool for utilizing space, providing animal enrichment, and educating the public about sympatric or taxonomically similar species. Slow and pygmy lorises have successfully been housed in enclosures with the following other animals: aye ayes, bush babies, pottos, slender lorises, mouse lemurs, Malayan mouse deer, echidna, aardvark, giant fruit bats, Egyptian fruit bats, bettong, Asian crested porcupine, tree shrews, and sugargliders (Lester, 2001). Slender

128

lorises have additionally been kept with pygmy lorises, fat-tailed lemurs, and hedgehogs (Lester, 2001). Mixed species exhibits should always be closely monitored to ensure one species does not interfere with reproduction or other natural behaviors. Visual barriers are useful for reducing stress. A disadvantage of mixed species exhibits is that it may be difficult to prevent different species from eating each other’s food. Scent marks, feces, or other activities of one species may be disturbing to the other. This may be especially true of the other species that are not native to Asia, such as the lemurs and tenrics. Mixing different loris taxa together is not advisable if there is any risk of inbreeding. Analysis of loris reproductive data (Fitch-Snyder, 1997) indicates that Loris tardigradus, Nycticebus pygmaeus, and N. bengalensis do not interbreed in captivity. However, N. bengalensis and N. coucang varieties will readily reproduce together. Disease transmission is a potential risk of mixed-species exhibits. Two lorises died of infection with Pasteurella after contracting it from a Provost’s squirrel housed in the same enclosure and showing no signs of illness (Fitch-Snyder and Schulze, 2001). It is not advisable to house lorises with any carnivore, predatory bird, snake or potential prey. Pythons (Wiens and Zitzmann, 1999) and orangutans (Utami and Van Hooff, 1997) are known to prey on lorises. Lorises will kill and eat small birds, lizards, and rodents. Indoor exhibit design Lorises are traditionally kept in wire or concrete enclosures. These types of housing are functional and inexpensive and therefore often the most feasible option for rescue centers and sanctuaries. In contrast, for zoos and facilities aiming to exhibit lorises, open indoor exhibits can enhance both the visitor’ experience and create a more stimulating environment for the lorises. (Figure 3.)

Primates of The O riental Night

Figure 3. Layout of an open indoor exhibit. (Drawing by H. Schulze.)

Figure 4. Feeding station design. Foliage on top of the station makes it more natural looking and provides a more protected feeding environment. (Drawing by H. Schulze.)

129

Snyder , Schulze & Streicher- Enclosure Design For Captive Slow and Pigmy Lorises

Outdoor Enclosures Advantages In outdoor environments in their native country, lorises are exposed to the natural lighting and weather conditions to which they are already well adapted. Exposure to these natural conditions encourages normal behavior and may influence important physiological patterns such as those associated with reproduction.

which the body weight is supported by one leg above and one grasping below. This allows the animal to assume a horizontal posture for some time before seizing a new support. In Table 1, bridging distances for various Nycticebus taxa are presented. When planning open enclosures, one should also recognize that lorises have the ability to use flexible twigs to enhance their reach to more distant supports. (F. Wiens, pers com)

Disadvantages A major disadvantage of outdoor enclosures is that lighting cannot be controlled. Lorises are strictly nocturnal, so they normally sleep through daylight hours. This makes a boring exhibit for visitors, and it is also more of a challenge for keepers to monitor the behavior of the animals and to check their health. Keepers and researchers will need to modify their normal work schedule in order to provide optimum care and to observe the lorises during their active period. With an outdoor enclosure, there is also a much greater risk of animal escape, predation, and transmission of disease from wild species.

Lorises may also have escaped from the island by climbing on over-grown branches that extended outside the enclosure. Before the second loris group was released, keepers modified the exhibit, trimming overgrown vegetation, adding climbing structures made out of rope, and modifying the water level in the moat to approximately 20 cm deep. It has been reported that lorises are unable to swim (Ryley, 1913), so keepers thought that the lorises would avoid the moat. However, two lorises had to be retrieved from the water: One loris went directly into the moat after the release. The other apparently fell into the water and was unable to get out without intervention (F. de Haas, pers com).

Island Exhibits Several SE Asian Zoos had problems maintaining lorises on outdoor islands (personal observation). For example, an island at the Schmutzer Primate Center in Jakarta was used to release groups of confiscated slow lorises. Most of the lorises disappeared or died within a few weeks, and the remaining lorises had to be removed from the island (F. de Haas, pers com). Possible factors responsible for the failure of the island enclosure are escape of the lorises, being preyed upon, unstable social groups, and poor initial health. Escape Lorises may have been able to escape from the island by grasping small concrete or metal outcroppings along the perimeter wall. Although they can’t jump, lorises have a remarkable ability to bridge gaps by stretching their bodies or bracing themselves along uneven vertical surfaces. If a vertical support is available, the loris may use a “cantilever posture”, in

130

Predation Wild animals or feral cats inside the zoo could have preyed upon the lorises. Wild predators such as civets, owls, or pythons might be responsible for some of the losses. One keeper reported finding clumps of loris fur on the island enclosure, which suggests possible predation by either a carnivore or raptor (F. de Hass, pers com). However, clumps of fur without the skin is more likely due to intraspecific fighting. Unstable social groupings In the wild, slow and pygmy lorises are normally found alone or in small groups of two to four (Fitch-Snyder and Thanh, 2002; Wiens, 2002). Six to eight lorises have been known to live successfully under captive conditions (Fitch-Snyder, pers obs), but these groupings are not usually stable and need to be closely monitored. In general, the larger and more complex the environment, the more likely that a loris will be able to establish their own territory and tolerate con-specifics. With the occasional exception of male

About 17 cm 1 Up to about 21 cm 2 Up to about 22 cm 2 Up to about 24 cm 2

26.3 / 29.3 cm

Up to 33 cm / - 3

Up to 34.6 cm / - 4

29.1-36.7 cm 3

N. coucang (small) N. coucang (medium) N. coucang (large) N. bengalensis Up to about 52 cm 2

Up to about 49 cm 2

Up to about 46 cm 2

About 37 cm 1

Average about 30 cm, up to about 41 cm 2

Reach of hand, starting from a vertical branch for bridging

Up to about 45 cm 2

Up to about 43 cm 2

Up to about 41 cm 2

About 32.5 cm 1

Average about 27 cm, up to about 36 cm 2

Maximum distance between attached feet during bridging

Up to about 63 cm 2

Up to about 59 cm 2

Up to about 57 cm 2

About 45 cm 1

Average about 37 cm, up to about 50 cm 2

Maximum reach vertically up or downward

Up to about 66 cm 2

Up to about 62 cm 2

Up to about 59 cm 2

About 47 cm 1

Average about 39 cm up to about 52 cm 2

Length during hanging, fully extended

1 Measurements from photos, still video and / or dead specimens. 2 Measurement of a dead N. coucang and head-body size. 3 C. Groves, pers. comm. 4 U. Streicher, pers. comm. 5 From literature and museum specimen label

Average about 14 cm, up to about 19 cm in large animals

Average 21.6 cm 4, 18 - 29 cm 5 / - 2

N. pygmaeus

Shoulder height when walking on a branch

Size of animal: head-body-length/ sitting height

Nycticebus taxon

size may be too wide). (Drawings by H. Shulze.)

Table I. Bridging distances for Nycticebus. Maximum breadth of substrate gaps that can be crossed by “bridging”. (For females carrying heavier offspring, gaps of this

Primates of The Oriental Night

131

Snyder , Schulze & Streicher- Enclosure Design For Captive Slow and Pigmy Lorises

littermates of pygmy lorises, males will nearly always fight with each other if housed together. While the initial injuries may not appear to be severe, these bite wounds usually become infected and often result in death. In is important that any loris social grouping is closely monitored for compatibility for at least several weeks before and after releasing them to a lessstructured environment. The stress of a new environment can also contribute to fighting after release. Poor physical health Lorises must be in excellent physical condition before being released into a large outdoor enclosure. The stress of this new environment can easily worsen any pre-existing health problems. Animals with dental disease, advanced age, or other potential handicaps are not good candidates for this type of housing. (See Streicher, et al. (this volume) regarding assessing loris health.) Lorises who are in otherwise good health but are lacking most of their front teeth could be candidates for this type of enclosure providing that they are able to consume a healthy diet. These lorises should not be mixed with dentally intact lorises that could harm them in the event of an aggressive encounter. Recommendations Island exhibits have the potential to be an excellent method of housing loris groups. However, no one has systematically tested this type of enclosure to resolve potential problems. In addition, the full extent of a loris’s capabilities is not known. The best way to determine if island housing is suitable is to systematically study and thoroughly document the results of any future releases. This information is essential for planning future outdoor or other naturalistic enclosures, and it will provide new insight about natural loris behavior that cannot be learned by studying lorises housed in artificial indoor cages. Questions for future inquiry regarding island housing 1. How do lorises escape from their enclosures? 2. What kinds of animals are preying on lorises?

132

3. How do lorises respond to potential predators? 4. How do lorises use their enclosures, for example, where they sleep? Where do various behaviors occur? 5. What types of fauna do lorises prey upon in their enclosure? 6. What kinds of moats or retaining walls work best to keep inside the enclosure? 7. How do the lorises interact with each other? 8. What method(s) are best to identify individual lorises? 9. What is the maximum population density possible under various housing conditions? 10. How can keepers identify signs of social stress, and how can it be minimized by factors such as group composition or environmental design? 11. How do confiscated lorises with absent teeth (usually removed by dealers) adapt to this disability? Which food items are they still able to eat? Some essential release requirements when housing confiscated lorises 1. Ensure that lorises are in good health and individually recognizable through colored plastic leg bands or stainless steel ball-chain bracelets with attached colored tassels. (See: http://www.loris-conservation.org/database/for more information.) 2. Sort loris group by taxon and avoid mixing different subspecies together in the same enclosure. (See: Schulze and Groves, (submitted).) 3. Ensure that each loris is able to obtain food and water. This is especially important for compromised animals such as those with missing teeth. 4. Do not attempt to house a group of male lorises together in the same enclosure. They are likely to fight and inflict severe injury. In order to establish a small loris group, try housing one male and several females together under close supervision for several weeks before releasing them. 5. The island should be checked for potential problems such as escape routes (including the steps in the moat area), drowning opportunities,

Primates of The O riental Night

6.

7.

8.

9.

loose electrical wires, and places where animals could become entangled. It is strongly recommended that all feral cats be removed from zoo grounds. Not only are they predators, but they can carry potentially fatal zoonotic diseases. Check the enclosure at night to make sure that there is enough light available for observers to clearly see the lorises and their activities. (The light should be no brighter than a full moon.) Colored filters should cover torches to keep the light from disturbing the lorises. Make sure that all areas of the enclosure are visible. Systematic behavioral data should be collected on each loris throughout the first few nights after release. This could be a good project for college students. Ideally, two people at a time should watch to make sure all parts of the island are monitored. The observers should be as quiet as possible so that the lorises and any potential predators will not modify their behavior in their presence. Observers may need to rotate after four-hour shifts to prevent boredom. It might also be useful to monitor the island by infrared video cameras. Locate each loris on a daily basis and check for injuries with as little disturbance as possible. Consider training lorises to come to specific areas of their enclosure whenever they hear a

whistle or other unique noise. Provide special food rewards whenever they respond to this auditory cue. 10. Watch feeding stations to make sure all lorises get food and ensure that wild animals do not eat the lorises’ food. Once the initial group adjusts to the island and an effective monitoring system is in place, additional lorises can be added to the island. Aroundthe-clock monitoring should be done again for at least the first few days after each new loris is added. Loris Sanctuaries and Free-ranging Enclosures Suggested design A large semi-natural enclosure could provide a suitable environment for a loris sanctuary or research center. A two-meter fence made from smooth plastic or metal sheeting may be sufficient to serve as a perimeter wall, but observation is necessary to ensure that lorises cannot leverage themselves up any connection points to climb up to climb to the top of the wall. An electric fence may be necessary to keep the lorises from escaping and to prevent potential predators from climbing into the enclosure. The electrical wires should be installed both inside and outside of the fence to be effective. Otherwise, predators can become trapped inside if there is electric wiring on the inside alone.

Figure 5. Loris feeding station. Roof protects food from rain and dirt and lorises from raptors. Branches connect trees with feeding stations. (Illustration by H. Schulze.).

133

Snyder , Schulze & Streicher- Enclosure Design For Captive Slow and Pigmy Lorises

The minimum size for a naturalistic enclosure such as this is 0.5 hector. The interior should be generously planted with trees and bushes. The goal is for the branches and other climbing structures to provide a continuous arboreal environment without the vegetation being so dense that it is difficult to see the lorises. The part of the enclosure closest to the keeper entrance should contain several feeding stations (Figure 5). These stations should be located in a fairly open area, but all be connected by branches or ropes to nearby trees or bushes. The number of feeding stations is dependant on the size of the enclosure and the number of lorises. Each feeding station should be build on stilts and consist of a platform large enough to hold several food and water bowls. The roof above the platform could be designed to include nest boxes to provide additional shelter. Like the island environment discussed earlier, the success of this enclosure depends on the keeper’s ability to monitor each individual. Radio collars may be useful for this purpose, along with bird rings or other methods of easy recognition by keepers and researchers. Each loris should also have a permanent means of identification such as a tattoo or transponder. It is especially useful to train the lorises to routinely approach the feeding station in response to an auditory cue from the keepers. Keepers can then do periodic inventories of the colony and be aware of any problems or changes in the population. ACKNOWLEDGEMENTS The authors thank Kay Izard, Lisa BottcherLaw and Myron Shekelle who reviewed this paper and gave valuable suggestions for improvement. We are grateful to Femke de Haas who provided historical data about lorises housed at Schmutzer Primate Center. Sincere thanks goes to Myron Schekelle who organized the workshop and coordinated this publication. Schmutzer Primate Center and the Indonesian Institute of Sciences financed and hosted the workshop that makes this publication possible.

134

REFERENCES Anonymous. 1993. IPS International guidelines for the acquisition, care and breeding of nonhuman primates: Codes of Practice 1-3. Primate Report 35: 3-29. Bottcher-Law, L. 2001. General Habitat Design. In: Management of Lorises in Captivity, A Husbandry Manual for Asian Lorisines (Nycticebus & Loris ssp.) H. Fitch-Snyder and H. Schulze eds. Zooloological Society of San Diego, San Diego. pp 72-73. Duckworth, J. 1994. Field sightings of the pygmy loris, Nycticebus pygmaeus, in Laos. Fol. Primat. 63: 99-101. Fisher, HR. Swaisgood, and H. Fitch-Snyder, H 2003. Countermarking by male pygmy lorises (Nycticebus pygmaeus): do females use odor cues to select mates with high competitive ability? Behav. Ecol. Soc. 53: 123-130. Fitch-Snyder, H. 1997. Asian Prosimian North American Regional Studbook. Zoological Society of San Diego. San Diego, USA. Fitch-Snyder, H, & A. Ehrlich. 2003. Mother-infant interactions in slow lorises (Nycticebus bengalensis) and pygmy lorises (Nycticebus pygmaeus). Fol. Primatol. 74:259-271. Fitch-Snyder, H & M. Jurke 2003. Reproductive patterns in pygmy lorises (Nycticebus pygmaeus): Behavioral and physiological correlates of gonadal activity. Zoo Biology 22:15-32. Fitch-Snyder, H. & H. Schulze (eds.) 2001. Management of lorises in captivity: A husbandry manual for Asian Lorisines (Nycticebus & Loris ssp.) Center for Reproduction of Endangered Species (CRES), Zoological Society of San Diego, USA. Fitch-Snyder, H. & V. Thanh. 2002. A preliminary survey of lorises (Nycticebus ssp.) in northern Vietnam. Asian Primates. 8(1 and 2):1-3.

Primates of The O riental Night

Frederick, C. & D. Fernandes. 1994. Increased activity in a nocturnal primate through lighting manipulation: the case of the Potto Perodicticus potto. Int. Zoo Yearbook 33:219228. Glassman, D. & J. Wells. 1984. Positional and activity behavior in a captive slow loris: a quantitative assessment. Amer J Primatol. 7: 121-132. Izard, M. & K. Weisenseel. 1989. Comparative reproduction of the lorisidae. Amer. J. Primatol. 18:140. Keeling, M. 1974. Housing Requirements. Primates Handbook of Laboratory Animal Science. E. Melby Jr., and N. Altman Eds. Cleveland, Ohio. CRC Press, Inc. Volume I: 97-104. Lester, B. 2001 Mixed Species Housing. In: Management of Lorises in Captivity, A husbandry Manual for Asian Lorisines (Nycticebus & Loris ssp.) H. Fitch-Snyder and H. Schulze eds. Zoological Society of San Diego, San Diego. pp. 88-92. Pariente, G. 1980. Quantitive and qualitative study of light available in the natural biotope of Malagasy prosimians. In: Nocturnal Malagasy primates: ecology, physiology and behavior: P. Charles-Dominique, H. Cooper, A. Hladik, C. Hladic, E. Pages, G. Pariente, A Petter-Rousseaux, A. Schilling, eds. New York: Academic Press. 117-134. Ryley, K. 1913. Bombay Natural History’s Society´s mammal survey of India. J. Bombay Nat. Hist. Soc., 22: 283-295. Schulze, H. 1998. Examples for the slender loris Loris tardigradus nordicus from Ruhr-Universität Bochum, Int. Zoo Yearbook 36: 34-48. Schulze, H. (compiler), H. Fitch-Snyder, C. Groves, K, Nekaris, R. Plesker, K. Petry, M. Singh, & U. Streicher et al. 2003. Loris and potto conservation database. http://www.lorisconservation.org/database/ Schulze, H. & C. Groves, (Submitted January 2004). Asian Lorises: taxonomic problems caused

by illegal trade. In: Proceedings of the International Symposium: Conservation of Primates in Vietnam, at Cuc Phuong National Park, 18. - 20. November. Streicher, U., H. Fitch-Snyder, H. Schulze. 2005. Confiscation, rehabilitation, and placement of slow lorises. This volume. Streicher, U. 2003. Saisonale Veraenderungen in Fellzeichnung und Fellfaerbung beim Zwergplumplori Nycticebus pygmaeus und ihre taxonomische Bedeutung. Zoolog Garten N.F. 73, 6: 368-373. Streicher, U &, H. Schulze, 2002, Seasonal changes in fur pattern and colouration in the pygmy loris (Nycticebus pygmaeus) Caring for primates. Abstracts of the XIXth Congress of the International Primatological Society, Mammalogical society of China Beijing. Tan, C. & J. Drake 2001. Evidence of tree gouging and exudates eating in pygmy slow lorises (Nycticebus pygmaeus). Fol. Primat. 72: 3739. Utami, S., & J. Van Hooff. 1997. Meat eating by adult female Sumatran Orangutans (Pongo Pygmeaus abelii). Amer. J. Primatol. 43:159165. Wiens, F. 2002. Behavior and ecology of wild slow lorises (Nycticebus coucang): Social Organization, infant care system, and diet. PhD dissertation, Bayreuth University, Frankfurt. Wiens, F., A. Zitzmann. 1999. Predation on a wild slow loris (Nycticebus coucang) by a reticulated python (Python reticulatus). Fol. Primat. 70 (6): 362-364. Weisenseel, K., M. Izard, L. Nash, R. Ange, & P. Poorman-Allen. 1998. A comparison of reproduction in two species of Nycticebus. Folia Primatol. 69:321-324. Zhang, Y., G. Quan, T. Zhao, & C. Southwick. 1995. Distribution of primates (except Macaca) in China. Acta Theriol Sin 12:85-95.

135

Primates of The O riental Night

CONFISCATION, REHABILITATION AND PLACEMENT OF SLOW LORISES: RECOMMENDATIONS TO IMPROVE THE HANDLING OF CONFISCATED SLOW LORISES NYCTICEBUS COUCANG Ulrike Streicher1, Helga Schulze2, & Helena Fitch-Snyder3 1 Endangered Primate Rescue Center, Cuc Phuong National Park, Nho Quan District, Ninh Bin, Vietnam 2 Ruhr-University, MA 6/161 b, 44780 Bochum, Germany 3 Loris Conservation International, 5624 Jockey Way, Bonita, California 91902, USA, Email: [email protected]

ABSTRACT Lorises are common in the wildlife trade all over Indonesia. They are traded as pets but also used in traditional medicines and are destined for the national as well as the international market. Large numbers of lorises are exported to the Middle East and Asian countries and confiscations of up to one hundred of animals in a single shipment are known to occur. The care, rehabilitation and final placement of these animals seem currently not addressed in a satisfying way. Animals die due to lack of adequate facilities, diseases, or ill managed releases. Some recommendations to reduce these problems have been developed during the workshop on tarsier and loris taxonomy, husbandry and conservation in Jakarta, 15.-25. February 2003. Materials herein are only brief and it is strongly recommended to access the mentioned sources for more detailed information. Keywords: Slow Loris, Wildlife Trade, Confiscations, Wildlife Rescue.

Note: The following recommendations address specifically the lorises confiscated in animal markets and in the trade (animals with unknown geographical origin). Animals, which are confiscated by rangers directly from hunters (animals with known geographical origin), require a different proceeding. These cases are not addressed here and recommendations for these cases can be obtained from the authors.

INTRODUCTION Lorises occur from China west to India, and south to Indonesia. A number of different species and genera occur in ten different countries. Despite different political structures and economic situations in these countries there is agreement on the protection of lorises. In most countries hunting or capture of lorises is prohibited and the keeping of lorises illegal. Keeping lorises is subject to permits that can usually be issued only by high governmental authorities and these permits are often restricted to specific purposes (e.g. scientific). In some countries there are no specified exceptions and all keeping of lorises is illegal. Fines can be surprisingly high and imprisonment up to 6 years can in theory be imposed. In Indonesia lorises are protected under Decree of Agriculture Ministry No.66 of 1973, the Government Regulation No. 7 of 1999 concerning the Protection of Wild Flora and Fauna and Act No. 5 of 1999 concerning the

Biodiversity Conservation. Consequently all catching, killing, keeping, hurting, transporting and trading of live or dead lorises, parts of their bodies and derivates or products made of them is prohibited. Imprisonment up to five years and a fine up to 100.000.000 rupiah can be implemented in case of violation. Lorises can only be kept with a permission from the Directorate of General Forest Protection and Nature Conservation and only for the purpose of captive breeding. The international trade with lorises is restricted by CITES and most loris species are listed in appendix II. A major task of conservation must be to raise awareness on the legal status of lorises and to encourage strict law enforcement in order to prevent hunting and illegal keeping. Handling of lorises after confiscation Lorises are nocturnal primates and are very susceptible to stress. Direct handling of confiscated

137

Streicher, Schulze & Snyder - Confiscation, Rehabilitation and Placement of Slow Lorises:

animals should be reduced to a minimum and reducing stress must be a main concern during all stages of confiscation and rehabilitation. The environment of the cages should always be quiet and cages should be at least partly covered and provide enough hiding possibilities (branches with foliage, sleeping boxes). Lorises should be handled preferably with gloves for the safety of the people handling them. Little is known about the diseases carried by wild lorises but potentially dangerous viruses can never be excluded. In addition bites of lorises can be toxic and are known to occasionally cause anaphylactic reactions. Transport containers for lorises should be dark, maybe with mesh wire on one of the sides to allow control of the animals. Two layers of mesh wire at a distance of at least one centimetre prevent injuries on the lorises’ hands during transport. Cages should contain some branches with foliage for the animals to hide. Ideally containers should at least measure 300 mm x 300 mm x 300 mm so lorises can be kept in them for several days if necessary. Lorises are fairly solitary animals and should be kept in separate transport containers if possible. However animals that have been together in a shipment or in one cage in a market can be kept in the same transport container. Separating them from a familiar cage mate might increase the stress for the animals, but the number of animals should not exceed four animals per cage/ transport unit. Animals should be transported in a non airconditioned vehicle. Exposing them for several hours to an air-conditioned environment where temperature and humidity are considerably lower than outside can cause severe health problems. Emergency health care After transport and confiscation, lorises should be given some time to recover before they are inspected or handled. After the recovery period a careful inspection without direct handling of the animal should be conducted (see attachments).

138

Common health problems in confiscated lorises include stress, dehydration, injuries, parasites and dental problems. Consequently emergency treatment will comprise the following: 1. Rest and removal of possible stress factors 2. Rehydration, orally or subcutanously 3. Antibiotic treatment 4. Wound hygiene/ dental problems 5. Antiparasitic treatment Details on treatment are included in the attachments. Identification 1. Taxonomic Identification At present knowledge of taxonomy of lorises is fragmentary. The number of species and possible subspecies of slow lorises in Indonesia is uncertain. A classification proposed in 1972 recognised three subspecies: Nycticebus coucang coucang (Sumatra, North Natuna), Nycticebus coucang menagenis (Borneo, Bangka, Belitong) and Nycticebus coucang javanicus (Java). But this arrangement may be too simple and needs to be checked by examination of larger samples of known origin and by molecular and karyotype work. Intensified surveys for lorises are required and the data collection should be standardized in order to allow better comparison of the gathered data. A suggested standard data collection sheet is attached to these materials (see attachments). It is important to emphasize that the data collection on confiscated animals is particularly important in order to assess later placement possibilities but also to identify trade source areas. A preliminary identification key for the different loris subspecies is shown in attachment 4 and a detailed version is currently in preparation. But a wide variety of colourations occur even within the subspecies level and exact taxonomic identification based on the outer appearance might be difficult. A dry hair sample or faecal sample in 100% Ethanol should be taken for genetic identification. The hair sample should be taken with a pair of tweezers or an artery forceps. It is important to assure the hairs are collected with the roots and care must be taken

Primates of The O riental Night

not to touch the hairs in order not to contaminate the sample. Hair samples are subject to CITES regulations and their export and import requires official permits. 2. Individual identification Lorises might be confiscated in large numbers and it is important to mark the animals individually. Implanting a transponder (microchip) is the marking method of choice. If lorises are kept in large groups a visible marking method might additionally be required. Coloured plastic bird rings are widely available and have been used in several institutions successfully. The number of the transponder respectively the colour of the arm ring must be noted on the measure sheet and track must be kept of every single individual from the moment of confiscation until its death or return to the wild. Quarantine and health screening In view of the high risk of primates carrying zoonotic diseases, quarantine requirements are stringent (Woodford 2001). As other primates lorises should be held in quarantine for at least 30 days after arrival. Quarantine cages must be easy to clean and sufficiently large to keep the animal for several weeks. Though hygiene plays an important role during the quarantine period, but in order to reduce stress, nest boxes must also be provided in quarantine cages and cages should be covered with drapes at least on three sides. Animals of one shipment can be quarantined together, but in order to reduce social stress, it is recommended to keep only small groups of animals, which are familiar with each other in the same quarantine cage. However it is difficult to confirm that each individual is eating normally. Animals that have been found in a poor state and/or animals with dental problems should be housed separately during quarantine in order to assure they feed normally. At least one general health check under anaesthesia should be performed during the quarantine period. Health checks in trade confiscated

lorises must include full dental checks, since removing teeth is a common procedure in animals destined for the pet trade. Suggestions on basic treatment and a recommended quarantine protocol are included in the attachments. Long-term placement options 1. Re-introduction Re-introduction is often considered a suitable placement option for confiscated lorises. However there are many concerns related to such a proceeding. The provenance of animals confiscated from the trade and from transports is usually not known. Furthermore the taxonomy of Indonesian lorises is still insufficiently known. Releasing individuals from a different race or sub-species might be a threat to the local population. Little is known about lorises specific adaptations to the ecological requirements of their habitat and individuals released into the wrong habitat type might not have the necessary skills to survive. Due to limited awareness and facilities animals hardly undergo a thorough health check and quarantine period before being returned to the wild. But animals that have been in the transport and trade might carry diseases and if they are returned to the wild they might spread these diseases to the wild populations. Thus the released animals can become a serious health threat to the wild populations. In addition animals might not have the necessary skills to survive in the wild. In most cases the animals are not monitored after the release and nothing is known about their survival in the wild. Returning animals to the wild in a responsible manner might be expensive and might limit the resources available for the conservation of the wild populations (IUCN, 2002). Due to the lack of detailed taxonomic information and the lack of capacity for adequate health care (screening, quarantine) re-introduction into the wild is at present not considered a feasible option (see 2.A). It might be possible in the future and in this respect it is important to assure the genetic resources (see 6.2).

139

Streicher, Schulze & Snyder - Confiscation, Rehabilitation and Placement of Slow Lorises:

2. Captivity Captive options are described in detail in this volume in Fitch-Snyder, H., Schulze, H., Streicher, U.. 2003: Enclosure design. Indonesian Prosimian Workshop. Schmutzer Primate Center, Jakarta, Indonesia. 2.1. Keeping in semi-natural conditions The establishment of semi-wild areas for the permanent keeping of larger groups of confiscated lorises (sanctuary) is an option worth pursuing for the immediate solution of the problem. Animals could be kept in a large fenced area and supplementary fed. There is no experience yet with such a facility for lorises and in particular and optimal feeding management for the lorises has to be developed. Optimal group size and potential social stress factors should be studied in such a facility in order to find suitable group sizes. Detailed recommendations on the establishment and management of such an enclosure can be found in the mentioned document. As long as there are no other placement solutions for confiscated lorises, breeding should be restricted for common lorises’ taxa in a rescue centre situation. The least invasive way to do so is by separating the sexes and housing only all female or all male groups. Though animals can’t be reintroduced at present, they can still be a valuable resource for educational purposes. In addition they provide important study possibilities and they assure that genetic resources are not lost. 2.2. Captive breeding programme At present some taxa of Indonesian lorises do not exist in captivity and little is know about their status in the wild. Confiscated individuals of such insufficiently known taxa should be used to establish a captive breeding population. Such a breeding population can be established and maintained with minimal cost and should not necessarily be used for exhibition.

140

3 Euthanasia Euthanasia might become necessary in some cases. Aspects of animal health and conservational considerations as well as lacking financial and logistical capacity to cope with the number of confiscated lorises suggest euthanasia as feasible option. But it should always be considered the last option and should be only pursued if all other options have been thoroughly explored. Public opinion and political aspects might well be contradictory and care must be taken to carefully justify any case of euthanasia, in order to maintain the credibility of conservation efforts. Further information on lorises First aid, taxonomy, husbandry, measuring standards, diseases: http://www.loris-conservation.org/database/ Quarantine: Woodford, M.H. .2001.Quarantine protocols and health screening protocols for wildlife prior to translocation and release in to the wild. Office International des Epizooties. Paris, France. Diseases: Goeltenboth, R. Kloes, H.-G. . 1995. Krankheiten der Zoound Wildtiere. Blackwell Wissenschaftsverlag, Berlin. Husbandry: Fitch-Snyder, H. and Schulze, H. (eds.). 2001. Management of Lorises in Captivity. A Husbandry Manual for Asian Lorisines (Nycticebus & Loris spp.) Center for Reproduction in Endangered Species (CRES), Zoological Society of San Diego, USA. Fitch-Snyder, H., Schulze, H., Streicher, U. 2003. Enclosure design. Indonesian Prosimian Workshop. Schmutzer Primate Center, Jakarta, Indonesia. In prep. Placement: Baker, L. R. (ed.). 2002. Guidelines for Nonhuman Primate Re-introductions. In Soorae, P.S., Baker, L.R. (eds.). Re-introduction News:

Primates of The O riental Night

Special Primate Issue, Newsletter of the IUCN/SSC Re-introduction Specialist Group, Abu Dhabi, UAE. No. 21: 29-53. IUCN. 2002. Guidelines on the Placement of Confiscated Animals. IUCN/SSC Reintroduction Specialist Group, Abu Dhabi, UAE. IUCN. 1998. IUCN Guidelines for Re-introductions. Prepared by the IUCN/SSC Re-introduction Specialist Group. Abu Dhabi, UAE. ACKNOWLEDGEMENTS Though three authors a signing responsible for this paper, it includes contributions from many participants of the Indonesian Prosimian Workshop in Jakarta in February 2003. Particularly Irene Arboleda, Colin Groves, Sharon Gursky, Alexandra Nietzsch and Myron Shekelle have provided valuable inputs to this paper. Staff of the Schmutzer Primate Center, Pro Fauna Indonesia and WWF Indonesia have contributed in many discussions. Sincere thanks go to LIPI and the Schmutzer Primate Center, which financed and hosted this workshop. Very personal thanks go to Myron Shekelle for his efforts in organizing the workshop. REFERENCES Baker, LR. (ed.) 2002. Guidelines for Nonhuman Primate Re-introductions. In Soorae, PS & LR. Baker. (eds.). Re-introduction News: Special Primate Issue, Newsletter of the IUCN/SSC Re-introduction Specialist Group, Abu Dhabi, UAE. No. 21: 29-53 Fitch-Snyder, H. (eds.) 2001. Management of Lorises in Captivity.

A Husbandry Manual for Asian Lorisines (Nycticebus & Loris ssp.) Center for Reproduction in Endangered Species (CRES), Zoological Society of San Diego, USA. Gass, H. 1987. Affen. In: K. Grabisch & P. Zwart (eds.): Krankheiten der Wildtiere. Schluetersche Verlagsanstalt, Hannover, Germany. Goeltenboth, R & HG. Kloss.. 1995. Krankheiten der Zoound Wildtiere. Blackwell Wissenschaftsverlag, Berlin, Germany. IUCN. 2002. Guidelines on the Placement of Confiscated Animals. IUCN/SSC Re-introduction Specialist Group, Abu Dhabi, UAE. IUCN. 1998. IUCN Guidelines for Re-introductions. Prepared by the IUCN/SSC Re-introduction Specialist Group. Abu Dhabi, UAE. Schulze, H. & H. Fitch-Snyder. 2001. Conservation database for lorises and pottos. http:// www.loris-conservation.org/database/ Sutherland-Smith, M. 2001. Review of Loris Clinical information and Pathological Data From The San Diego Zoo:1982-1995. In: Management of Lorises in Captivity. A Husbandry Manual for Asian Lorisines (Nycticebus & Loris ssp.). Fitch-Snyder, H. and Schulze, H. (eds.). 2001. Center for Reproduction in Endangered Species (CRES), Zoological Society of San Diego, USA. Wiens, F. 1995. Verhaltensbeobachtungen am Plumplori Nycticebus coucang (Primates: Lorisidae) im Freiland. Diplomarbeit. Johann Wolfgang Goethe-Universitaet, Frankfurt a.M., Germany. Woodford, MH. 2001.Quarantine protocols and health screening protocols for wildlife prior to translocation and release in to the wild. Office International des Epizooties. Paris, France.

141

Streicher, Schulze & Snyder - Confiscation, Rehabilitation and Placement of Slow Lorises:

Attachment 1- Performing a basic health check Healthy animals should be alert, when disturbed Signs of diseases: Animal seems very sleepy Animals hold the head wobbly or not straight. Note: Freezing to motionlessness may be normal camouflage behaviour, indicating stress The eyes should be bright and look around actively, when disturbed Signs of diseases: There is blood, pus or other secretion around the eye. Animal is unable to open both eyes. There is a swelling in the surrounding of the eye. The eye seems dull, greyish or sunken. The nostrils should be clean and the animal breathes normally Signs of diseases: There is blood, pus or other secretion around the nostrils. The animal is breathing very heavily. The animal makes sounds while breathing. Note: Feeling threatened lorises might utter growls, that must not be misinterpreted as raspy breathing and a sign of disease. The muzzle should be symmetrical and the animal should feed normally Signs of diseases: Swollen or enlarged muzzle. Blood or pus on the visible gums. Difficulties to bite and chew. The ears should be symmetrical and clean Signs of diseases: There is blood, pus or other secretion from one or both ears. One ear or both ears hanging not normally There are injuries or swellings around the ear. They should have a healthy looking coat Signs of diseases: There are injuries or wounds There are bald patches The animal shows excessive scratching. They should have a normal body shape and move normally Signs of diseases: There are swellings or lumps The position of one or more limbs is abnormal. There is lack of function of a limb Note: During daytime lorises will present themselves mostly tightly curled up, holding on to the cage or the cage furnishing. Thus it might be difficult to correctly assess the reactions of the animal. Repeating the inspection of the animal at night with the help of a torch might be necessary.

142

Primates of The O riental Night

Animal access sheet Access data: Taxon name:

Identification number:

Date of birth (est.):

Sex:

Weight:

Source:

Arrival date: ___________________________________________________________________ External check on arrival (not under anaesthesia): Nutrition state: Care state: Skin: Eyes: Nose: Mouth: Movements: Digestive system: Respiratory system: Circulatory system: ___________________________________________________________________ Emergency measures:

___________________________________________________________________ Feeding instructions:

___________________________________________________________________ Duration of quarantine:

143

Streicher, Schulze & Snyder - Confiscation, Rehabilitation and Placement of Slow Lorises:

Attachment 2 – Emergency care of slow lorises Emergency care The below mentioned drugs are only some suggestions out of a variety of possibilities. Rehydration:

Oral rehydration solution or Ringer’s lactate orally or subcutaneous (at body temperature), up to 10 % of the bodyweight distributed over the day

Antibiotic treatment:

Antibiotic treatment is required in case of obvious (wounds, gingivitis, pneumonia) infections in order to prevent further spreading of bacteria. Enrofloxacin (Baytril®) has been used with good results at a dosage of 5mg/kg intramuscular.

Wound hygiene:

Wounds should be cleaned with clean water and iodine solution (Betadyne®). Application of antibiotic creams is not recommended, since lorises lick themselves thoroughly and will ingest the cream. Wound in trade animals will in most cases not be fresh and aggressive treatment (e.g. suturing) might be delayed for several days until the health check under anaesthesia.

Dental problems:

Dental problems occur mostly in animals from the pet trade, in which teeth have been removed. These animals might suffer from severe gingivitis and sinusitis. Antibiotic treatment is necessary in order to control the infection. In addition such animals might be reluctant to feed on hard or firm food items and might require specific food preparation (blander). In some cases removal of teeth or tooth fragments might be necessary but this can only be done under anaesthesia.

Antiparasitic treatment:

Ivermectin (Ivomec®) 0.2-0.4 mg/kg (Goeltenboth et al., 1990)

144

Primates of The O riental Night

Attachment 3 – Quarantine and health screening in slow lorises Quarantine Based on international guidelines (Woodford, 2001) 30 days quarantine with the following protocol is recommended. During anaesthesia at least one full health check must be performed under anaesthesia. 1.

Faecal examination (direct and flotation) for endoparasites, especially Entamoeba sp. which often infect primates, causing diarrhoea in animals subjected to stress. Since Entamoeba sp. are shed intermittently, several samples should be examined.

2.

Check for ectoparasites.

3.

Appropriate serology, based on history and origin. Health screening for lorises might include screening for Hepatitis B and C and Herpes simplex and Herpes B viruses.

4.

Intracutaneous tests for tuberculosis (using avian, bovine or mammalian tuberculine). In most primates this test is routinely performed in the eyelid. However this location might in lorises be difficult to monitor and a shaved spot on the flank or abdomen might be preferable.

5.

Complete blood chemistry

Note:

False positive tuberculine reactions have been reported from lorises at San Diego Zoo. A new Herpes virus variety has been identified in lorises at the same facility. Not all rescue facilities might be able to perform all these tests for several reasons. Thus the quarantine period should be absolutely strictly kept.

Anaesthesia Anaesthesia should only be performed by qualified veterinarians. Ketamine is most commonly used for anaesthesia in lorises. Recommended dosage:

23 mg/kg intramuscular Literature: 5 mg/kg (Gass, 1978) 8-12 mg/kg (Sutherland-Smith et al., 2000) 20 mg/kg (Goeltenboth et al., 1990) 11.3-33.3 mg/kg (Wiens, 1995) Note: females might require lower dosages than males

Health check in anaesthesia: Taxon name:

Identification number:

Date of birth (est.):

Sex:

Weight:

Origin:

Arrival date:

Way of application:

145

DEDICATION

This book is dedicated to Yakob Muskita, one of the world’s few tarsier field biologists, who passed away on February 13, 2003, a few days before the workshop began, from complications of diabetes at the age of thirty-eight. He is survived by his wife and two children.

Yopie Muskita, on his wedding day, in Batuputih, North Sulawesi, gateway to Tangkoko Nature Reserve. Photo Myron Shekelle (c) 1996.