Bunomys karokophilus, Musser, 2014

Musser, Guy G., 2014, A Systematic Review Of Sulawesi Bunomys (Muridae, Murinae) With The Description Of Two New Species, Bulletin of the American Museum of Natural History 2014 (392), pp. 1-313 : 237-272

Bunomys karokophilus , new species

HOLOTYPE: AMNH 225038, the skin, skull, and fluid-preserved carcass of an adult male (original number ASE 2322) collected November 6, 1974, by G.G. Musser. Standard external measurements, weight and other data, and measurements of the skull and dentition are listed in table 40. The stuffed skin is complete, the cranium and mandible are intact (figs. 84–86), and the carcass was originally preserved in formalin but is now stored in a mixture of water and ethanol.

TYPE LOCALITY: Tropical lowland evergreen rain forest along the Sungai Sadaunta (01 ° 23 9 S, 119 ° 58 9 E), a tributary on right side of the Sungai Miu , at 3300 ft (1007 m; locality 8 in gazetteer and on the map in fig. 50), in the northern portion of the westcentral mountain block of the island’s core, Propinsi Sulawesi Tengah, Indonesia .

DIAGNOSIS: A member of the B. fratrorum group that is moderately large in physical size (LHB 5 150–190 mm, W 5 95–175 g, ONL 5 39.0– 42.4 mm) and further characterized by the following combination of traits: (1) a short and wide muzzle and small external pinnae relative to body size; (2) dorsal fur dark gray or steel-blue speckled with pale buff, ventral fur varies from dark grayish white to dark gray or blue-gray speckled with white, dorsal carpal and metacarpal surfaces dark gray to brownish gray, digits and claws unpigmented; (3) tail typically shorter than length of head and body (LT/LHB 5 94 %), grayish brown to brownish gray on dorsal surface, white through mottled to brown over ventral surface (4) white tail tip characterizing all specimens in sample, short to moderately long relative to length of tail (mean 5 18.2 %, range 5 2 % –31 %); (5) testes small relative to body size (10 %); (6) spermatozoa head and tail length similar in shape and dimensions to spermatozoa of B. penitus , but the apical hook on the head is shorter and the tail is connected to middle of the concave surface; (7) chunky skull with a wide and moderately short rostrum, relatively narrow interorbit, narrow and upright zygomatic plate, short and broad incisive foramina, long bony palate (related to the short incisive foramina), and small ectotympanic bulla relative to skull size; (8) molars wide relative to lengths of maxillary and mandibular molar rows; (9) cusp t3 occurs infrequently on second upper molar (33 %) and third upper molar (22 %); (10) occlusal cusp pattern of third upper molar simple, cusp t1 typically not present or reduced to a low cingular ridge or pimplelike cusp on the cingular ridge in most specimens; (11) anterior labial cusplets absent from first lower molar, posterior labial cusplets present on first lower molars in half of sample and on second lower molars in three-fourths of sample; (12) anterolabial cusp present on second and third lower molars in about one-third of sample; and (13) karyotype, 2N 5 42, FNa 5 56, FNt 5 60.

REFERRED SPECIMENS AND DATES OF COLLECTION: Total 28, including the holotype (coordinates for collection localities are provided in gazetteer; also see distribution map in fig. 50). Sungai Sadaunta: 2700 ft (823 m), AMNH 225027–33 (September 25– December 3, 1974) ; 2850 ft (869 m), AMNH 225034 , 225039 (October 1, 7, 1974) ; 2900 ft (884 m), AMNH 225035 (October 2, 1974) ; 3000 ft (915 m), AMNH 225040 (October 30, 1974) ; 3200 ft (976 m), AMNH 224772 , 226833 (November 11, 1974, March 19, 1976) ; 3250 ft (991 m), AMNH 225036 , 225037 (November 3, 4, 1974) ; 3275 ft (999 m), AMNH 225041 (November 19, 1974) ; 3300 ft (1006 m), AMNH 225038 (holotype, November 6, 1974). Tomado: 1000 m, AMNH 223045 , 223046 , 223056 , 223058 , 223060 , 223072 , 224153 , 226931 , 257190 (January 25, 1972; July 26–27, 1973; August 2–14, 1973; May 14, 1974). Sungai Tokararu: 1150 m, AMNH223305 , 223316 (September 22, October 8, 1973) .

GEOGRAPHIC AND ELEVATIONAL DISTRI- BUTIONS: The only known sample of Bunomys karokophilus consists of the specimens collected along the Sungai Sadaunta (823– 1006 m) in the drainage of the larger Sungai Miu, and in the watershed of Danau Lindu in the vicinity of Tomado (1000 m) and Sungai Tokararu (1150 m); see gazetteer and the map in figure 50. No specimens exist in older collections stored in the world’s museums I have visited.

All collection sites are in the west-central mountain block of Sulawesi’s core (see the map in fig. 50), and are characterized by wet and cool habitats along shaded streams and wet ravines in tropical lowland evergreen rain forest. At least three species of shrews ( Crocidura ), two or more tarsiers ( Tarsius ), a tree squirrel ( Prosciurillus topapuensis ) and ground squirrel ( Hyosciurus heinrichi ), and 20 species of rats and mice in Crunomys , Maxomys , Bunomys , Sommeromys , Melasmothrix , Tateomys , Margaretamys , Taeromys , Paucidentomys , Haeromys , Eropeplus , an undescribed genus related to Eropeplus and Paruromys , an undescribed shrew rat, and the amphibious rat Waiomys are also endemic to the west-central mountain block (table 65).

No present evidence indicates the range of B. karokophilus to extend beyond the western mountainous region of central Sulawesi. However, its distributional limits in that vast highland block have yet to be determined. If the collection sites along my transect reliably reflect its altitudinal range, B. karokophilus should be sought in tropical lowland evergreen rainforest habitats elsewhere in the west-central region and not in montane forest formations.

SYMPATRY WITH OTHER BUNOMYS: In addition to B. karokophilus , two other members of the B. fratrorum group are found along my transect, but neither occurs syntopically with the gray rat (fig. 103; table 6). We trapped B. andrewsi along the Sungai Sadaunta but only as high as 675 m; 823 m is the lowest collection site of B. karokophilus on the Sungai Sadaunta. The montane B. penitus drops to 1285 m on Gunung Kanino, which is the lower boundary of lower montane forest along the transect. The highest collection site for B. karokophilus is in tropical lowland evergreen rain forest at 1150 m (Sungai Tokararu) at the base of the high ridge forming Gunung Kanino (see the map in fig. 50).

Bunomys karokophilus lives in the same habitat as B. chrysocomus , the only representative of the B. chrysocomus group we trapped on the transect line between 823 m and 1150 m (fig. 103; tables 6, 20). We caught examples of each species between 823 m and 1006 m in streamside forest along the Sungai Sadaunta (both were taken at 823 m in the same trapline on September 25 and 28, 1974; and at 854 m on October 1, 1974); at Tomado, 1000 m (taken in same trapline on July 26 and 27, and August 2–4, 10, and 14, 1973); and at 1150 m near the Sungai Tokararu (both were caught in the same trapline on October 8, 1973).

ETYMOLOGY: ‘‘ Karoko,’’ in the local language (bahassa Kulawi) in the area of central Sulawesi where I worked, refers collectively to ‘‘ear and ear lobes’’ and is also used for the gelatinous brownish-purple earlike lobes forming the fruiting body of the ear fungus Auricularia delicata (in the past also called a ‘‘jelly’’ fungus). This information was imparted to me by Panca, the father of Aminudi, one of my helpers, who also told me the fungus grew only on wet and decomposing wood and was sometimes eaten by the villagers. The fungus is also the primary ingredient in the diet of the gray rat, and I combine karoko with the Greek philia to indicate this gastronomic fondness.

DESCRIPTION: Bunomys karokophilus is of medium body size (LHB 5 150–190 mm, LT 5 135–205 mm, LHF 5 36–44 mm, LE 5 22–25 mm, W 5 95–175 g, ONL 5 39.0– 42.4 mm) with dark gray or bluish-gray fur and a moderately long tail with a white tip (see the rendition based on a live animal in fig. 6)—a dark gray animal with a whitetipped tail. The long (15–20 mm) dorsal coat is smooth and soft to the touch. The dorsal coat of a few individuals is very dark gray speckled with pale to very pale buff (produced by the combination of dark gray underfur, overhairs that are dark gray for most of their lengths and tipped with short and pale buffy bands, all intermixed with black guard hairs); sides of the body are paler (because of the paler gray and buffy bands). Because the guard hairs are only slightly longer than the overhairs, the surface of the coat is smooth, and the glistening guard hairs and dark gray of the underfur and overhairs produce a subdued sheen to the coat. Most specimens exhibit a darker steel-blue dorsal pelage; the buffy bands are narrower, and the combination of the long dark gray overhairs mingled with the glistening black guard hairs imparts a deep, bluish-gray cast to the fur, a tone we came to associate with ‘‘ tikus abu-abuan. ’’ An old female has unpigmented (silver) guard hairs scattered through the fur. The nose, lips, and lower half of each side of muzzle are white in some individuals, but sides of the muzzle are gray in others. Rest of the muzzle, upper cheeks, and around the eyes are darker, approaching grayish black. Forearms are dark gray, contrasting with the dorsal surfaces of the feet and rest oftops of feet and rest of the arms.

Fur covering the underparts of the head and body is also soft and dense, but shorter (8–10 mm long) than the dorsal fur, the usual pattern in murids. Four specimens out of the series have a grayish-white ventral coat with whiter areas on the throat, chest, and inguinal region (the long tips of the gray hairs are unpigmented). The rest of the specimens show very dark gray or bluish gray lightly speckled with white (the dark gray hairs have short unpigmented tips). The contrast between dorsal and ventral coats is barely evident.

In life, the ears (pinnae) feel and appear rubbery, and the color ranges from shiny dark gray through bluish gray to grayish black. They seem naked, but are scantily covered by short hairs. Ears of the stuffed voucher skins lack the rubbery texture of the live animal and have dried to brownish black.

Typically the tail is nearly equal to or shorter than the combined length of head and body (LT/LHB 5 95 %). Tail patterning is individually variable. All specimens show a distal white tail segment (unpigmented on all surfaces) ranging in the sample from 2 % to 31 % of the tail length (mean 5 18.2 %; table 8). Behind the white tip the top and sides of the tail range from glossy dark

TABLE 65

Nonvolant Mammal Species Currently Recorded Only from the West-Central Region in Sulawesi’s Core a

TABLE 65

(Continued)

d Documented in Musser et al. (2010).

e Nearly all records for these species come from montane forest habitats, but a few specimens of each have also been collected recently at 1600 m in the transition between tropical lowland evergreen and lower montane rainforest formations on Gunung Gandangdewata in the Quarles Range (K.C. Rowe, in litt., 2012).

f In lowland forest on Gunung Gandangdewata, an undescribed shrew rat (a new genus and species) was collected by Kevin Rowe and colleagues (K.C. Rowe, in litt., 2012; Esselstyn et al., 2012).

g The holotype of dollmani is from Rantekaroa (Bulu Karua), Pegunungan Quarles, in the west-central mountain

block, and until the 1970s was the only specimen reported from that mountainous region. I had identified a sample of Maxomys collected by Gerd Heinrich from Pegunungan Mekongga on the southeastern peninsula as the same species ( Musser, 1969), but that assessment is incorrect. Along my transect I collected a small series of a Maxomys exhibiting traits similar to the holotype of dollmani . Three specimens collected in 2010 by J.L. Patton and his colleagues from Gunung Balease (MVZ 225723, 225724, and 225821 also closely resemble the holotype, as does one specimen (KCR 1494) collected by K.C. Rowe and Anang Setiawan Achmadi from Gunung Gandangdewata in Pegunungan Quarles; these two samples are from the west-central mountain block. Heinrich’s collection from Pegunungan Mekongga is a distinct species endemic to that range but still more closely related to M. dollmani than to any other Sulawesian species of Maxomys (Musser, MS.) .

h This undescribed species of Maxomys is based on a collection I made in the forested basin of Danau Lindu, 955– 1000 m. A sample from Gunung Rorekatimbo in the northern part of the west-central mountain block has been identified as an undescribed species by analysis of mtDNA ( Achmadi et al., 2013); it is not included in the table because I don’t know what forest formation the specimens came from.

grayish brown to brownish gray. The ventral surface varies considerably as follows: (1) same color as or only slightly paler than top and sides (monocolored); (2) basal half is the same color as the top and sides, but the distal half behind the white tip is paler or the entire ventral surface is pale brownish gray, so the tail is conspicuously bicolored; (3) white that is densely speckled with tan; (4) apparently white but is lightly speckled, usually only over basal half of tail; (5) glossy white from base to tip (so tail is bicolor except for distal white segment; this last pattern is also typical of B. penitus ).

Metacarpal and metatarsal surfaces range from dark gray to brownish gray (the integument is white but densely covered with dark gray or brownish hairs), with silvery highlights in some specimens. Palmar pads are gray, the rest of the palmar surface ranges from pale gray to unpigmented; all the plantar surface ranges from gray to dark gray. Digits of front and hind feet are white on some specimens, on others the distal half of digits are white, the basal half speckled gray. All claws are unpigmented, those on the front digits are not concealed by ungual tufts, but comparable short tufts of silvery hairs sparsely cover the hind claws.

Females exhibit the number of teats usual for all species of Bunomys : four, arranged in two inguinal pairs. The scrotal sac of males is gray and sparsely haired (appears naked), and the testes are small relative to body size (10 %; table 9). Spermatozoan morphology is described by Breed and Musser (1991; under Bunomys sp. ).

Dorsal pelage of the three juveniles in the sample is denser to the touch than that of adults, appears woolly, and is dark gray with a flat tone (lacking the glossy sheen of the adult coat). Underparts are dark grayish white. The range in color of the ears and feet is similar to that of adults. All three juveniles have a white-tipped tail that is dark brown on all surfaces behind the white tip.

The stocky skull is smaller than that of most examples of other species in the Bunomys fratrorum group (figs. 84–86). The short and wide rostrum of B. karokophilus is distinctive, as is its short incisive foramina ending well anterior to front surfaces of the first molars (with a correspondingly long, bony palate). Relative to overall size of the skull, the interorbital region and zygomatic plate are narrow and the ectotympanic bullae small. Configuration of each dentary resembles the shape of this element in other members of the B. fratrorum group.

Compared with other members in the B. fratrorum group, B. karokophilus has a shorter molar row (indicated by the maxillary molars, the row measured), but the teeth are wider relative to length of the toothrow, a proportion imparting a chunky aspect to them.

Cusp patterns forming the occlusal surfaces are also unelaborate and after much use the cusp rows wear to irregular basins surrounded by dentine (figs. 87, 88).

Another distinctive dental feature is the occlusal configuration of the third upper molar. In all species of Bunomys except for B. karokophilus , cusp t1forms the anterolingual border of the third molar and is large relative to overall size of that tooth, nearly as relatively large as cusp t1 on the second upper molar (see the molars of B. penitus , for example, portrayed in figs. 74 and 75, or any of the images of maxillary molars from the other species of Bunomys ). Cusp t1 is either missing or variable in its expression in the sample of B. karokophilus (fig. 87), as results of the following survey of 22 specimens demonstrate: (1) the occlusal surfaces of four individuals are too worn to discern cusp patterns; (2) there is no sign of cusp t1 on the third upper molar in five specimens, the occlusal surface is composed of an anterior lamina formed by cusps 4 and 5 and a posterior cusp; (3) a low cingular ridge forms the anterolingual margin of the molar in one rat; (4) in four specimens, cusp t1 is very small and takes the shape of a deformed pimplelike cusp on the cingular margin, tiny compared with the normal-sized cusp t1 on the second molar; (5) seven specimens bear a definite cusp of medium size, but it is relatively much smaller than the large cusp t1 on the second molar and is distorted, configured among the seven rats as a ridge, a comma, or a ridge with two enamel bumps; (6) finally, only a single specimen has a large cusp t1 that relative to the other cusps forming the third molar is as large as cusp t1 on the second molar.

The absence of cusp t3 from second and third upper molars in a significant portion of the sample also signals an uncomplicated chewing surface. The cusp is present on the second molar in 33 % of the sample and occurs on the third molar in only 22 % of all specimens surveyed (table 10).

Missing or low frequency of occurrence of particular cusps and cusplets on the lower (mandibular) molars contributes to their simple occlusal topography (fig. 88, table 11). An anterolabial cusp does not occur on the second and third molars in about three-fourths of the sample. The first lower molar does not support an anterior labial cusplet (usual for most other species of Bunomys except B. chrysocomus ), but does exhibit a posterior labial cusplet in about half of the sample; this cusp occurs somewhat more frequently on the second molar.

KARYOTYPE: 2N 5 42, FNa 5 56 and FNt 5 60, comprised of six pairs of metacentric chromosomes, two pairs of subtelocentrics, and 12 pairs of acrocentrics; the sex chromosomes are submetacentrics (table 12).

COMPARISONS: Bunomys karokophilus requires comparisons with four other species of Bunomys , beginning with B. fratrorum . That species is endemic to the northern arm of Sulawesi east of the Gorontalo region; summarizing its phenetic contrasts with B. karokophilus is relevant to demonstrate that the latter is not just a population of B. fratrorum occurring in central Sulawesi. This contrast is especially relevant viewed in context of the UPGMA clustering of population samples representing all species of Bunomys that indicates B. karokophilus to be phenetically more closely related to B. fratrorum than to B. andrewsi or B. penitus , the other members of the B. fratrorum group (fig. 21).

Bunomys chrysocomus , B. andrewsi , and B. penitus are the others to be contrasted with B. karokophilus . The gray rat is syntopic with B. chrysocomus and is altitudinally parapatric with samples of B. andrewsi (at lower elevations) and B. penitus (at higher elevations) along my altitudinal transect.

Bunomys karokophilus and B. fratrorum: Both species are similar in body size, as estimated by average lengths of head and body and hind foot (because weight is unavailable for B. fratrorum , I cannot assess similarity or difference in mass between it and B. karokophilus ), but differ in tail length— B. karokophilus has a shorter tail both absolutely and relative to length of head and body (LT/LHB 5 94 % as opposed to 99 % –101 % in the three population samples of B. fratrorum ; table 41). Bunomys karokophilus also has, on average, a slightly longer white tail tip relative to tail length (mean 5 18.2 %, range 5 2 % –31 % in B. karokophilus ; mean 5 17.7 %, range 5 6 % – 75 % for B. fratrorum ; table 8), and an average absolutely longer white tip (means 5 35.6 mm for B. karokophilus and 30.2 mm for B. fratrorum ). The dorsal fur is dark grayish brown or steel-blue and the ventral coat dark grayish white in B. karokophilus , but B. fratrorum has rich, brownish gray upperparts speckled with buff and black, and grayish white underparts speckled or washed with buff.

Testes size relative to body size is similar in the two species (table 9); spermatozoan morphology is different. In B. karokophilus , the falciform sperm head is long and moderately wide with a long apical hook and moderately long tail attaching near the middle of the convex surface of the sperm head. The spermatozoal tail attaches to about the middle of the sperm head in B. fratrorum , but the head is shorter and gently curved with a shorter apical hood, and the tail is shorter ( Breed and Musser, 1991; Breed and Taylor, 2000). It may be significant that among the species of Bunomys surveyed for their spermatozoa morphology, in only B. karokophilus and B. fratrorum does the tail attach to the middle of the sperm head; in B. chrysocomus , B. andrewsi , and B. penitus , the tail attaches to the caudal portion of the sperm head. The shared spermatozoan configuration in B. karokophilus and B. fratrorum reinforces the morphometric alliance between these two species (fig. 21).

Bunomys karokophilus has a smaller skull than B. fratrorum , as attested by the smaller univariate mean values for almost all cranial and dental dimensions except for the interorbit and the bony palate, which are slightly wider and longer, respectively, in B. karokophilus (table 66); mean values for breadth of the upper molars is the same in both. Overall, B. karokophilus possesses a relatively wider interorbit and longer bony palate (reflecting shorter incisive foramina compared to those openings in B. fratrorum ) relative to skull size as compared with B. fratrorum , and wider molars relative to length of the molar row. These absolute and proportional differences can be appreciated by examining the images of skulls portrayed in figures 84–86.

Relationships among the cranial and dental variables in the context of quantitatively discriminating between the two species are summarized by results of two multivariate analytical approaches. The first comprises a scatter plot where individual specimen scores representing the sample of B. karokophilus and all population samples of B. fratrorum are projected onto first and second principal components (fig. 89, upper graph). Along the first axis, scores for B. karokophilus generally cluster to the left of the cloud of

TABLE 66 Descriptive Statistics for Cranial and Dental Measurements (mm) Contrasting Samples of Bunomys karokophilus with Samples of B. fratrorum from the Northeastern Peninsula (East of Gorontalo), and B. chrysocomus , B. andrewsi , and B. penitus from the West-Central Mountain Block Mean ±1 SD, and observed range (in parentheses) are listed.

points representing B. fratrorum , a positional response to the many greater dimensions typical of B. fratrorum as indicated by the positive moderate to high correlations (r 5 0.53–0.90): covariation in all variables except interorbital breadth, length of bony palate, and breadth of first upper molar influence the distribution and segregation of scores along the first component (table 67). The regression lines of the second principal component on the first are phenetically distinct: their Y-intercepts are just greater than conventionally considered significantly different between the two species (+0.096 versus ‾ 0.012; F 5 3.10, P 5 0.081), but their slopes are not (0.130 versus 0.215; F 5 0.19, P 5 0.667).

Individual specimen scores for each species projected onto first and second canonical variates form another visual summary of morphometric distinctions between the two species (fig. 89, lower graph). Two nonoverlapping clusters of scores identify each species, and their separation along the first axis is stongly influenced by correlations among the many variables in which B. fratrorum exceeds B. karokophilus in size (r 5 0.21–0.76; table 67), along with the wider interorbit and longer bony palate of B. karokophilus (r 5 ‾ 0.21 and ‾ 0.41, respectively).

The two species also differ in occlusal patterns formed by certain cusps and cusplets on maxillary molars (table 10). In the sample of B. karokophilus , cusp t3 occurs at a greater frequency on the second (33 % in B. karokophilus , 11 % in B. fratrorum ) and third upper molars (22 % as opposed to 9 %). Cusp t1 on

TABLE 67 Results of Principal-Components and Discriminant- Function Analyses Performed on All Population Samples of Bunomys karokophilus and B. fratrorum Correlations (loadings) of 16 cranial and two dental log-transformed variables are based on 17 B. karokophilus and 100 B. fratrorum ; see figure 89.

the third upper molar is typically not present or reduced to a low cingular ridge or pimplelike cusp on the cingular ridge in most specimens of B. karokophilus (fig. 87), but takes the form of a large prominent cusp in all examples of B. fratrorum (fig. 12).

Coronal topography of the lower molars differs slightly between the species (table 11). An anterolabial cusp is about twice as frequent on the second molar in B. karokophilus (27 %) as in B. fratrorum (12 %), and occurs at the same percentage on the third molar in B. karokophilus (fig. 88), but is absent from the third molar in all specimens of B. fratrorum surveyed (fig. 12).

The combination of external, cranial, and dental traits characterizing B. karokophilus , along with its distinctive diet, clearly separates it from B. fratrorum . Although B. karokophilus , living in the west-central moun- tain block, is not just a geographic population of B. fratrorum , which is endemic to the northern peninsula east of the Gorontalo region, the two may be more closely allied to one another than to other species of Bunomys if morphometric aspects of skulls and molars, along with gross conformation of spermatozoa, reliably mirror genetic alliance. The phenetic relationships among all population samples of Bunomys analyzed that are illustrated in the cluster diagram in figure 21, which is derived from discriminantfunction analyses of covariation in cranial and dental variables, links the sample of B. karokophilus with those representing B. fratrorum . An independent test of this phenetic kinship by analyzing DNA sequences is required.

Bunomys karokophilus and B. chrysocomus: Samples of tikus abu-abuan were collected on the Sungai Sadaunta between 823 and 1006 m, and at Tomado (1000 m) and Sungai Tokararu (1150 m) in the Danau Lindu valley. Bunomys chrysocomus lives in the same damp forest habitats in these areas and was collected along the Sungai Sadaunta through the elevational range 823–960 m, and also at Tomado (1000 m) and Sungai Tokararu (1150 m). My comparisons are based primarily on specimens of each species representing this syntopic distribution. Descriptive univariate statistics derived from external, cranial, and dental measurements from these samples are listed in table 68. An additional comparison will contrast the sample of B. karokophilus with all the examples of B. chrysocomus collected along my transect that extended from Sungai Oha Kecil to Gunung Kanino (the two populations samples, ‘‘Sungai Oha Kecil + Sungai Sadaunta’’ and ‘‘Danau Lindu Valley + Gunung Kanino’’; table 2); table 66 contain the relevant univariate statistics. A final contrast involves a ratio diagram pitting the sample of B. karokophilus against eight population samples of B. chrysocomus (identified in table 2).

I thought the first few specimens of B. karokophilus I trapped to be B. chrysocomus for at first glance both species appear externally much alike, with their dark dorsal coats of similar length and texture and grayish underparts. But more critical inspection reveals strong dissimilarities. Bunomys

TABLE 68 Descriptive Statistics for External, Cranial, and Dental Measurements (mm) derived from Samples of Bunomys karokophilus and B. chrysocomus Collected at the Same Places in Sulawesi’s West-Central Region Specimens representing each species are from Tomado (1000 m), Sungai Tokararu (1150 m) and through an overlapping elevational range along Sungai Sadaunta (823–1006 m for B. karokophilus ; 823–960 m for B. chrysocomus ). Mean ± 1 SD and observed range (in parentheses) are listed. Measurements (excluding WT and LT/LHB) from specimens in these two samples were used to derive the results from the principalcomponents analyses summarized in figure 90 and table 69.

karokophilus has darker fur, the upperparts are dark grayish brown or blue-gray, the underparts are dark grayish white or bluegray flecked with white; B. chrysocomus has rich, dark brownish-gray dorsal fur finely spotted with dark buff and underparts that are grayish white, or grayish white suffused with buff or ochraceous hues. The living B. karokophilus has a short and broad face anterior to the eyes, but that of B. chrysocomus is long and slender by comparison. Nose, lips, and lower half of each side of the muzzle are white in some individuals of B. karokophilus , the muzzle is gray in others; nose and muzzle are brownish gray in chrysocomus , and only the lips and very tip of the nose are unpigmented. In the living animals, ears (pinnae) of B. karokophilus are dark gray to blue-gray, those of B. chrysocomus are dark brownish gray (dried ears of specimens in museums are dark brown to blackish in both species). Dorsal surfaces of the carpal and metacarpal regions are gray and the digits unpigmented in B. karokophilus , while the entire dorsal surfaces of the feet are lightly to intensely speckled dark gray to brown in B. chrysocomus . Bunomys karokophilus has short and delicate front claws compared to the noticeably longer and more robust claws of B. chrysocomus .

In both species, the tail is typically shorter in relation to length of head and body, but B. karokophilus has, on average, a relatively longer tail (LT/LHB 5 95 %) than does B. chrysocomus (mean of nine samples 5 90 %). Length of the white tail tip relative to length of tail is greater in B. karokophilus (mean 5 18.2 %, range 5 2 % –31 % for B. karokophilus ; mean 5 5.5 %, range 5 1 % –25 % for B. chrysocomus ; see table 8), and frequency of a white tip differs. All specimens of B. karokophilus (N 5 24 with intact tails) exhibit a white distal segment while only 20 % of the sample of B. chrysocomus (N 5 396) show a white tip.

Bunomys karokophilus is physically larger than B. chrysocomus (comparing adults), averaging greater in body mass and in all external dimensions except length of ear, which is about the same in absolute values, making it smaller relative to body size (LE/LHB 5 13 % for B. karokophilus , 14 % – 16 % for four samples of B. chrysocomus ). This disparity in body size is quantified in two tables where the sample of B. karokophilus (table 41) is contrasted with all population samples of B. chrysocomus (table 19), and in table 68 where B. karokophilus is compared with the sample of B. chrysocomus from the transect where the two species overlap in range (specimens of each species are from Sungai Sadaunta, Tomado, and Sungai Tokararu). Differences in external variables for the syntopic samples are summarized by individual specimen scores projected onto first and second principal components (fig. 90, upper graph). High and positive loadings for lengths of head and body, tail, and hind foot (r 5 0.71–0.97) power the scores along the first axis and segregate those for B. chrysocomus with the shorter body and appendages on the left from scores representing B. karokophilus with longer body and appendages on the right; length of ear is negligible in influencing the spread of points, which reflects the relatively smaller ears of B. karokophilus (table 69).

Relative size of testes clearly distinguishes adult males of the two species (table 9). The testes are small in relation to body size in B. karokophilus (length of testes/length of head and body 5 10 %), but conspicuously much larger in B. chrysocomus (22 %).

External morphology of spermatozoa is also different. Breed and Musser (1991: 10) noted that compared with the spermatozoa of B. chrysocomus , those of B. karokophilus (identified as ‘‘ Bunomys sp. A’’ in their report) have ‘‘a wider sperm head, shorter apical hook, and shorter sperm tail.’’ Also, the tail attaches to about the middle of the head in B. karokophilus but at the ventrocaudal portion of the head in B. chrysocomus .

Both species have a 2N of 42, but FNa 5 56 and FNt 5 60 for B. karokophilus and FNa 5 56 and FNt 5 58 for B. chrysocomus (table 12).

The two species do not share a similar diet. The karoko fungus, Auricularia delicata , comprises the bulk of B. karokophilus ’s diet; B. chrysocomus does not consume fungi (see Natural History section in the accounts of each species). This dietary contrast proved useful in separating two juveniles I caught in live traps along the Sungai Sadaunta. They were similar in body size, color and texture of their juvenile fur, and color of ears and feet; tails of both were dark gray along their dorsal surfaces and white on the ventral surface, but one had a white tip and the other did not. They could be young of either B. karokophilus or B. chrysocomus , or the two together a sample of one of those species. I offered insects, snails, earthworms, and karoko to each rat. The juvenile with the white tail tip accepted and ate only karoko; the other juvenile rejected the karoko but consumed the invertebrates. Subsequent study showed the former to be B. karokophilus , the latter B. chrysocomus .

The large, robust skull of B. karokophilus with its stocky rostrum, narrow zygomatic plate, spacious incisive foramina, and large, chunky molars contrasts with the smaller skull of B. chrysocomus that is more elongate in overall shape, appearing gracile in comparison, and with weak molars (figs. 99–101). Univariate means for breadths of interorbit, zygomatic plate, and bony palate, along with length of bullar capsule are less or the same in B. karokophilus compared with the various samples of B. chrysocomus , but the means are greater for all other cranial and dental dimensions in the sample of B. karokophilus (tables 66, 68).

How the cranial and dental variables interact in distinguishing the two species is summarized below in results of two analyses. First, individual specimen scores from the syntopic samples (Sungai Sadaunta, To-

TABLE 69 Results of Principal Components Analysis Comparing Samples of Bunomys karokophilus and B. chrysocomus The samples of B. karokophilus from Sungai Sadaunta (823–1006 m), Tomado (1000 m), and Sungai Tokararu (1150 m) are compared with samples of B. chrysocomus from the Sungai Sadaunta (823–960 m), Tomado (1000 m), and Sungai Tokararu (1150 m). Correlations (loadings) of log-transformed values for lengths of head and body, tail, hind foot, and ear (upper graph); and for 16 cranial and two dental variables (lower graph) are based on 99 specimens; see figure 90.

mado, and Sungai Tokararu) are projected onto first and second principal components where they form two oblique elliptical spreads (fig. 90, lower graph). The major axes are phenetically distinct: their Y-intercepts are significantly different between the two species (‾ 2.254 versus +0.500; F 5 23.98, P 5 0.000), but not their slopes (0.867 versus 0.873; F 5 0.00, P 5 0.988). Loadings for nearly all variables influence the distribution of scores along the first axis, reflecting the many greater internal dimensions and overall larger skull and molars of B. karokophilus compared with the smaller skull and molars of B. chrysocomus ; loadings (correlations) for overall skull size (occipitonasal length and zygomatic breadth), breadths of rostrum, mesopterygoid fossa, incisive foramina, and first upper molar, along with lengths of bony palate, postpalatal region, and maxillary molar row, are especially high (r 5 0.57–0.87; table 69). There exists but slight overlap between the two clouds of points along the first component. A principal-components analysis employing all 10 population samples of B. chrysocomus and the sample of B. karokophilus resulted in a closely similar pattern of scores (not illustrated here).

Second, relevant proportional distinctions are recorded in a ratio diagram where the sample of B. karokophilus is compared with all 10 population samples combined of B. chrysocomus (fig. 91). The interorbit of B. karokophilus is significantly narrower relative to overall skull size (reflected by occipitonasal length and zygomatic breadth); the rostrum is shorter relative to skull length and appreciably wider relative not only to its length, but to size of skull; the zygomatic plate is narrower relative to skull size and most other cranial and dental variables; the bony palate is longer but much narrower, and the mesopterygoid fossa is wider not only relative to skull size but to bony palate; the incisive foramina is markedly wider than long; the bulla, about the same size in univariate means, is smaller relative to size of skull; and the molar row is longer relative to many cranial variables, and the teeth are wider relative to length of molar row.

In addition to the larger molars of B. karokophilus , there are contrasts between it and B. chrysocomus in the frequencies of occurrence of several cusps and cusplets. Cusp t3 occurs at a greater frequency on the second (33 % in B. karokophilus , 17 % in B. chrysocomus ) and third (22 % as opposed to 5 %) upper molars (table 10). Cusp t1 on the third upper molar is typically absent or consists of a distorted cusp, cingular ridge, or pimple in B. karokophilus (fig. 87), but the third molar of B. chrysocomus sports a prominent cusp t1 (fig. 12), which is also usual in all the other species of Bunomys except B. karokophilus . In the lower molars, an anterior labial cusplet is absent from the first molar in specimens of B. karokophilus , but occurs in 52 % of the B. chrysocomus sample; about half the sample of B. karokophilus shows a posterior labial cusplet, but nearly all specimens of B. chrysocomus exhibit the cusplet (table 11). On the second molar, 27 % of the series of B. karokophilus have an anterolabial cusp, but most specimens (90 %) of B. chrysocomus have this cusp; 64 % of the sample of B. karokophilus show a posterior labial cusplet, but the frequency is 98 % in the sample of B. chrysocomus . An anterolabial cusp is less frequent on the third molar in the sample of B. karokophilus (27 %) than it is in B. chrysocomus (65 % of the sample).

Contrasts between the two species described in the preceeding paragraphs are drawn from my observations in the forest of freshly caught animals and my study in the museum of preserved voucher specimens. The results presented here are somewhat detailed, but two approaches, one in the field, the other in the laboratory, can be used to distinguish the two species. Should one be surveying for small mammals in tropical lowland evergreen rain forest in the western mountain block of Sulawesi’s core and trapping alive what may be examples of each but the identification is uncertain, find some karoko and offer it to the rats. In the laboratory or museum, measure the maxillary molar row and breadth of the first upper molar of the samples. Bunomys karokophilus has larger molars and there is little or no overlap in the range of variation within my different sets of samples. Ranges for CLM1–3 and BM1 are 6.8–7.3 mm and 2.3–2.6 mm, respectively for B. karokophilus . Values for B. chrysocomus in the syntopic samples are 5.7– 6.5 mm and 1.8–2.2 mm; in the two population samples, ‘‘Sungai Oha Kecil + Sungai Sadaunta’’ and ‘‘Danau Lindu Valley + Gunung Kanino,’’ the ranges are 5.7–6.8 mm and 1.8–2.2 mm; and for all 10 population samples of B. chrysocomus the respective ranges are 5.7–6.8 mm and 1.8–2.3 mm.

Bunomys karokophilus and B. andrewsi: These two species are generally physically similar in body size, although animals in certain geographic areas may average larger or smaller than B. karokophilus . For example, the samples of B. andrewsi from Kuala Navusu, Gunung Balease, and the Malili Area weigh more, on average, than B. karokophilus (mean 5 131.5 g for B. karokophilus , range of means 5 145.0– 154.6 g for B. andrewsi ), but body size is comparable (mean for length of head and body 5 172.9 mm for B. karokophilus , range of means 5 169.9–177.4 mm for B. andrewsi ; and mean for length of hind foot in B. karokophilus 5 39.3 mm as opposed to 40.2– 40.5 mm for B. andrewsi ); however, animals in the small sample from ‘‘Puro-Sungai Miu’’ weigh less (mean 5 113.7 g) and average slightly smaller in lengths of head and body and hind foot (means 5 163 mm and 38.1 mm, respectively).

Bunomys karokophilus has small ears (external pinnae) in relation to body size (LE/LHB 5 13 %) compared with the relatively larger ears of B. andrewsi (means 5 14 % –16 % for four samples; see table 41). This proportional contrast is also reflected by the slightly smaller (relative to skull length) bullar capsule of B. karokophilus compared with the relative size of this structure in B. andrewsi , as seen on the ratio diagram in figure 93.

No matter their provenances, all adult examples of B. andrewsi can be distinguished from B. karokophilus by coat color and tail characteristics. Fur covering the upperparts is a rich dark brown speckled with buff in B. andrewsi , the ventral coat ranges from grayish white to bright buffy gray; most specimens from lowlands have a gray ventral coat washed with pale or rich buff; many have reddish stains on the chest and throat. Bunomys karokophilus is a dark gray rat. The dorsal coat is either dark gray speckled with pale buff or a very dark blue-gray; underparts are dark grayish white in all examples at hand.

Tails are relatively longer and white-tipped in B. karokophilus , shorter and mostly without white tips in B. andrewsi . In both species, the tail is typically shorter than length of the head and body, but it is relatively longer in B. karokophilus (mean 5 95 %) than in the samples of B. andrewsi (range of means 5 75 % –92 %). All examples of B. karokophilus with intact tails (N 5 24) have a white tail tip (mean length 5 28.7 mm, range 5 4–50 mm) that comprises 2 % –31 % of the tail length (mean 5 18.2 %); of the 133 specimens of B. andrewsi , tails of only 20 % exhibit a white tip (mean length 5 8.6 mm, range 5 1–20 mm), and when present it is typically relatively shorter in relation to overall length of tail (mean 5 7.7 %, range 5 1 % –13 %); see table 8.

Diet differs between the two species. Ear fungi, primarily Auricularia delicata , comprise the bulk of B. karokophilus ’s diet; B. andrewsi consumes fruit, arthropods, earthworms, and small vertebrates but not fungi (table 13).

There is an average contrast in relative testes size (table 9). The testes of B. karoko- philus are smaller relative to combined length of head and body (10 %) compared with samples of B. andrewsi (8 % –15 %).

External morphology of spermatozoa is slightly different. Breed and Musser (1991: 10) noted that spermatozoa of B. karokophilus (identified as Bunomys sp. A ) have ‘‘a wider sperm head, shorter apical hook, and shorter sperm tail’’ compared with those of B. andrewsi ; also, site of attachment of tail to the sperm head differs.

Both species have a 2N of 42, but an FNa 5 56 and FNt 5 60 for B. karokophilus and FNa 5 56 and FNt 5 58 for B. andrewsi (table 12).

Comparisons in cranial and dental morphometrics are subjects of the following paragraphs. With the exception of a ratio diagram, the contrasts are between B. karokophilus and the samples of B. andrewsi from the west-central region where the two species are regionally sympatric but not syntopic in distribution (see gazetteer and the map in fig. 50). Skulls of the two species are similar in overall size (estimated by occipitonasal length) and many internal cranial and dental dimensions are equivalent (table 66). The few differences, however, are noticeable, as seen in tables of measurements and side-by-side comparisons of skulls: B. karokophilus typically has a narrower zygomatic breadth and interorbital region, shorter rostrum, narrower zygomatic plate, shorter incisive foramina, and longer bony palate (a reflection of the shorter incisive foramina).

The general similarity in cranial dimensions between most individuals of B. karokophilus compared with the specimens representing B. andrewsi (from Puro-Sungai Miu, Tamalanti, Tuare, Mamasa Area, and Gunung Balaese) is summarized by individual specimen scores projected onto first and second principal components where they form parallel oblique elliptical clusters (fig. 92). The major axis of each ellipsoid is phenetically discrete: their Y-intercepts are significantly different between the two species (1.493 versus ‾ 0.580; F 5 12.48, P 5 0.001), but not their slopes (0.573 versus 0.388; F 5 0.46, P 5 0.499). The high positive correlations for most cranial and dental variables (r 5 0.52–0.92; table 70) spread the scores along the first axis where there is broad overlap between the groups of scores for each species, a reflection of their similarity in magnitude of most cranial and dental variables (table 65).

Segregation of the scores into two clusters along the second component highlights shape differences (table 70). Compared with B. andrewsi , the gray rat has a relatively narrow- er interorbit (r 5 ‾ 0.41); wider rostrum, narrower zygomatic plate, longer diastema and postpalatal region, longer and wider bony palate, wider mesopterygoid fossa (r 5 0.39–0.75); shorter incisive foramina (r 5 ‾ 0.55), and broader molars (r 5 0.47).

Relevant proportional distinctions are recorded in a different form in a ratio diagram (fig. 93). Here, I combined all seven population samples from throughout the range of B. andrewsi (see table 2). Compared with this larger sample of B. andrewsi , the interorbit of B. karokophilus is significantly narrower relative to skull length (indexed by occipitonasal length); the rostrum is wider relative to its length and to length of skull; the zygomatic plate is narrower relative to skull size and most other cranial and dental variables; the bony palate is longer, and the mesopterygoid fossa is wider relative to skull length; the incisive foramina are markedly shorter than wide; and the molars are wider relative to both length of molar row and length of skull.

There are differences in occlusal cusp patterns of upper (maxillary) and lower (mandibular) molars (tables 10, 11). Compared with the sample of B. andrewsi from the central core of Sulawesi, cusp t3 occurs at a greater frequency on the second (33 % in B. karokophilus , 17 % in B. andrewsi ) and third upper molars (22 % as opposed to 4 %). Cusp t1 on the third upper molar is typically absent

TABLE 70 Results of Principal-Components Analyses Comparing the Sample of Bunomys karokophilus with Those of B. andrewsi from the West-Central Mountain Block Correlations (loadings) of log-transformed values for 16 cranial and two dental variables are based on 17 B. karokophilus and 40 B. andrewsi ; see figure 92.

or consists of a cingular ridge or distorted pimples in B. karokophilus , but the third molar of B. andrewsi exhibits a prominent cusp t1, which is usual in all the other species of Bunomys except B. karokophilus . Only about half of the sample of B. karokophilus (55 %) has a posterior labial cusplet on the first lower molar, but nearly all specimens of B. andrewsi from the core of the island bear this cusplet; 64 % of the sample of B. karokophilus has a posterolabial cusplet on the second lower, but all specimens of B. andrewsi show this cusplet. An anterolabial cusp occurs at a frequency of 27 % on both second and third lower molars in B. karokophilus , the frequency for the second molar is 94 % for B. andrewsi and 18 % for the third molar.

All qualitative and quantitative comparisons of phenetic traits attest to the distinctive phenetic (and presumably genetic) separation of B. karokophilus from B. andrewsi .

Bunomys karokophilus and B. penitus: The only sample of B. karokophilus comes from collection localities along my transect line in tropical lowland evergreen rainforest habitats between 823 and 1150 m. The site at 1150 m (locality 10 in the gazetteer and on the map in fig. 50) is at the base of Gunung Kanino. Upslope on that mountainous ridge at 1285 m marks the lower boundary of lower montane forest and the lowest spot on the transect line where B. penitus was trapped (locality 1 in the gazetteer and on the map in fig. 51). At higher elevations on Gunung Kanino and all the way to the summit of adjacent Gunung Nokilalaki, B. penitus was common in lower and upper montane forest habitats (see the map in fig. 51). Most of my comparisons will be between B. karokophilus and the samples of B. penitus from Gunung Kanino and Gunung Nokilalaki; one multivariate analysis and a ratio diagram are based on all population samples of B. penitus (identified in table 2).

Bunomys karokophilus is physically smaller than B. penitus (table 41). Mean values for lengths of head and body, hind foot, and ear are less in B. karokophilus , than in the samples of B. penitus from Gunung Kanino and Gunung Nokilalaki; body mass is not so different (133.2 g for B. karokophilus , 132.3 and 133.3 g for the two samples of B. penitus ). Mean length of tail is similar in the two species; however, relative to length of head and body, B. karokophilus has a slightly longer tail (95 %) than in either sample of B. penitus (93 % and 88 %). Relative size of ears also differs with B. karokophilus having slightly smaller ears in relation to body size (LE/LHB 5 13 %) as contrasted with the larger ears of B. penitus (14 % for two samples; table 41).

Bunomys penitus has brighter pelage than B. karokophilus . Its luxuriant dorsal coat is soft, long (up to 25 mm), densely silky to the touch, and dark brownish gray with buffy highlights; the soft, dense fur covering the underparts is whitish gray ( B. karokophilus has a harsher and shorter dorsal coat—up to 15 mm long—that is dark grayish brown or dark blue gray finely speckled with buff; the ventral coat is dark grayish white or blue-gray lightly speckled with white). Ears are tan (dark gray to grayish black in B. karokophilus ). Digits are white, dorsal surfaces of the carpal and metacarpal regions are typically white, speckled with gray in some specimens, front claws are short and appear delicate (grayish

TABLE 71 Results of Principal-Components Analyses Comparing the Sample of Bunomys karokophilus with Those of B. penitus from the West-Central Mountain Block Correlations (loadings) of log-transformed values for 16 cranial and two dental variables are based on 17 B. karokophilus and 174 B. penitus ; see figure 94.

white to grayish brown in B. karokophilus , claws longer and more robust relative to body size). Glistening white below and around the tip, then tan on the proximal dorsal strip is the typical tail color pattern (tail white tipped in B. karokophilus , but the ventral surface ranges from glistening white through varying intensities of brown speckling and mottling to solid brown). Length of the white tip relative to length of tail averages shorter in B. karokophilus (mean 5 18.2 %, range 5 2 % –31 %) than in B. penitus (mean 5 21.0 %, range 5 3 % –68 %) as does actual length of the white tip (means 5 28.7 mm for B. karokophilus and 35.6 mm for B. penitus ).

External morphology of spermatozoa is similar in B. karokophilus and B. penitus . Each has a wide sperm head, short apical hook, and short sperm tail, but the apical hook of B. karokophilus is shorter than that of B. penitus , and the tail connects to near the middle of the head rather than at its base ( Breed and Musser, 1991; B. karokophilus was reported as Bunomys sp. A ).

Both species have a 2N of 42, but an FNa 5 56 and FNt 5 60 for B. karokophilus and FNa 5 58 and FNt 5 60–61 for B. penitus (table 12).

The smaller physical size of B. karokophilus compared to B. penitus is mirrored by the cranial and dimensions measured: except for breadths of zygomatic plate and bony palate, which are statistically the same in both species, univariate means for all the other variables average less in the sample of B. karokophilus , whether compared with all population samples of B. penitus (table 42) or only those from Gunung Kanino and Gunung Nokilalaki (table 73). Dimensional contrasts that are especially noteable are the markedly shorter rostrum, incisive foramina, and molar row of B. karokophilus as compared with B. penitus .

The discriminating qualitative associations among the cranial and dental variables are summarized by the pattern of individual specimen scores projected onto first and second principal components (fig. 94). In this exercise, I expanded the sample of B. penitus to include all population samples from the west-central mountain block (Gunung Kanino, Gunung Nokilalaki, Rano Rano, Gunung Lehio, Mamasa Area, and Pegunungan Latimojong; see the univariate statistics in table 73). Bunomys karokophilus is represent- ed by an elliptical group of scores in the lower-left quadrant of the ordination, B. penitus by a parallel elliptical cloud in the upper right. The regression lines of the second principal component on the first are phenetically distinct: their Y-intercepts are significantly different between species (‾ 3.309 versus +0.279; F 5 31.74, P 5 0.000), but their slopes are equivalent (‾ 0.701 versus ‾ 0.564; F 5 0.18, P 5 0.370). The spread and partial separation of clusters along the first component is influenced by positive and moderate to high loadings for all cranial and dental variables except breadth of first upper molar (r 5 0.31– 0.87; table 71) and reflects the generally greater cranial and dental dimensions of B. penitus , especially occipitonasal length, length and breadth of rostrum, length and breadth of incisive foramina, breadth of mesopterygoid fossa, and length of molar row.

Proportional (shape) dissimilarities are summarized in a ratio diagram (fig. 95). Bunomys karokophilus has an absolutely shorter and narrower rostrum than B. penitus , but the rostrum is significantly shorter relative to skull size (indicated by occipitonasal length and zygomatic breadth) and wider relative to rostral length, a proportion that signals the relatively chunky rostrum characteristic of B. karokophilus that is unlike the relatively long and narrow rostrum of B. penitus . Absolute average measurements are equivalent for breadth of the zygomatic plate in the two species, but the plate is much wider relative to size of skull in B. karokophilus compared with B. penitus . The bony palate of B. karokophilus is wider relative to its length, the incisive foramina are relatively shorter but wider across the openings relative to their lengths, the bulla are slightly smaller in relation to skull length, and the molars are wider relative to length of the maxillary molar row.

These size and shape distinctions between B. karokophilus and B. penitus as summarized by univariate statistics for the cranial and dental variables, graphic ordinations derived from multivariate analyses, and the ratio diagram can also be detected in the illustrations where skulls of the two species are contrasted (figs. 99–101).

In addition to the difference between the two species in size of molars, the frequency of certain cusps and cusplets is variable (tables 10, 11). Cusp t3 is found on the second upper (maxillary) molar in over half the sample of B. penitus (62 %) as opposed to a third of the sample of B. karokophilus (33 %), but occurs infrequently on the third molar in B. penitus (5 %) compared to B. karokophilus (22 %). Cusp t1 on the third upper molar is typically absent or consists of a cingular ridge or pimplelike shapes in B. karokophilus (fig. 87), but the third molar of B. penitus , like all the other species of Bunomys excepting B. karokophilus , exhibits a prominent cusp t1 (figs. 74, 75). On the lower molars, a posterior labial cusplet is present on the first molar in about half of the sample of B. karokophilus (55 %) but occurs on all specimens of B. penitus examined; this cusplet is present on the second lower molar in 64 % of the sample of B. karokophilus but is more frequent in B. penitus (97 % of the sample). About half of the sample of B. penitus has an anterolabial cusp on the second molar, but no specimen exhibits this cusp on the third molar; 27 % of the sample of B. karokophilus has the cusp on both second and third molars.

GEOGRAPHIC AND ELEVATIONAL VARIA- TION: Because all the specimens of B. karokophilus come from a few localities in the northern portion of the west-central mountain block, there is nothing serious to report concerning geographic variation in external, cranial, and dental traits. The specimens comprise a single population sample; any analysis of geographic variation will have to wait until samples from else-

TABLE 72

Summary of Microhabitats at Trapping Sites, Stomach Contents, and Other Relevant Information for

Specimens of Bunomys karokophilus Collected on My Transect in Sulawesi’s West-Central Region, 1973–1976

Descriptions of the trapping sites and contents of stomachs are summarized from my field journals (in

Mammalogy Archives at AMNH). All collection localities are in old-growth tropical lowland evergreen rain

forest (see the habitats depicted in figures 96–98). All the rats were caught during the night in Conibear traps

unless indicated otherwise. In the description of stomach contents, ‘‘ karoko ’’ is the local name for the purple ‘‘ear’’ fungus, Auricularia delicata .

Locality, AMNH

and (ASE-field) Elevation

numbers (m) Date Trap site and other information

Sungai Sadaunta

225027 (2051) 823 Sept. 1974 In a damp runway beneath crisscross of large rotting tree limbs on a narrow rocky terrace above Sungai Sadaunta. The limbs are scattered about from an old giant that lies decaying on the hillside above the stream terrace. Moss-covered rocks and limbs on the terrace are covered by leafy vines, shrubs, and ferns; the streamside environment is wet and cool. Stomach: packed with cut pieces of karoko only.

225028 (2073) 823 Sept. 1974 On wet, mossy trunk extending from one terrace of Sungai Sadaunta to the opposite terrace and resting just above water level on rocks in the stream; trunk would be inundated in high water. The tree squirrel Rubrisciurus rubriventer was taken on same spot during the day and the rat Maxomys hellwaldii on a different night. Stomach: stuffed with finely chopped karoko; in middle of the mass is a hard piece of dark brown wood (15 mm long, 10 mm wide) that probably was detached along with the fungus from the rotting trunk or limb.

225034 (2090) 854 Oct. 1974 On top of moss-covered decaying trunk lying at 30 ° angle across a small creek at the bottom of a steep, wet, and mossy ravine tributary to Sungai Sadaunta. Slopes above ravine covered with tall hillside forest. Stomach: full of karoko.

225035 (2100) 884 Oct. 1974 On wet, smooth, rotting trunk of a wanga palm ( Pigafetta filaris ) bridging the Sungai Sadaunta and resting on rocks on the stream (would be inundated after rains). The rocky, wet, and mossy stream banks are shaded by understory figs, tall shrubs, endemic banana trees ( Musa sp. ), and tall wanga palms. Stomach: full of karoko; also scanty remains of a large insect (a few abdominal sclerites and parts of a wing).

225029 (2116) 823 Oct. 1974 On damp ground beneath base of large, decaying tree lying on terrace above Sungai Sadaunta. Now rotten and covered with moss, shrubs, ferns, and vines, the trunk provides protected cover and pathways beneath the area where it splintered from the adjacent stump; stump and trunk are surrounded by dense and tall shrubs, tall gingers, and rattan. The rat Paruromys dominator was caught on same spot on a different night. Other P. dominator , Bunomys chrysocomus , Rattus facetus , and Maxomys hellwaldii were trapped on the forested hillside above the stump. Stomach: some karoko, small fig seeds and bits of rind; few chitinous segments of adult beetle abdomen and a leg.

225039 (2135) 869 Oct. 1974 On half-shredded, rotting, and slick wanga palm trunk lying across Sungai Sadaunta, extending from a low terrace (muddy, rocky, and eroded) on one side of stream across to the opposite rocky bank. Bunomys chrysocomus and the tree squirrels Prosciurillus topapuensis and P. murinus were caught in same spot during different days and nights. A few feet upstream a rat, Paruromys dominator , was trapped on a different trunk bridging the stream.

Stomach: full of chopped karoko; thin flakes of the woody holdfasts are mixed with the karoko.

TABLE 72 (Continued)

Locality, AMNH

and (ASE-field) Elevation

numbers (m) Date Trap site and other information

224933 (2250) 823 Oct. 1974 ASE 2250, a Taeromys celebensis , and Rattus hoffmanni were trapped within 10 ft of each other on trunks or limbs from an old and rotting treefall lying across Sungai Sadaunta. ASE 2250 was taken on a trunk (6–8 in. diameter) 2 ft above stream; T. celebensis was trapped on an adjacent branch (4–5 in. diameter) next to a trunk (12 in. diameter) on which the R. hoffmanni was caught; branch and trunk about 4 ft above stream surface. All three elements are wet, slightly mossy, and free of other vegetation. Terraces on either side of the stream are rocky (the rocks covered with moss), muddy, and covered with chunks and tangles of rotting tree trunks and limbs, wanga palm trunks and fronds, all nearly concealed by thick shrubbery, ferns, and vines winding through the shrubs. Gingers are common and several banana trees were emerging above the shrubs. The higher terraces above the stream support understory figs, and tall emergent trees that canopy the stream. The limbs, branches and trunks that bridge the stream also rest on the bordering terraces, providing runways across the higher terraces above the stream. Stomach: partly full of chopped karoko.

225031 (2260) 823 Oct. 1974 On moss-covered, rotting limb lying over rocky terrace next to stream and extending across the water to opposite terrace. A rat, Rattus hoffmanni , was taken on same spot on a different night. The limb is 20 ft upstream from the treefall on which ASE 2250 was caught during the previous evening. Stomach: packed with pieces of karoko.

225040 (2272) 915 Oct. 1974 On thick, decaying limb bridging tributary of main stream. Limb is free of vegetation and moss, extends from one high eroded bank across the stream 2 ft above water level to a lower opposite bank; stream flows through deep ravine in hillside forest. Stomach: full of cut pieces of karoko.

225036 (2302) 793 Nov.1974 On wet, moss-covered decaying limb lying across main headwater tributary of Sungai Sadaunta. The canyon is V-shaped, covered with short forest typical of steep slopes near ridgetops. Understory figs are common along stream and form canopy over the water; other components of the streamside forest are scattered fishtail palms ( Caryota mitis ) and sugar palms ( Arenga pinnata ), with wanga palms ( Pigafetta filaris ) higher on the banks; shrubs, ferns, and rattan provide a dense understory; woody and monocot vines thread through the trees and palms. Stomach: nearly full with karoko.

225037 (2307) 991 Nov. 1974 On rotting, wet and smooth trunk bridging banks of tributary stream (suspended about a foot above the water surface) and downstream from where ASE 2302 was caught the previous night. Stomach: distended with freshly consumed karoko; mouth was also full of large chunks of the fungus; looks like rat had just finished eating, then ran across trunk lying across the stream.

225038 (2322) 1006 Nov. 1974 On wet, moss-covered tree root growing across tributary creek in steep ravine; caught rats ( Taeromys celebensis , Rattus facetus , and Paruromys dominator ) and Prosciurillus murinus just downstream on rotten trunk lying across stream. Stomach: half full, chunks of karoko.

224772 (2353) 976 Nov. 1974 A juvenile caught in live trap set on wet ground beneath canopy of wanga palms on narrow terrace 4 ft above stream; little ground cover, mostly scattered ferns and shrubs. Kept captive: offered a variety of foods but accepted and consumed only pieces of karoko (see text).

TABLE 72 (Continued)

Locality, AMNH

and (ASE-field) Elevation

numbers (m) Date Trap site and other information

225041 (2374) 999 Nov. 1974 Adult caught in live trap set on rotting, moss-covered wet trunk (5 in. diameter) lying across headwater tributary of Sungai Sadaunta, suspended about 2 ft above water surface; ravine slopes rocky and muddy, steep hillside forest above ravine. A rat, Paruromys dominator , was caught in same spot about a week before; kept captive. Kept captive: accepted mostly karoko, and a few invertebrates (see text).

225032 (2387) 823 Nov. 1974 On rotting, clean trunk (1.5 ft diameter) extending from one terrace across the Sungai Sadaunta to the other terrace. Karoko usually found growing on sides of the wet trunk. About two weeks before, caught a tree squirrel, Prosciurillus murinus in same trap. On the wet terrace beneath the trunk caught several Bunomys chrysocomus and a Rattus hoffmanni ; on limbs bridging the stream about 10 ft upstream caught a Maxomys hellwaldii and a M. musschenbroekii ; and about 30 ft downstream caught two Bunomys karokophilus , a Maxomys musschenbroekii , Rattus hoffmanni , Crunomys celebensis , Taeromys celebensis , and the shrew Crocidura elongata on decaying limbs lying across the stream. Streamside terraces are rocky, mossy, wet, and densely covered with shrubs, ferns, vines, scattered banana trees, rotting, moss-covered debris from old treefalls, and palm fronds; terraces and stream shaded by understory fig trees wanga palms ( Pigafetta filaris ) with tall forest in background. Stomach: packed with karoko.

225033 (2398) 823 Dec. 1974 On same rotting trunk where ASE 2260 was caught. Stomach: distended with chunks of karoko, looks freshly consumed.

226833 (4340) 976 Mar. 1976 On large decaying trunk (2 ft diameter, 20 ft long) covered with wet moss and small plants lying across a ravine containing a tributary of the Sungai Sadaunta, and suspended about 4 ft above the water surface. Ravine sides are steep and muddy; tall wanga palms dominant in a short streamside forest; ground washed, cover sparse. Stomach: full of karoko plus one partly digested segment of an earthworm; also a few pure white fibrous pieces in which are trapped small fly larvae.

Tomado

223045 (118) 1000 July. 1973 On ground near stream in primary forest. Stomach: partly full of chopped karoko; also a little bait.

223046 (139) 1000 July. 1973 Along wet, decaying tree trunk in understory of forest. Stomach: partly full of chopped karoko; one adult ant.

223056 (325) 1000 Aug. 1973 On ground deep in primary forest besides stream. Stomach: empty except for a few remnants of karoko and hairs ingested during grooming.

223058 (330) 1000 Aug. 1973 In dense undergrowth near small stream. Stomach: partly full of karoko; many oblong, flat, hard orange seeds (surrounded by translucent tissue); chewed remains of a large orthopteran (mostly semidigested tissue, parts of legs, and corregated wings).

223060 (342) 1000 Aug. 1973 In forest near small stream beneath shrubs. Stomach: full of karoko; also a few small flat, elliptical, flat orange seeds from some kind of fruit.

223072 (401) 1000 Aug. 1973 Beneath rotting, wet trunk lying on ground in tall forest partly and newly cleared of understory (preparatory to planting coffee trees). Stomach: partly full of karoko; large, long-stemmed fig seeds; remains of small spider (pedipalp, leg segments, bits of chitin).

226931 (469) 1000 Aug. 1973 On branch (8 in. diameter) that is part of a pile of branches, larger limbs, and trunk of a decaying treefall lying in understory of deep primary forest. Stomach: empty except for a bit of karoko remnants.

TABLE 72 (Continued)

where in the mountain block are obtained. Within the population sample I employed here the primary expression of variation in the variables measured is a reflection of the range in age from young adults to old adults.

All examples of B. karokophilus come from lowland tropical evergreen rain forest along the Sungai Sadaunta (823–1006 m) in the drainage of the larger Sungai Miu, and the vicinity of Tomado (1000 m) and Sungai Tokararu (1150 m), within the watershed of Danau Lindu. These voucher specimens document the known elevational range of B. karokophilus (table 5).

NATURAL HISTORY: This section contains the information I recorded in the field covering habitat, diet, and behavior.

Habitat: Bunomys karokophilus inhabits streamside forests (table 72; see description of streamside habitat under Natural History in the account of B. chrysocomus ). Most specimens were trapped on wet and decaying tree trunks and limbs, as well as rotting palm trunks, spanning streams, narrow tributary creeks, and wet ravines in dense primary forest—the habitat images captured in figures 96 and 97 are typical. Several rats were encountered on forested terraces above streams where traps were placed beneath wet, decaying trunks and limbs from old treefalls lying on the ground and partially concealed by dense vegetation (fig. 98). Air is cool and the ground wet in these shaded environments, wet and moss-covered rocks and remains of treefalls are nestled among gingers, palm rosettes, ground and tree ferns along with a variety of tall shrubs. We never caught B. karokophilus in traps placed on forested ridgetops or hillsides away from stream and wet ravine terraces. Along Sungai Sadaunta, at Tomado, and Sungai Tokararu, the three areas we encountered B. karokophilus , mean ambient air temperatures ranged from 61.1 ° to 80.9 ° F with relative humidity reaching 100 % (table 3).

All but two rats were caught in primary forest that was to my eyes undisturbed by human activities. One of those two exceptions was trapped on a decaying tree trunk lying across a small stream covered by intact forest on one bank but a coffee grove on the other. Species of large canopy and emergent trees shade the coffee trees and in their structure the groves are not unlike undisturbed forest with the coffee forming the understory beneath a tall canopy. Another animal was trapped just back from a stream flowing through forest being cleared prior to the planting of coffee trees. These groves were surrounded by tall forest. Our trapping efforts focused on undisturbed or only slightly modified forest habitats and we did not go out of our way to survey cut-over areas or secondary growth. Whether B. karokophilus can survive in such places is unknown. Certainly their presence would be influenced by secondary forest tall and dense enough to provide shade and the right ambient conditions for Auricularia delicata (karoko), their primary food, to flourish.

Our trapping records indicate that B. karokophilus is terrestrial. We never caught it in trees or on woody vines hanging low to the ground. The few captures above ground level were on decaying tree trunks lying on the forest floor where the rats could scamper along the ground and onto the trunk (the microhabitat configuration shown in fig. 98 is an example).

Diet: Bunomys karokophilus is mycophagus, concentrating on one kind of ear fungus; invertebrates, vertebrates, and fruit form a minor if not insignificant part of its diet. The dietary information discussed here and summarized in tables 13 and 72 is derived from the contents of stomachs removed from preserved specimens and the range of foods I offered two captives, an adult and a juvenile.

Fungi— Auricularia delicata is an ‘‘ear fungus’’ (Division Basidiomycota, Class Agaricomycetes, Order Auriculariales , Family Auriculariaceae ; species of Auricularia are also informally referred to as ‘‘jelly fungi’’ along with species in the orders Dacrymycetales and Tremellales ). The fruiting bodies range in size from small rubbery and oblong capsules (10 mm long, 5 mm wide) to broad and thick ear-shaped lobes (up to 50 mm wide), rubbery to the touch. The top surface is smooth and purplish or purplish brown, the undersurface white and irregularly ribbed and veined. The inside is gelatinous. Individual small and broad lobes may occur on the same wet and decaying tree trunk or limb, but usually several lobes form a cluster originating from a single point of attachment. The North American Auricularia auricula , known commonly as ‘‘tree-ear’’ or ‘‘wood-ear’’ (see plate 617 in Lincoff, 1997) is very similar in morphological form to the Sulawesi A. delicata . In central Sulawesi, ‘‘ karoko ’’ of the Kulawi language, refers both to human ears and A. delicata (see etymology for Bunomys karokophilus ).

The foods accepted by two captive B. karokophilus and contents of stomachs (table 72) indicate that karoko ( Auricularia delicata ) is the primary dietary component, at least during the seven months (March, August to January) in which my sample of B. karokophilus was collected. Except for one stomach containing only bait, all the other stomachs held some remains—from traces to masses—of karoko. Karoko grows only on wet and decaying tree trunks and limbs lying on the ground, usually on woody substrates free of other kinds of ear fungi, are not extensively covered in moss, and have not started to decay to the degree where soil has formed along the sides where the wood touches the wet ground supporting nurseries of small plants. Apparently the fungus requires a clear but irregular and cracked, wet, and decaying surface. Karoko also grows on trunks and limbs spanning streams, so the rats not only use these as bridges but also feed along them when there is a suitable fungal crop. Karoko is absent from wet and decomposing wanga palm trunks ( Pigafetta filaris , the common tall palm growing on stream and river terraces), which are invariably smooth, from decaying rattan vines, and from rotting pieces of limbs and trunks that lay close to the stream surface and are inundated by periodic flooding. My two helpers and I found karoko to be more common after rains and located on woody substrates in protected wet pockets on streamsides, stream terraces, and hillsides that remain wet and cool between rains.

During our daily forays in the forest, we always looked for karoko to gain an impression of its abundance. One day we walked about 3 km along the Sungai Sadaunta from camp at 913 m up to about 1100 m looking for karoko on decaying and wet wood bridging the stream and resting on banks and edges of terraces above the stream. We attempted to examine all rotting and saturated wood in those areas, but did not climb away from the stream onto the hillsides; previous to our search, there had been two full days and one night without rain. The fungus was scarce. We collected enough to fill stomachs of about three adult B. karokophilus . Much of it was in the form of young capsules that usually appear along one side of a decaying trunk. Presumably after a rain these would have grown into clusters of larger lobes. While the fungus at times may increase in abundance after periods of heavy rains, along the stream banks and terraces it was not common during the period of our survey, and I gained the impression that karoko is never plentiful because when checking traplines we always looked for fungal clusters and only infrequently saw them. A crop in one place may last several nights but then disappears after sporulation (some sections of trunks and limbs where we found karoko just after it sporulated remained bare of fungi for several weeks), so the availability is erratic and the source shifting. It seems that during the night rats would have to hunt constantly and perhaps over long sections of streams and ravines. This may explain in part their low population density compared with other species of murids in the same areas. For example between 800–1100 m along the transect, the elevational extent in which we encountered the 27 representatives of B. karokophilus , we trapped 184 examples of Bunomys chrysocomus in the same streamside habitats, and on stream terraces and the base of hillsides we obtained 118 Rattus hoffmanni , 91 Rattus facetus , and 84 Paruromys dominator . We used one kind of bait for all our traps and it seemed to attract B. karokophilus because bait was found in the stomach of one rat and unhesitatingly consumed by a captive adult.

During subsequent surveys, we discovered karoko growing on wet and shaded tree trunks lying on steep hillsides 75–100 ft above stream banks and terraces, but it was uncommon compared to the density near the streams and creeks. We set traps on the hillsides near the karoko sites but did not capture B. karokophilus .

Our experience with karoko was primarily in habitats within tropical lowland evergreen rain forest. My helpers and I looked for it in montane forest where we found the fungus only at one spot along the transect, growing on a wet and decomposing tree trunk spanning a narrow stream (Sungai Salubeka) at 1418 m on Gunung Kanino. A small cluster was growing next to a large white karoko ( Auricularia fuscosuccinea ), which although possessing a similar rubbery texture and gelatinous insides, takes the form of a bracken fungus; clumps of this white karoko were scattered along the trunk. The narrow valley of the stream is always wetter than the adjacent upper slopes and ridgetops and the forest mossier. During the short periods without rain (usually less than a week in my experience), the streamside environment remained wet while the higher slopes dried out. During the usual bouts of rainfall, every day or every other day during our camp on the ridge forming Gunung Kanino, the slopes remain wet and muddy and the forest constantly drips. Bunomys karokophilus does not live in this montane forest, but its relative, B. penitus , does and is also a consumer of fungus.

I watched the captive adult eat karoko. With its incisors the rat grasped the fungus and then transferred it to the front feet. Small capsules were held near the mouth and each piece was completely consumed. Large lobes were rested against the floor of the cage. The rat chewed around the edge of each rubbery lobe, then ate the folded whitish underside along with some of the inner gelatinous layer. Some of the larger lobes were first stripped on the underside where the skin is softer. The tougher upper skin was either discarded or folded between the front feet and then consumed in small bites. The rat ate only the firm, rubbery lobes and ignored those lobes beginning to transform into a jellylike mass (the transformation when the fungus sporulates). In most of the stomach contents I examined, the karoko consisted of the gelatinous insides, the cut and chewed sections of the white undersurface, and chunks of amber or purplish pieces from the lobe’s margin. The same adult rat was offered two kinds of reddish bracken fungi, a brown ear fungus with a thick rubbery rind and thin gelatinous inner layer (not a karoko), and a white bracken fungus—all were rejected.

The captive juvenile consumed only karoko, ignoring all other categories of food offered.

Karoko ’s strong appeal to Bunomys karokophilus is emphasized by a rat caught in a Conibear trap that had been set on a wet and decaying section of tree trunk bridging the Sungai Sadaunta. Its mouth was crammed with large chunks of the fungus, and the stomach was distended with freshly consumed karoko. Apparently the rat had just finished gorgeing on karoko to the degree that it could not swallow any more pieces.

Contents of the stomach from another rat caught along the Sungai Sadaunta provide a typical example of the amount of karoko in relation to fruit and invertebrates seen in those few stomachs that held remains of fungus along with other items. The stomach was full of karoko and also contained remains of fig (rind with small black seeds adhering to inner surface of the rind) and a chewed centipede (antennae, legs, and masticated body). Only eight of the 26 stomach samples I examined held remains of karoko and bits of fruit and/or invertebrates; one stomach contained bait only, and remains of karoko was the sole constituent in 17 others.

Earthworms —Judged from stomach contents and food preferences of the adult captive, earthworms are recognized as food but are an uncommon constituent of the diet. I offered a small earthworm (2 inches long) to the adult rat. With his incisors, he grabbed the worm from my fingers, transferred it to the front feet and consumed the entire worm in less than 30 seconds. On subsequent days I offered several earthworms of about the same size, but all were rejected. One day I placed a larger worm (4–5 inches long) in the cage, which the rat chased, now and again pulling it along the cage floor; finally the rat bit the worm in several places, grasped it with his front feet, and proceeded to devour the worm, taking 2–3 minutes to do so. Biting and pulling the worm through his front feet seemed awkward and slow compared to the actions characteristic of B. chrysocomus , an aggressive and voracious invertebrate predator that can consume an earthworm 4– 5 inches long in 10–20 seconds. I found a short, mascerated segment of an earthworm in one stomach (AMNH 226833) that otherwise contained mostly remains of karoko (see table 72), and a short piece from the posterior end of a worm in the stomach of another rat that also held some karoko, an intact ant and fragments of another, and pieces of a figlike fruit (AMNH 223316).

Snails —No stomach contained remains of snails. I offered small live snails to the captive adult. He ignored all of them. Finally I cracked open a shell and gave the extracted snail to the rat who ate only the soft portions, leaving on the cage floor the muscular foot and tougher tissue adhering to the foot. Bunomys chrysocomus , occurring in the same habitat and a voracious consumer of snails, not only bites open the shell but consumes the entire snail body.

Insects and other arthropods —Evidence from contens of stomachs suggests that insects and other arthropods are a minor part of B. karokophilus ’s diet (table 72). Of the 26 stomachs examined, only four held arthropod remains: one stomach contained an ant; another a few insect sclerites and parts of a wing; the antennae, legs, and partly digested body of a centipede was in a different stomach; and another contained bits of wing, legs, and tissue from a large orthopteran. The captive adult ate an adult cicadid offered. It ignored beetle larvae placed in the cage.

Vertebrates —I did not find remains of vertebrates in stomachs. I offered a small frog (1 inch long) to the adult rat, which he grabbed with his incisors and with his front feet manipulated the frog until its head was near the rat’s mouth. He began chewing on the head and consumed about half the frog, discarding the rest (part of the body and legs). The rat seemed to recognize the frog as food but handled it awkwardly and ate slowly, chewing on the frog rather than cutting it into pieces.

Fruit —Fruit, like invertebrates, is apparently not a staple in the diet of B. karokophilus . Of the 26 stomachs, one contained the remains of fig, two others held gelatinouscovered seeds from an unidentified fruit, and remains of a figlike fruit was found in another (table 72). I offered the captive adult the large fruits (1–2 inches long) from two species of understory streamside figs ( Ficus sp. ), and fruit from the wanga palm ( Pigafetta filaris ); all were ignored. However, the small, quarter-inch, wedge-shaped fig from Ficus minahassae was eaten. This is a common understory tree forming a partial understory canopy along some sections of stream banks and wet ravines. The figs are bright red and soft when ripe; the thick rind encloses a syrupy fluid filled with tiny pale orange seeds. The rat bit chunks of the rind off and discarded them, ingesting only the seeds and thick liquid in which they were suspended.

Nests and behavior: In addition to offering the captive adult different foods, I recorded the following observations. I kept the rat in a moderately large cage in which I placed dry leaves. For the first few days he would walk over the top of the leaf pile and eventually curl his body on top to sleep. Then one night he pushed the leaves into a pile at one end of the cage and burrowed into it forming a deep depression where he slept during the day. The nest was a mound of leaves 8 inches high and 10 inches wide in which there was an opening in the center and only a pink nose visible deep within. For the next few days the rat curled up in its nest with the head under the body in a vertical position rather than curled on the side so the head was lost to view and the rat became nothing more than a bluegray furry ball nestled deeply in the leaf pile.

The adult rat, like those I observed representing other species of Bunomys , groomed itself by first licking the palms repeatedly, often holding the wrist of one hand with the other to steady it while it wets the palm; then he rubbed around the nose and muzzle, licked the palms again, then reached back behind the ears and with both arms brushed forward over the ears down the sides and top of the head, over the eyes to the nose, with the fore-and upper arms acting as brushes and the palms on top sweeping over the head. Then he licked sides of the body near the thighs, the back and rump, and finally licked the inguinal area and genitals. I never saw him clean his tail.

During their captivity I never heard the adult or juvenile vocalize.

ECTOPARASITES: Records include only two species of fleas (table 14). Sigmactenus alticola pilosus (Leptopsyllidae) is also recorded from 14 other species of endemic Sulawesi murine rodents ( Bunomys penitus , B. prolatus , and B. chrysocomus ; Margaretamys elegans ; Maxomys hellwaldii , M. wattsi , and Maxomys sp. ; Melasmothrix naso and Tateomys rhinogradoides ; Paruromys dominator ; Taeromys celebensis and Taeromys sp. ; Rattus hoffmanni and R. facetus [recorded as R. marmosurus ]) and one nonnative rat, Rattus exulans ( Durden and Beaucournu, 2000) .

In addition to Bunomys karokophilus , five endemic Sulawesi murids ( Rattus hoffmanni ; Bunomys penitus and Bunomys chrysocomus ; Maxomys sp. ; Paruromys dominator ) and two nonnative rats ( Rattus exulans and R. nitidus ) are also hosts for Stivalius franciscae ( Stivaliidae ; Beaucournu and Durden, 2001).

SYNONYMS: None.

SUBFOSSILS: None.

With the discussion of Bunomys karokophilus , exposition of the eight species of Bunomys is complete. The four species of Bunomys we sampled within the murine community along the transect in the northern portion of the west-central mountain block, the single species in the lowlands of northeast-central Sulawesi, and all eight species as elements of the total murine fauna endemic to the island are the subject of the following two accounts.

BUNOMYS AND TWO MURINE COMMUNITIES IN CENTRAL SULAWESI

This section places Bunomys andrewsi , B. karokophilus , B. penitus , and B. chrysocomus in particular faunal contexts. I use results gained from surveys in two regions. One is along the transect in the northern part of the west-central mountain block in the forested landscape stretching between the lowlands in the lower drainage of the Sungai Miu and the mountains northeast of Danau Lindu (see the map in fig. 4), the other in coastal lowlands of the northeast-central region at Kuala Navusu (see the map in fig. 5). In both areas Aminudi, Usma, and I set the traps ourselves on the ground, in the understory, and on remains of treefalls spanning streams and ravines; recorded where each rat was caught, noting, among other observations, when examples of different species were taken in the same trap set during different nights; collected ectoparasites; noted reproductive characteristics; kept rats alive to watch their behavior, particularly the differences between arboreal and terrestrial species; identified contents of stomachs from preserved animals, and offered a range of forest foods to captives and recorded what was either accepted or rejected to enhance our understanding of diets; and learned firsthand something of the diverse floristic components comprising the habitats in which the rats were encountered. Information from the present report, published documents, and my field journals are summarized to form a general picture of the murine communities along my transect in the west-central mountain block, and east of there in the lowlands at Kuala Navusu.

Each of the other four species of Bunomys — B. fratrorum , B. prolatus , B. torajae , and B. coelestis —is also part of a murine community. I have not worked in the places where samples of those taxa were collected and, other than study of museum specimens, lack personal exposure in the field to their habitats and ecologies. An estimate of the species with which each is regionally sympatric can be determined from the information listed in table 81.