Novltates PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

Home , Crunomys, Echiothrix, Haeromys, Hapalomys, Malayan tree rat, Murinae

AMERICANt MUSEUM Novltates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2989, 18 pp., 10 illustrations, 1 table October 25, 1990

Sulawesi Rodents: Species Traits and Chromosomes of Haeromys minahassae and Echiothrix leucura (Muridae: Murinae)

GUY G. MUSSER'

ABSTRACT Results from analyses of chromosomal spreads pairs, two telocentric pairs, and three small met- ofa male tree mouse, Haeromys minahassae, and acentric pairs; a small submetacentric X chro- a male shrew rat, Echiothrix leucura, are described mosome; and small telocentric Y. Although the and illustrated. A 2N of 48, FN of 54, autosomes karyotypes constitute additional sets of data to consisting of 21 pairs oftelocentric chromosomes that provided by morphology of skins, skulls, and and two pairs ofsmall metacentrics, and sex chro- teeth (also briefly described here), the information mosomes made up of a large submetacentric X they provide on the phylogenetic relationships of and smaller submetacentric Y characterize the either species is ambiguous. Rather than indicat- karyotype of H. minahassae. The chromosomal ing closest relatives, these chromosomes may con- set of E. leucura encompasses a 2N of 40; FN of tribute more to polarizing karyotypic characters 75; an autosomal component of 14 submetacentric in certain groups of Indo-Australian murids. INTRODUCTION The small tree mouse, Haeromys mina- countered by collectors -a reflection of the hassae (fig. 1), and large shrew rat, Echiothrix arboreal life style and spotty distribution of leucura (fig. 3), are among the more than 40 the granivorous tree mouse, and the special- species of murid rodents which occur on Su- ized and secretive habits of the vermivorous lawesi and nowhere else. Both species are rep- shrew rat. resented by samples collected only in the Phylogenetic relationships of each species northern and central parts ofthe island, have have yet to be determined. Haeromys mi- been found only in tropical lowland ever- nahassae is related to other species of Hae- green rain forest, and are infrequently en- romys which are found on the Sundaic is- I Archbold Curator, Department of Mammalogy, American Museum of Natural History.

Copyright © American Museum of Natural History 1990 ISSN 0003-0082 / Price $2.60 2 AMERICAN MUSEUM NOVITATES NO. 2989 lands of Palawan and Borneo. The genus is Karyotypes were prepared and stained in characterized by retention ofmany primitive the field from chromosomes ofbone marrow, morphological traits and a few derived fea- and processed by employing colchicine, hy- tures. Some of these specializations are also potonic citrate, and flame-drying, procedures shared by species of Chiropodomys, arboreal that were described by Patton (1967). mice indigenous to Indochina and islands on I use four terms to describe the shape of the Sunda Shelf (Musser, 1979; Musser and each chromosome relative to position of the Newcomb, 1983), but this link is weak and centromere: metacentric (the chromosome is whether or not it reflects close relationship biarmed, one arm being about the same length has to be tested with additional data. In what as the other); submetacentric (one arm is has been documented ofits morphology based shorter than the other, about a third of its on museum study skins, crania and mandi- length); subtelocentric (one arm is very short bles, and dentitions, E. leucura retains some relative to the long arm on the other side of primitive characters but most of its diagnos- the centromere); and telocentric (the centro- tic traits are derived. Its closest relative is mere is at the tip of the chromosome or near unknown, and the species seems as isolated enough that any portion on the other side of within a phylogenetic framework as it is in the centromere is so short as to be indistin- its geographic distribution (Misonne, 1969; guishable or nearly so). Musser, 1969). I will try to clarify these muddled views in a systematic revision of Haeromys (Musser, ACKNOWLEDGMENTS in prep.) and a morphological analysis of the Curators at the institutions cited above al- specialized traits peculiar to Echiothrix lowed me access to collections under their (Musser and Sawitzke, in prep.). Here I pre- care and loaned me specimens; I appreciate sent results from analyses of a different data their generous help. set-chromosomal number and morpholo- I am grateful to Charles J. Cole and Carol gy-and discuss its significance for inferring Townsend who taught me the techniques re- phylogenetic relationships between these two quired to obtain and preserve chromosome Sulawesian endemics and other elements of spreads, procedures I could use in isolated the murine fauna native to the Indo-Austra- forest camps. When I returned with my boxes lian region. First, I will introduce the rodents ofslides, Jay allowed full access to equipment by presenting brief descriptions of their in- and supplies in his laboratory; he and Carol sular distributions, morphological character- tutored me in the craft of extracting the sig- istics, forest habitats, and habits. nificant data pressed between glass slide and cover slip. The chromosomal spreads were photographed by Cameron Newcomb. SPECIMENS AND METHODS The manuscript has been strengthened by Specimens cited here are in collections of the intelligent criticisms and contributions of the American Museum of Natural History, Audrone Biknevicius, Charles J. Cole, Law- New York (AMNH); the British Museum rence R. Heaney, and James L. Patton; I thank (Natural History), London (BM); the Muse- them for their time and effort. um Zoologicum Bogoriense, Bogor (MZB); My work in Sulawesi was financially sup- the Rijksmuseum van Natuurlijke Historie, ported by the Celebes Fund ofthe American Leiden (RMNH); and the National Museum Museum of Natural History and Archbold of Natural History of the Smithsonian Insti- Expeditions Inc. I lived in Indonesia under tution, Washington, D.C. (USNM). the auspices of the Lembaga Ilmu Pengeta- I measured external dimensions-total, tail, huan Indonesia and the Museum Zoologi- hind foot, and ear lengths-and took weights cum Bogoriense. The Director of the Muse- shortly after the animals were captured; length um at that time, Dr. Sampurno Kadarsan, ofhead and body was derived by subtracting was my amiable and gracious host as well as tail length from total length. Crania and mo- sponsor. lar rows were measured by me in the labo- This report is Number 120 in Results of ratory. the Archbold Expeditions. 1 990 MUSSER: SULAWESI RODENTS 3

250 and 500 ft in the Malakosa region of ANIMALS AND HABITATS central Sulawesi (see the map in Musser and Haeromys minahassae is the smallest na- Dagosto, 1987: 30). Two specimens from for- tive murid living on Sulawesi (fig. 1; table 1). est along the Sungai Sadaunta (1°22'S, Its short head and body-about as long and 119°59'E) at 3050 and 3300 ft (AMNH thick as my thumb -is clothed by a soft taw- 226817 and 226816) constitute the only oth- ny brown dorsal coat, a white or whitish gray er records from the central part ofthe island. ventral coat, and counterbalanced by a long Elsewhere in Sulawesi, H. minahassae has and brown whiplike tail. been collected in the northeastern arm, but The tree mouse has a gracile cranium (fig. nowhere else. One example, the holotype (BM 5) in which the rostrum is short and wide; 97.1.2.39), comes from Rurukan (1°21'N, the zygomatic plates are narrow, and the zy- 124°52'E); two others (RMNH 21159 and gomatic notches very shallow; the interorbit 21160) were taken near Tumaratus (1°09'N, is wide; the braincase is globular, its dorso- 124°48'E) at 700-800 m; and another (MZB lateral margins weakly defined by indistinct 4840) was obtained from 700-800 m near ridges; the zygomatic arches are slender and Tomohon (1°19'N, 124°49'E). delicate; the incisive foramina are short, as Haeromys minahassae is rarely encoun- is the wide bony palate; the sphenopalatine tered by biologists and infrequently collected, vacuities are small, set in a wide mesopter- possibly because of its arboreal habits and ygoid fossa; the pterygoid plates are wide; and patchy distribution within the forest. As in- the auditory bullae are small. The mandible dicated by specimens I have studied, the al- is stocky, and the ramus is nearly square in titudinal distribution of the tree mouse ex- side view; the coronoid is a low projection, tends from 250 to 3300 ft, a range contained as are the condylar and angular processes. within the elevational brackets of tropical Incisors are short and orange, the molars lowland evergreen rain forest on Sulawesi. In low-crowned and white. Coronal cusp pat- this type of forest along Kuala Navusu, rel- terns are simple. A cusp is retained on the ative humidity is about 100 percent and am- anterolabial corner of each second and third bient air temperatures are moderately high. upper molar, a prominent posterior cingulum For example, during 95 days in 1975, the forms the back ofeach first and second upper period when the mice were collected, the av- tooth, an anterocentral cusp connects large erage minimum air temperature was 73.6°F anterolingual and anterolabial cusps at the (range, 72-76°) and the average maximum front of each first lower molar, and labial temperature was 80.9°F (range, 74-87o). The cusplets are few (fig. 7). Three roots anchor canopy is 20-40 m high, with emergent trees each upper molar, two roots each lower tooth, reaching 50-70 m. The high diversity of spe- which is a primitive condition (Musser and cies includes not only the tall canopy-formers Newcomb, 1983). but understory trees, shrubs, palms, and Characteristics of the Sulawesian tree rat woody vines as well; Musser and Dagosto are generally representative ofthe genus Hae- (1987: 41-44) provided a fuller description. romys. Morphological aspects of the group Haeromys minahassae was collected dur- are a composite of many retained primitive ing the night in traps set above ground on traits and a few derived ones (three pairs of woody vines in tangles like those shown in mammae, whiplike tail, large opening be- figure 2. The scansorial facility of this tree tween bullae and squamosal, spacious sphe- mouse (fig. 1)-its ability to climb and scam- nopterygoid vacuities, a ridge bounding the per along vines and other cylindrical sub- pterygoid bridge between foramen ovale and strates above ground-is impressive. In this bulla, and cusplets on anterior margins offirst understory beneath the canopy, the mice have molars are derived examples; Musser and access to fruits of herau palms (Areca ves- Newcomb, 1983). tiaria) and species offigs (Ficus), which form The specimen ofH. minahassae from which a large part oftheir diet. The pulp ofthe herau the chromosomal spreads were obtained is fruit is consumed, but pulp offigs is discarded part of a sample (table 1) collected from the and only the seeds are eaten. Other frugivo- forested slopes above Kuala Navusu between rous species of rats living on Sulawesi also 4 AMERICAN MUSEUM NOVITATES NO. 2989

71h

Fig. 1. An adult female Haeromys minahassae (AMNH 226044), close to natural size. Caught October 17, 1975, at 250 ft along Kuala Navusu, central Sulawesi. The long, flexible tail is a scansorial special- ization. 1 990 MUSSER: SULAWESI RODENTS 5

Fig. 2. Woody vines on which Haeromys minahassae were caught (arrows) in the understory of tropical lowland evergreen rain forest along Kuala Navusu, 500 ft. The tangles are loosely coiled about trunks of fig trees (Ficus spp.) and hang from their crowns by viney suspensions. Seeds from fig fruit are a primary dietary component of the tree mouse. Photographed October, 1975. 6 AMERICAN MUSEUM NOVITATES NO. 2989

Fig. 3. A young adult male Echiothrix leucura (AMNH 225682), about two-thirds natural size. Caught October 15, 1975, at 400 ft above Kuala Navusu, central Sulawesi. The elongate head, large eyes, long and pointed ears, spiny fur, delicate front feet, long and robust hind feet, and long sturdy tail are distinctive characteristics. eat figs, but they digest the pulp and pass the samples from the Malakosa region and Sung- intact seeds in their feces; H. minahassae is ai Sadaunta includes the features described distinguished from those fig consumers as the for the holotype ofMus minahassae (Thom- only seed predator (Musser, personal obs. and as, 1896). I compared specimens from my fieldnotes). series directly with the holotype of minahas- The range in variation of morphological sae and found their characteristics to be sim- and chromatic traits that characterize the ilar. Venter coloration is the only conspicu- 1 990 MUSSER: SULAWESI RODENTS 7

>, -r n 4

'{ o FE't.A' 'S |r .- Ad> .i *

. "

w .,O. P416 VH

-Ar 31

I .. A % Lrs.

Fig. 4. Tropical lowland evergreen rain forest above Kuala Navusu, 400 ft, the habitat ofEchiothrix leucura. The broad leaves ofa young palm, Livistona rotundifolia, partially shade the leaf-covered forest floor. Photographed November, 1975. 8 AMERICAN MUSEUM NOVITATES NO. 2989

Fig. 5. Cranium and mandible of an adult Haeromys minahassae (AMNH 226047) from Kuala Navusu in the Malakosa region of central Sulawesi. About x 3. ous difference between the samples from ied species which is native to Sulawesi (fig. central Sulawesi and those from the north- 3; table 1). Its elongate head and body are eastern arm: chest and abdomen ofthe latter covered by a bristly, variegated gray and are whitish gray rather than white. Evidence brown dorsal pelage, and a white (stained with from museum skins and skulls convinces me russet in some individuals) semispinous ven- that the samples from these two regions of tral coat. The dark brown of its very large the island represent the same species. and soft ears contrasts with the predomi- Echiothrix leucura is one ofthe large-bod- nantly gray and spiny pelage. Its front legs 1990 MUSSER: SULAWESI RODENTS 9

1.1 , .1 .. LAI Voi .4 ,, t M:,; '4 Adw,.,.0. .I wt

Fig. 6. Cranium and mandible ofan adult Echiothrix leucura (AMNH 225681) from Kuala Navusu. About x 2. 10 AMERICAN MUSEUM NOVITATES NO. 2989

4.*is

-4 f

4, ,r--

-*..Iif /_,,

Fig. 7. Occlusal views of right upper (left) and lower (right) molar rows ofHaeromys minahassae (AMNH 226048; CLM1-3 = 3.3 mm, clml-3 = 3.4 mm) from Kuala Navusu. Fig. 8. Occlusal views ofright upper and lower molar rows ofEchiothrix leucura (AMNH 225681; CLM 1-3 and clm 1-3 each = 6.5 mm) from Kuala are thin and the feet small and delicate. The Navusu. hind legs are short and strong, the feet very large and very long, with short outer digits and long central ones. Its thick tail is longer flects the protracted configuration of the cra- than the head and body and white every- nium. where except for the proximal one-third to Teeth are unique. The cream-colored up- one-half, which is dark gray on dorsal and per incisors are short and delicate, and their ventral surfaces. Females have two pairs of front faces are scored by either one or two inguinal teats, no more, no less. shallow grooves. Lower incisors are also The distinctive cranium of the shrew rat cream colored, very long and thin, and very (fig. 6) has a high, squarish, and slightly ridged sharp. Molars are very small in relation to braincase; a wide interorbit; graceful, elon- the surface areas ofbony palate and dentaries. gate rostrum with stretched incisive forami- Their crowns are low and the coronal cusp na; gracile zygomatic arches; a long palatal patterns simple (fig. 8). The low cusps are bridge; moderately large sphenopalatine va- discrete in young animals but wear quickly cuities; a simplified pterygoid region in which with advancing age to form basined occlusal much of the pterygoid shelf is gone, and the surfaces. Upper molars are anchored by three surfaces for muscle attachment significantly roots under each tooth, lowers by two. reduced; and moderately large auditory bul- Chromosomal spreads were obtained from lae. The elongate and gracile mandible, with a specimen in a sample of eight individuals its low coronoid projection and caudally di- collected in the Malakosa area between 100 rected condylar and angular processes, re- and 400 ft along the Kuala Navusu (table 1). 1 990 MUSSER: SULAWESI RODENTS I1I

TABLE 1 Selected Measurements (in millimeters) and Weight (in grams) from Samples of Haeromys minahassae and Echiothrix leucura Collected Along Kuala Navusu in the Malakosa Region of Central Sulawesia Species and individual Sex Age LHB LT LHF LE WT GLS CLMI-3 H. minahassae AMNH 226043b M A 78 131 21 15 14 - - AMNH 226047 F A 70 125 20 15 13 23.4 3.1 AMNH 226051 F A 75 128 19 15 15 23.7 3.1 AMNH 226050 M A 74 133 19 16 15 23.5 3.1 AMNH 226045 F A 76 127 19 15 13 23.3 3.1 AMNH 226044 F A 75 120 19 16 12 22.7 3.1 AMNH 226048 F YA 70 120 20 15 11 22.9 3.3 AMNH 226049 F J 64 111 19 14 10 21.1 3.1 AMNH 226046 M J 62 104 18 14 7 20.1 3.1 E. leucura AMNH 225684 M OA 225 258 53 33 310 54.1 6.0 AMNH 225681 M A 219 242 51 34 290 53.6 6.2 AMNH 225683 F A 221 245 52 34 275 52.8 6.4 AMNH 225685 F A 221 241 53 34 293 54.2 - AMNH 225682b M A 227 - 53 33 - - - AMNH 225679 F YA 208 247 52 33 248 51.5 6.4 AMNH 225678 M YA 213 253 53 32 235 52.4 6.3 AMNH 225680 M YA 200 235 52 33 220 50.2 6.4 a Abbreviations: M, male; F, female; OA, old adult; A, adult; YA, young adult; J, juvenile; LHB, length of head and body; LT, length of tail; LHF, length of hind foot; LE, length of ear; WT, weight; GLS, greatest length of skull; CLM 1-3, crown length of maxillary molar row. b Karyotyped. Morphology of skins and skulls in this series and between them and leucura from the is like that described for specimens diagnosed northeastern arm of the island (east of Bum- and named Echiothrix brevicula by Miller and bulan, 0029'N, 122053'E), involve size and Hollister (1921). The 13 individuals upon pelage coloration. These differences have not which this taxon was based were collected at always been interpreted as those which de- Pinedapa (1°25'S, 120°35'E), 100 ft, which is limit species, for in their list of Indo-Austra- south ofthe Malakosa region and in the same lian mammals, Laurie and Hill (1954) re- forest formation (see the map in Musser and garded brevicula and centrosa as subspecies Dagosto, 1987: 30). My comparison of the ofEchiothrix leucura. In identifying the sam- rats from Kuala Navusu with the holotype of ple from Kuala Navusu as E. leucura, I have brevicula (USNM 219744) and the rest ofthe followed their arrangement. Whether or not Pinedapa series confirms this identity. the morphological differences among sam- In contrast to the meager samples ofHae- ples represent geographic variation within one romys minahassae, the shrew rat is repre- species or define limits of more than one resented in museums by larger series collected productively isolated entity cannot be deter- at more places. Still, all the localities are only mined without a rigorous systematic revision in the northern arm and central core of Su- of the genus. lawesi. Sufficient geographic variation in cra- Echiothrix leucura is another inhabitant of nial and chromatic characters exists, enough tropical lowland evergreen rain forest; all to have stimulated Miller and Hollister (1921) specimens I have studied come from locales to not only name E. brevicula, but also E. between 100 and 3600 ft. The shrew rat is centrosa, which was founded on specimens terrestrial and nocturnal. I trapped it on the from higher elevations in central Sulawesi. forest floor (fig. 4), in runs alongside fallen The distinctions between those two forms, tree trunks, and on rotting limbs and trunks 12 AMERICAN MUSEUM NOVITATES NO. 2989 that lay across ravines and streams. Earth- logenetic relationships between this cluster worms are the primary dietary component of and other genera are unclear. Tate (1936) sug- E. leucura. The morphological specializa- gested that Haeromys was a derivative ofthe tions in the body conformation, as well as in Rattus cremoriventer group. Initially, Eller- skull and teeth, reflect its ability to forage for man (1941, 1949) had also allied Haeromys earthworms, run along substrates, leap away with Rattus but later considered it to be more from predators, and consume soft-bodied in- closely related to Lorentzimys of New Guin- vertebrates. ea. Misonne (1969) was unsure about the af- finities of Haeromys but included it within his Lenothrix-Parapodemus Division which, CHROMOSOMES in the context ofhis inquiry, meant that Hae- Haeromys minahassae. - Chromosomal romys was not closely allied to Rattus or the spreads were obtained from a male (AMNH cremoriventergroup or any other cluster, only 226043) collected in the Malakosa area of "that it was one ofthose genera that are relict central Sulawesi along the Kuala Navusu at in some way, arboreal to some degree, lives 250 ft (0058S5, 120'27'E). Twenty-five cells in high mountains or are isolated on islands, were studied. The diploid number is 48; the and dwell in deep forest" (Musser and New- fundamental number is 54 (fig. 9). The au- comb, 1983: 559). tosomal set consists of 21 pairs oftelocentric In a study of relationships among Sundaic chromosomes that range in size from a lapge murines that was based on morphological first pair to succeeding smaller pairs in a grad- characters derived from skins, skulls, and ed series, and two pairs ofsmall metacentrics. teeth, Musser and Newcomb (1983: 557) re- The presumed X chromosome (largest chro- jected any close ties between Haeromys and mosome of the heteromorphic pair) is sub- either Rattus or Niviventer (which contains metacentric and also the largest chromosome the cremoriventer group). Among the rats and in the karyotype; the Y chromosome is a mice known to occur on the Sunda Shelf, smaller submetacentric. Haeromys, they suggested, may be more Echiothrix leucura. - Spreads were ob- closely related to Chiropodomys than to any tained from a male (AMNH 225682) trapped other genus. Even between those two the link along the Kuala Navusu at 400 ft. Twenty was weak, as the hypothesized phylogenetic cells were examined. The diploid number is alliances need to be further tested with data 40; the fundamental number is 75 (presum- from other characters and Haeromys should ably 76 in females; fig. 10). The autosomes be compared with species occurring outside consist of 14 pairs ofsubmetacentrics in which the Sundaic region. What may have been the first pair is the largest and the others range Musser and Newcomb's primary contribu- in size from moderately large to small, two tion was demonstrating that the very ambi- pairs of small telocentrics, and three pairs of guity characterizing Haeromys's phylogenet- small metacentrics. The presumed X chro- ic position stemmed from its retention of so mosome is a small submetacentric; the Y a many primitive characters. Of all the traits small telocentric. associated with skins, skulls, and dentitions to which polarities could be hypothesized, the fewest derived features of any Sundaic SIGNIFICANCE species are possessed by Haeromys. The many The chromosomal data provide another set primitive traits retained by species of Hae- of characters that might point to the closest romys are shared by other species that are phylogenetic relatives of H. minahassae and native not only to the Sunda Shelf but also E. leucura. But do they? to archipelagos and continents outside that Morphology of museum skins and associ- area. It was this primitive suite that formed ated skulls and dentitions of the Sulawesian the foundation upon which the conflicting Haeromys clearly tie it to the Bomean H. statements of relationships were constructed pusillus and H. margaretae as well as a Pa- by Tate (1936), Ellerman (1941, 1949), and lawan species, which together form a mon- Misonne (1969). Shared specializations sug- ophyletic group (Musser, in prep.). The phy- gest Haeromys to be phylogenetically closer 1 990 MUSSER: SULAWESI RODENTS 13 to Chiropodomys than to any other Sundaic minahassae. The diploid number is 42 and murid (Musser and Newcomb, 1983). the fundamental number is 46 for M. bec- If the karyotype derived from one speci- carii, and its chromosomal package consists men ofH. minahassae is even generally rep- of 19 pairs of telocentric chromosomes and resentative of the genus, the chromosomal two pairs of small metacentrics (sex chro- information-at the level of gross morphol- mosomes have yet to be identified). The first ogy as well as diploid and fundamental num- pair of telocentrics is the largest in the kar- bers-does not sharpen the resolution of yotype and the others decrease in size with Haeromys's position within the diversity of pair 19 the smallest. Aside from the differ- Asian murines, and only reinforces the view ences in diploid and fundamental numbers that it is characterized by the retention of (48 and 54, respectively, in H. minahassae), many primitive traits and few derived ones. and a possible difference in conformation of In its high diploid number, low fundamental sex chromosomes, the combination ofall tel- number, and its composition ofbasically telocentric and two small metacentric pairs in ocentric pairs of chromosomes, the karyo- M. beccarii is morphologically more remi- type represents what is likely a primitive niscent of H. minahassae than is the karyo- chromosomal condition, assuming that high type of Chiropodomys gliroides, which has number of telocentrics and low number of telocentric autosomes but either no small biarmed elements is more primitive than a metacentric pairs (Thailand) or small meta- complement characterized by mostly biarmed centric and submetacentric pairs (Malay Pen- elements (for example, see Viegas-Peuignot insula). et al., 1983). However, except for being arboreal and in- This presumed primitive state defined by digenous to Sulawesi, M. beccarii and H. a high number of acrocentric autosomes is minahassae are not especially closely related. the primary karyotypic feature that is shared No derived morphological traits (as seen in by Haeromys, as represented by the Sulawe- skins, skulls, and dentitions) are shared by sian H. minahassae, and Chiropodomys, as the two species, and the characters which are sampled by Thai and Malay specimens of C. held in common are primitive ones (compare gliroides. Chromosomal spreads of Thai C. table 11 in Musser, 1981: 324, with table 38 gliroides consist of 21 pairs of telocentric in Musser and Newcomb, 1983: 536). A kar- chromosomes that grade in size from large yotype in which all but two pairs are telo- to small; sex chromosomes have not been centric, which is common to the two species, identified; and both diploid and fundamental is possibly another shared primitive feature. numbers are the same, 42 (Marshall, 1977; In its diploid number and gross morphol- Tsuchiya et al., 1979). In samples of C. gli- ogy ofchromosomes, the autosomal comple- roides from the Malay Peninsula, spreads ment ofH. minahassae appears more similar contain a diploid number of 42, a funda- to the non-Rattus native Australian murines mental number of 48, 18 pairs of telocentric than to Chiropodomys or any other murine chromosomes, one pair of small metacen- indigenous to Indochina, the Sunda Shelf, or trics, one pair of small submetacentrics, a Sulawesi. The euchromatic autosomal com- large metacentric X chromosome, and a me- plements of species of Pseudomys, Leggadi- dium-size submetacentric Y (Yong, 1973, na, Notomys, Melomys, and Hydromys, as 1983). Haeromys and the Malayan Chiro- shown by Baverstock et al. (1977), have four podomys also share a large biarmed X chro- characteristics in common: (1) a diploid mosome, which is the largest in the comple- number of 48, (2) the first telocentric pair is ment. But it is metacentric in Chiropodomys "distinguishably" larger than any ofthe other (a "duplicate-type X chromosome," as Yong, pairs, (3) autosomal pairs 2-21 are telocentric 1973: 61, described it) and unlike the biarmed and grade in size from large to small, and (4) configuration of the X chromosome in H. there are two pairs of small metacentric au- minahassae. tosomes (pairs 22 and 23). The karyotype of Margaretamys beccarli The chromosomal characteristics shared (Musser, 1981: 288), a tree rat endemic to among the Australian species could "be due Sulawesi, also resembles that of Haeromys to karyotypic convergence," according to 14 AMERICAN MUSEUM NOVITATES NO. 2989 QnODIOgRO~^ |1 An no- 4)n nn 1

*fl *f 00 so eniflAne itA ARt

to a 4A 22 23 O, x Y Fig. 9. Karyotype from a male Haeromys minahassae AMNH 226043): 2N = 48, FN = 54. The first autosomal pair is the largest ofthe telocentrics, and pairs 22 and 23 are small metacentrics. Additional explanation is provided in the text.

Baverstock et al. (1977: 1 15), but that seemed Haeromys minahassae and the Australian unlikely to them and they argued "that the species indicate close relationship and com- common and widespread occurrence of such mon phyletic origin? Possibly. However, an a distinctive karyotype in the same zoogeo- explanation more plausible to me is that the graphic region (Australasia) emphasizes ho- karyotype of H. minahassae is conservative, mology." Furthermore, the species "appear retaining primitive characteristics which are to have a common phyletic origin in Aus- also found in most ofthe indigenous Austra- tralasia, and there seems little doubt that a lian rats and mice. If a diploid number of 48 similar karyotype must have been present in and a set ofchromosomes consisting ofmost- the ancestor of the entire group." Some of ly telocentrics and two pairs of small meta- the other Australian species (excluding Rat- centrics is primitive for Indo-Australian mu- tus) that were sampled exhibited karyotypes rines, it would not be surprising to find such that Baverstock and his colleagues suggested a karyotype shared among different species were modified from this presumed ancestral scattered throughout the region from South- configuration by what they believed to be fixed east Asia to New Guinea and Australia. A pericentric inversions. similar karyotype characterizes Hapalomys Haeromys minahassae has a diploid num- longicaudatus from the Malay Peninsula ber of 48, the first pair of telocentrics appear (Yong et al., 1982), for example. And it may slightly larger than the others, autosomal pairs not be a surprise to find it in at least one 2 through 21 are telocentric and form a grad- species ofHaeromys, a genus with a relictual ed series of decreasing size, and pairs 22 and distribution in tropical forests on the Sunda 23 are small metacentrics. The resemblance Shelf and Sulawesi and one that is charac- between Haeromys and the Australian cluster terized by few specialized features (Musser seems impressive if possession of identical and Newcomb, 1983). diploid numbers is significant. If such an Chromosomal data turn out to be ambig- identity is not especially significant, then the uous in detecting the nearest phylogenetic rel- resemblance between karyotypes of Marga- ative ofHaeromys. But the information from retamys beccarii and H. minahassae (2N of the one specimen of H. minahassae does 42 versus 2N of 48) is just as noteworthy. point to the next set of inquiries necessary to Do the karyotypic similarities between determine the real significance of the kary- 1 990 MUSSER: SULAWESI RODENTS 15 A^4 AA aa i^ Al ^^AAi

AAkj A A Ait A^ AA AL

1u me 0S 17 18 19 a' x Y Fig. 10. Karyotype ofa male Echiothrix leucura (AMNH 225682): 2N = 40, FN = 75. All autosomes are submetacentric except pairs 15 and 16, which are telocentric, and the small metacentric pairs 17- 19. See text for additional discussion. otype. First, the other species of Haeromys dicate the significance ofthe resemblance be- should be sampled to discover the range in tween Haeromys and the Australian murids chromosomal morphology and number with- which is now seen in gross chromosomal in the genus. Second, banding techniques structure. My preliminary studies, however, should be used to elucidate the amount and suggest the chromosomal resemblances be- distribution of constitutive heterochromatin tween Haeromys and the Australian genera in the karyotypes of Haeromys, and to de- to reflect shared primitive traits. termine whether or not the autosomal chro- In contrast to Haeromys, the shrew rat, mosomes are homologous with those in some Echiothrix, has not been found anyplace out- species of Chiropodomys, Margaretamys, or side mainland Sulawesi. Data derived from the Australian group. For example, after pro- museum study skins, skulls, and teeth have cessing chromosomes in a sample of C. gli- yielded different interpretations of the phy- roides from the Malay Peninsula through G- logenetic relationship between Echiothrix and and C-banding, Yong (1983) discovered that other Indo-Australian murids. Thomas (1898) one arm ofthe metacentric and the short arm regarded Rhynchomys, a large-bodied shrew of the submetacentric autosomes are heter- rat known only from the Philippine island of ochromatic, as is the entire Y-chromosome. Luzon (Musser, 1969), to be the closest rel- Finally, the morphological traits of Haero- ative of Echiothrix and even suggested that mys need to be compared with those of the the two be placed in the same subfamily. Australian cluster as well as Sulawesian Mar- Writing in 1936, Tate disagreed with Thomas garetamys to discover ifthere are shared der- and put forth the opinion that Echiothrix is ivations by which a phylogenetic relationship a specialized derivative of the Rattus xan- could be inferred indicating that the species thurus group, which is a cluster of species of Haeromys and those in Australia share a endemic to Sulawesi. Ellerman (1941), be- more recent common ancestor that was dif- lieving that Tate completely misjudged the ferent from the one shared by other Indo- distinctive and specialized features ofthe Su- chinese, Sundaic, and Sulawesian species, es- lawesian shrew rat, considered Echiothrix to pecially those in Chiropodomys and be morphologically far removed from any Margaretamys. Such an inquiry would in- species of Rattus. 16 AMERICAN MUSEUM NOVITATES NO. 2989

A study of molar occlusal patterns under- mal configuration compared with what I in- taken by Misonne (1969) could not identify terpret to be a primitive karyotype that is the closest relative ofEchiothrix. He includ- retained by the Australian rat. No chromo- ed the genus in a division basically defined somal evidence suggests a close phylogenetic by one characteristic: occlusal surfaces ofmo- relationship between Hydromys and Echio- lars become basined-not cuspidate-after thrix. only a little wear. To Misonne, the group con- Among the other Indo-Australian mur- sists of some members of the Murinae (the ids-aside from Hydromys-whose karyo- New Guinea Macruromys and Philippine types have been described and published, the Crunomys) and all of the Hydromyinae (the only ones with chromosomal compositions Australian and New Guinea Leptomys, Para- remotely similar to that possessed by leptomys, Xeromys, Hydromys, Pseudohy- Echiothrix are three species of Maxomys. A dromys, Microhydromys, Neohydromys, diploid number of 36, fundamental number Parahydromys, Crossomys, Mayermys, and of 71 (72 in females), autosomal set of all Philippine Chrotomys, Celaenomys, and biarmed chromosomes, metacentric X chro- Rhynchomys). Misonne concluded that mosome, and telocentric Y characterize M. Echiothrix belongs in the Murinae, but is not whiteheadi from the Malay Peninsula (Yong, dentally close to any of the other genera he 1969). Maxomys inas, also from the Penin- considered members ofthat group; noted that sula, has a diploid number of 42, fundamen- its "common trend towards a basin-shaped tal number of 83 (84 in females), 20 auto- molar structure, a character not met with in somal pairs of biarmed chromosomes (10 any other Murinae, allows a classification metacentrics and 10 submetacentrics), met- closer to the Hydromyinae" (p. 159); and even acentric X chromosome, and telocentric Y suggested that the shrew rat, along with Ma- (Yong, 1969). The Sulawesian M. hellwaldii cruromys and Crunomys, could be "an ex- has a diploid number of 34, a fundamental ample ofpossible transition" between the two number of 61 (62 in females); an autosomal subfamilies "but this cannot be clearly shown cluster consisting of four pairs of metacen- now" (p. 159). Clearly, his results are equiv- trics, seven pairs ofsubmetacentrics, two pairs ocal and he was really left in a quandry about ofsubtelocentrics, and three pairs oftelocen- the place of Echiothrix in the evolution of trics; a subtelocentric X chromosome and the morphological and cladistic diversity of telocentric Y (Musser, in prep.). In all the murines endemic to the archipelagos and other species of Maxomys sampled to date continents from Indochina to Australia. (M. bartelsii, M. moi, M. rajah, and M. sur- Chromosomal data are currently unreveal- ifer), telocentric chromosomes form a signif- ing about phylogenetic relationships between icant portion of the autosomal karyotype Echiothrix and other murids. Karyotypes (Yong, 1969; Duncan and Van Peenen, 1971; have not been published for Macruromys, Van Peenen et al., 1974; Markvong et al., Crunomys, or any of the Philippine shrew 1973). rats (Chrotomys, Celaenomys, and Rhyncho- These chromosomal resemblances be- mys) so chromosomal comparisons between tween Echiothrix and Maxomys are gross; at them and Echiothrix are not yet possible. Of more refined levels the similarities vanish. the other genera that Misonne (1969) grouped No species ofMaxomys, for example, has an within the Hydromyinae, only the Australian autosomal composition of mostly submeta- Hydromys chrysogaster is represented by centric chromosomes. Homologies among karyotypes (Baverstock et al., 1977). It has a chromosomes of the two groups are just a diploid number of 48, 22 pairs of telocentric guess without banding patterns, and it is un- autosomes, and two small pairs of metacen- known if the biarmed elements result from tric autosomes, a set ofchromosomes that is fusion of telocentric chromosomes, the ad- strikingly unlike Echiothrix with its diploid dition of heterochromatin, or a combination number of40, 14 pairs ofsubmetacentric au- of both processes. Without further informa- tosomes, two pairs of small telocentrics, and tion, the karyotypic resemblances between three pairs of small metacentrics. The Sula- Echiothrix and Maxomys are interesting but wesi animal possesses a derived chromoso- essentially uninformative; the chromosomal 1990 MUSSER: SULAWESI RODENTS 17 composition of E. leucura remains distinc- Boonsong and J. A. McNeely (eds.), tive, possibly highly derived, and enigmatic Mammals of Thailand, pp. 396-487. in the context of its significance for phylo- Bangkok, Thailand:Assoc. Conserv. of genetic inference. Wildlife. Karyotypes ofthe two endemic Sulawesian Miller, G. S., Jr., and N. Hollister murids here a 1921. Descriptions of sixteen new murine ro- reported provide window dents from Celebes. Proc. Biol. Soc. through which we can view possible phylo- Washington 34: 67-76. genetic relationships between them and other Misonne, X. species, those indigenous to Sulawesi and 1969. African and Indo-Australian Muridae. those occurring elsewhere. But the window is Evolutionary trends. Mus. R. l'Afrique small and the view cramped. In its low dip- Cent., Tervuren, Zool. 172: 219 pp. loid number (40), high fundamental number Musser, G. G. (75), and autosomal component of mostly 1969. Results of the Archbold Expeditions. biarmed chromosomes, the karyotype of No.91. A new genus and species ofmur- Echiothrix leucura is strikingly different from id rodent from Celebes, with a discus- that ofHaeromys minahassae, which is char- sion ofits relationships. Am. Mus. Nov- acterized by a higher diploid number (48), itates 2384: 41 pp. lower fundamental number (54), and pre- 1979. Results of the Archbold Expeditions. uniarmed The No. 102. The species of Chiropodomys, dominantly autosomes. two arboreal mice ofIndochina and the Ma- karyotypes share only the metaphorical sim- lay Archipelago. Bull. Am. Mus. Nat. ilarity ofbeing mirrors, one reflecting the spe- Hist. 162: 377-445. cialized essence of the shrew rat, the other 1981. Results of the Archbold Expeditions. the primitive nature of the tree mouse. Im- No. 105. Notes on systematics of Indo- ages of their closest relatives and their evo- Malayan murid rodents and descrip- lutionary histories remain unformed and tions of new genera and species from cloudy. Ceylon, Sulawesi, and the Philippines. Ibid., 168: 225-334. Musser, G. G., and M. Dagosto REFERENCES 1987. The identity of Tarsius pumilus, a pyg- my species endemic to the montane Baverstock, P. R., C. H. S. Watts, and J. T. Ho- mossy forests of central Sulawesi. Am. garth Mus. Novitates 2867: 53 pp. 1977. Chromosome evolution in Australian Musser, G. G., and C. Newcomb rodents. I. The Pseudomyinae, the Hy- dromyinae and the Uromys/Melomys 1983. Malaysian murids and the giant rat of Group. Chromosoma 61: 95-125. Sumatra. Bull. Am. Mus. Nat. Hist. 174: Duncan, J. F., and P. F. D. Van Peenen 327-598. 1971. Karyotypes of ten rats (Rodentia: Mur- Patton, J. L. idae) from Southeast Asia. Caryologia 1967. Chromosome studies of certain pocket 24: 331-346. mice, genus Perognathus (Rodentia: Ellerman, J. R. Heteromyidae). J. Mammal. 48: 27-37. 1941. The families and genera of living ro- Tate, G. H. H. dents, vol. 2. London: British Museum 1936. Results of the Archbold Expeditions. (Nat. Hist.), 690 pp. No. 13. Some Muridae ofthe Indo-Aus- 1949. Ibid., vol. 3, 210 pp. tralian region. Bull. Am. Mus. Nat. Hist. Laurie, E. M. O., and J. E. Hill 77: 501-728. 1954. List of land mammals of New Guinea, Thomas, 0. Celebes and adjacent islands 1758-1952. 1896. On mammals from Celebes, Borneo, and London: British Museum (Nat. Hist.), the Philippines recently received at the 175 pp. British Museum. Ann. Mag. Nat. Hist., Markvong, A., J. Marshall, and A. Gropp ser. 6, 18: 241-250. 1973. Chromosomes ofrats and mice ofThai- 1898. On the mammals obtained by Mr. John land. Nat. Hist. Bull. Siam Soc. 25: 23- Whitehead during his recent expedition 40. to the Philippines. Trans. Zool. Soc. Marshall, J. T., Jr. London 14: 377-412. 1977. Family Muridae: rats and mice. In L. Tsuchiya, K., T. H. Yosida, K. Moriwaki, S. Ohta- 18 AMERICAN MUSEUM NOVITATES NO. 2989

ni, S. Kulta-Uthai, and P. Sudto 1973. Chromosomes of the pencil-tailed tree- 1979. Karyotypes of twelve species of small mouse, Chiropodomys gliroides (Ro- mammals from Thailand. Rep. Hok- dentia, Muridae). Malayan Nat. J. 26: kaido Inst. Publ. Health 29: 26-29. 159-162. Van Peenen, P. F. D., R. H. Light, F. J. Duncan, 1983. Heterochromatin blocks in the karyo- R. See, J. Sulianti Saroso, Boeadi, and W. P. type of the pencil-tailed tree-mouse, Carney Chiropodomys gliroides (Rodentia, 1974. Observations on Rattus bartelsii (Ro- Muridae). Experientia 39: 1039-1040. dentia: Muridae). Treubia 28: 83-117. Yong, H. S., S. S. Dhaliwal, and B. L. Lim Viegas-Pequignot, E., B. Dutrillaux, M. 1982. Karyotypes of Hapalomys and Pithe- Prod'Homme, and F. Petter cheir (Rodentia: Muridae) from Pen- 1983. Chromosomal phylogeny of Muridae: a insular Malaysia. Cytologia 47: 535-538. study of 10 genera. Cytogenet. Cell Gen- Yosida, T. H. et. 35: 269-278. 1973. Evolution of karyotypes and differen- Yong, H. S. tiation in 13 Rattus species. Chromo- 1969. Karyotypes of malayan rats (Rodentia- soma 40: 285-297. Muridae, genus Rattus Fischer). Chro- mosoma 27: 245-267.

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates, Bulletin, and Anthropological Papers published during the last five years are available free of charge. Address orders to: American Museum of Natural History Library, Department D, Central Park West at 79th St., New York, New York 10024.

THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER.