Bunomys fratrorum ( Thomas, 1896 )

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 : 148-169

Bunomys fratrorum ( Thomas, 1896)

Mus fratrorum Thomas, 1896: 246 .

HOLOTYPE: BMNH 97.1.2.28, the skin and skull of a young adult female collected October 8, 1895, by C. Hose and E. Hose. Measurements (external, cranial, and dental), along with other information, are listed in table 40. The skin is stuffed in the conventional museum fashion and shows no significant alteration in color of fur. The cranium and mandible are intact, and all incisors and molars are present.

TYPE LOCALITY: Rurukan (01 ° 21 9 N, 124 52 9 E), 3500 ft (1068 m; locality 7 in gazetteer and on the map in fig. 50), northeastern tip of the northern peninsula of Sulawesi, Propinsi Sulawesi Utara, Indonesia .

EMENDED DIAGNOSIS: One of the largerbodied members of the B. fratrorum group ( LHB 5 157–190 mm, WT 5 175 g, ONL 5 41.5–46.5 mm) and further set apart by the following combination of traits: (1) dorsal fur brownish gray speckled with buff, ventral coat grayish white tinged with pale buff; (2) tail typically equal to combined length of head and body (99 % –101 %), brown on dorsal surface, white or mottled brown on ventral surface, most with a white or speckled tip (86 % of the 99 specimens surveyed) that is moderately long relative to tail length (mean 5 18.0 %, range 5 4 % –75 %); (3) testes small relative to body size (8 %); (4) sperm head short, asymmetrical and gently curved in outline, with a short tail attached near the middle of its concave surface; (5) cranium and mandible robust (6) rostrum and zygomatic plate wide, interorbit narrow, braincase deep, postpalatal region (basicranium) short, and ectotympanic bulla small relative to size of cranium; (7) molar row relatively long; (8) cusp t3 absent from second and third upper molars in most of sample (89 % and 91 %, respectively); (9) anterolabial cusp rare on second lower molars (12 %) and absent from third lower molars; and (10) posterior labial cusplet present on first lower molar in about half of sample and on third lower molar in three-fourths of sample.

GEOGRAPHIC AND ELEVATIONAL DISTRI- BUTIONS: Bunomys fratrorum is endemic to the northeastern section of the northern peninsula where it is documented by voucher specimens collected at localities scattered from Teteamoet, near the tip of the peninsula, west to near Gunung Mogogonipa (see the map in fig. 50). This region is on the peninsular mainland east of Gorontalo (00 ° 31 9 N, 123 ° 03 9 E) where the distribution of B. fratrorum is concordant with that of several other mammals: the macaques Macaca nigra and M. nigrescens ( Fooden, 1969; Groves, 1980, 2005); and the murids, Echiothrix leucurus , Taeromys taerae , Rattus xanthurus , and R. marmosurus ( Musser and Carleton, 2005; also see table 81). Bunomys fratrorum has not been collected from islands adjacent to the Sulawesi mainland.

Known elevational range extends from lowlands near sea level (Teteamoet, for example) to about 2000 m on Gunung Klabat (table 5).

SYMPATRY WITH OTHER BUNOMYS : Bunomys fratrorum is the only member of the B. fratrorum group found on the northern arm east of the Gorontalo region. There it is sympatric with B. chrysocomus , which also occurs elsewhere on Sulawesi (see that account and table 6). C.H.S. Watts collected examples of both species 1 km north of Gunung Mogogonipa at 250 m during August 2–8, 1985 (table 20). Bunomys fratrorum and B. chrysocomus are the only species in the genus recorded from the northeastern tip of the northern peninsula east of the Gorontalo region.

DESCRIPTION: As is usual in his descriptions of new taxa, Thomas (1896: 246) captured the essence of Mus fratrorum :

Size about as in Mus rattus . Fur of medium length, soft, without longer bristles intermixed. Ears large, evenly rounded. General colour above brownish grey, finely sprinkled with dull yellowish. Head rather paler, but the circumference of each eye slightly darker. Under surface dirty greyish yellow, the bases of the hairs slaty grey, their tips dull yellowish; line of demarcation on sides little marked. Hind feet rather elongate; fifth hind toe reaching to the middle of the first phalanx of the fourth; upper surface of hands and feet silvery white. Tail somewhat shorter than head and body, finely scaled (about 11scales to the centimetre), its proximal two thirds above brown, its end and the whole of its under surface white. Skull with rather a small cranial and long facial portion. Subraorbital beads distinct, although not strong. Front of zygoma-root little projected forwards. Anterior palatal foramina of medium length, not reaching back to thelevel of [first upper molar]. Molars large and heavy. Bullae small.

After providing measurements, Thomas noted (1896: 247) that besides ‘‘the type, there are several other specimens from Rurukan, two from Menado, and one from Mount Masarang. This species is apparently most closely allied to M. chrysocomus, Hoffm. , also a native of Celebes, but differs from it by its larger size, beaded supraorbital edges, and much heavier molars.’’ Apparently Thomas’s perception of the close morphological relationship between fratrorum and chrysocomus prompted him to propose ‘‘ fratrorum ,’’ which is derived from the Latin frater meaning ‘‘brother.’’

TABLE 41

Descriptive Statistics for Measurements (mm) of Lengths of Head and Body, Tail, Hind Foot, and Ear, and for

Weight (g), Derived from Samples of Bunomys fratrorum , B. andrewsi , B. penitus , and B. karokophilus n. sp.

Mean ± 1 SD, observed range (in parentheses), and size of sample are listed. Mean values were used to compute LT/LHB. Specimens measured are listed in footnotes.

Species and sample LHB LT LHF LE WT LT/LHB (%) a

B. fratrorum

Teteamoet 155.0 ± 5.91 178.5 ± 11.64 39.9 ± 1.24 — — 101 (165–187) 40 (150–200) 39 (38–44) 40

Kuala Prang 177.3 ± 6.99 179.0 ± 12.86 40.4 ± 1.27 — — 101 (165–190) 21 (155–195) 21 (38–43) 21

Temboan 165.8 ±.86 167.5 ± 8.06 39.6 ± 0.99 — — 99 (157–182) 34 (150–180) 33 (36–41) 34

Gunung Maujat 183.5 ± 3.54 159.0 ± 1.41 39.5 ± 0.71 27.0 ± 1.41 175.0 ± 0.00 87 (181–186) 2 (158–160) 2 (39–40) 2 (26–28) 2 2

B. andrewsi b

Pulau Buton 148.5 ± 2.12 111.5 ± 2.12 37.5 ± 2.12 24 — 75 (147–150) 2 (110–113) 2 (36–39) 2

Wawo + Masembo c 170.1 ± 15.20 148.6 ± 7.02 39.3 ± 2.25 — — 87 (153–186) 8 (144–160) 7 (36–41) 8

Kuala Navusu 177.4 ± 11.62 150.1 ± 9.92 40.5 ± 1.33 24.7 ± 1.00 154.6 ± 33.43 85 (157–195) 17 (130–161) 15 (38–42) 17 (23–26) 17 (98–222) 17

Pinedapa 163.1 ± 8.59 145.8 ± 8.93 37.1 ± 1.62 — — 89 (145–175) 17 (130–165) 17 (35–40) 17

Gunung Balease 171.9 ± 7.54 156.8 ± 5.96 40.2 ± 1.75 26.6 ± 1.93 149.6 ± 21.60 91 (157–181) 11 (148–165) 12 (37–44) 12 (24–30) 12 (121–177) 11

Malili Area 169.9 ± 9.81 153.0 ± 6.19 40.4 ± 1.08 27.1 ± 1.51 145.0 ± 19.22 90 (155–186) 12 (145–167) 12 (39–42) 12 (24–29) 11 (120–179) 12

Puro-Sungai Miu 163.1 ± 8.24 146.4 ± 10.86 38.1 ± 1.27 24.4 ± 1.24 113.7 ± 11.50 89 (151–177) 9 (130–165) 9 (36–40) 9 (23–27) 9 (95–135) 9

Tamalanti 162.5 ± 4.18 139.2 ± 5.85 36.0 ± 1.10 24.5 ± 1.23 — 86 (160–170) 6 (130–145) 6 (35–38) 6 (22–25) 6

Tuare 147.5 ± 3.54 135.0 ± 0.00 35.0 ± 0.00 — — 92 (145–150) 2 2 2

Gunung 156.1 ± 5.15 142.9 ± 5.52 35.4 ± 0.55 23.3 ± 0.76 111.7 ± 19.48 92

Gandangdewata (149–164) 7 (134–150) 7 (35–36) 7 (22–24) 7 (88–144) 7

Lombasang c 148.8 ± 7.74 137.4 ± 8.19 35.8 ± 0.80 — — 92 (137–162) 14 (129–153) 13 (35–37) 14

d

B. penitus

Gunung Lehio 168.0 ± 2.92 171.8 ± 19.47 40.4 ± 0.89 — — 102 (165–172) 5 (140–190) 5 (39–41) 5

Rano Rano 170.8 ± 7.89 161.3 ± 8.54 38.8 ± 1.89 — — 94 (160–178) 4 (150–170) 4 (36–40) 4

Gunung Kanino 182.6 ± 7.34 169.1 ± 8.24 41.8 ± 1.40 26.2 ± 1.03 132.3 ± 14.47 93 (167–199) 74 (153–188) 71 (38–45) 74 (24–29) 74 (100–170) 74

Gunung 184.0 ± 10.45 162.5 ± 8.63 41.4 ± 1.30 26.1 ± 0.91 133.3 ± 15.42 88

Nokilalaki (155–242) 82 (138–185) 76 (38–44) 82 (23–28) 82 (95–170) 76

Pegunungan 179.2 ± 10.21 171.7 ± 11.26 39.7 ± 1.37 25.0 ± 0.00 — 96

Quarles (160–190) 6 (155–175) 6 (38–41) 6 6

Gunung 176.8 ± 14.34 180.5 ± 14.48 42.0 ± 0.82 26.0 ± 0.82 169.3 ± 31.38 102

Gandangdewata (160–195) 4 (164–198) 4 (41–43) 4 (25–27) 4 (137–212) 4

Pegunungan 167.8 ± 10.21 159.8 ± 10.26 39.8 ± 0.45 — — 95

Latimojong e (154–179) 5 (143–171) 5 (39–40) 5

Pegunungan 159.3 ± 7.79 160.9 ± 8.29 37.8 ± 1.44 — — 101

Mekongga e (145–173) 19 (146–179) 18 (36–40) 19

TABLE 41

(Continued)

Thomas knew the characteristics of chrysocomus , at least as reflected in the type specimen, because during a tour of several European museums in the late 1800s, he examined the holotype of chrysocomus at Dresden during October, 1887, and recorded his observations and measurements in a journal, which is in the library of the mammal section at BMNH (a photographic copy was made by G.H.H. Tate in 1937 and is stored in the Mammalogy Archives at AMNH). It was clear to Thomas that fratrorum was a species distinct from chrysocomus and that both occurred on the northeastern region of the northern peninsula of Sulawesi.

The large-bodied B. fratrorum (LHB 5 157–190 mm, LT 5 150–200 mm, LHF 5 36–44 mm; ONL 5 41.5–46.5 mm) has soft, long (12–15 mm long), and glistening brownish-gray dorsal fur brightly speckled with buff (‘‘finely sprinkled with dull yellowish,’’ as Thomas wrote). In a large sample, the tonal range extends from brownish gray to a dark brownish gray with blackish highlights. Fur covering the underparts is dark gray and either speckled with buff (Thomas’s ‘‘dirty greyish yellow’’), which is usual, or heavily washed with buffy tones. Ears are brownish gray, rubbery in texture (noted in a fluidpreserved example), their inner and outer surfaces covered with short, fine hairs.

Dark brown above, white or speckled below, and around the tip is the basic coloration of the relatively long tail, which is typically as long as the head and body in most samples (LT/LHB 5 99 % –101 %). Most specimens (86 % of the 99 specimens surveyed) exhibit a white or speckled tip that is moderately long (mean 5 18.0 %, range 5 4 % –75 % of tail length in 85 specimens; table 8). Color pattern varies in any large sample. The dorsal surface and sides of the tail are brown in all specimens, the ventral surface is either glistening white the entire tail length or lightly to heavily speckled with tan; the tail tip is either white or speckled to some degree. Thirty-three males and females from Temboan (locality 17) illustrate the variation. The dorsal surface and sides of the tail range from brown to dark brown, the ventral surface is white in 16 specimens and lightly to heavily speckled in the other 17; 31 have a whitish tip that is entirely white in 25 rats, but speckled in six.

Dorsal surfaces of front and hind feet are white and covered with short, fine silvery hairs (some specimens are brown from grease stains). Claws are short and unpigmented, and not concealed by the meager silvery ungual tufts. Palmar and plantar surfaces are naked and gray.

Females possess four teats configured in two inguinal pairs. Males have small testes relative to body size (8 %; see table 9). Gross morphology and ultrastructural anatomy of the spermatozoa are described by Breed (2004).

Bunomys fratrorum has a large skull, its overall dimensions matched only by B. penitus , with a moderately long and narrow rostrum and a wide zygomatic plate (figs. 52– 54; tables 42, 45). The interorbit is narrow and its dorsolateral margins framed by moderately high ridges, the braincase wide and deep, the incisive foramina short and wide, and the bullae small. Each dentary is robust and similar in shape to those of the other species in the B. fratrorum group.

Large molars with simple occlusal patterns are typical of B. fratrorum (fig. 12). Contributing to this simplicity is the absence of cusp t3 from the second and third upper molars in 89 % –91 % of the sample (table 10). As in most other species of Bunomys , an anterior labial cusplet is absent from the first lower molar, but a posterior labial cusplet typically forms part of the chewing surface on the first lower molar in about half of the sample and on the third lower molar in 78 % of the sample (fig. 12; table 11). An anterolabial cusp is absent from the second lower molar in most specimens (present on only 12 of 58 individuals) and was not detected on any third lower molar.

KARYOTYPE: No information.

COMPARISONS: Bunomys fratrorum first requires comparison with B. chrysocomus .

TABLE 42

Descriptive Statistics for Cranial and Dental Measurements (mm) derived from Population Samples for Bunomys chrysocomus and the Four Species in the B. fratrorum Group

Mean ± 1 SD and observed range (in parentheses) are listed.

The two species occur sympatrically on the northern peninsula east of Gorontalo, and fratrorum for a while was regarded as a synonym of B. chrysocomus ( Ellerman, 1949; Laurie and Hill, 1954; table 4).

Demonstrating that B. fratrorum is not just a geographic population of one of the members of the large-bodied B. fratrorum group found on Sulawesi south of the northern peninsula demands its contrast with B. penitus , B. andrewsi , and B. karokophilus , n. sp. Here I will contrast the characteristics of B. fratrorum first with B. chrysocomus , then with B. andrewsi followed by B. penitus . Comparisons between B. fratrorum and B. karokophilus , n. sp., will be entertained in the account in which the new species is described.

Bunomys fratrorum and B. chrysocomus: In the years between 1896 and 1941, fratrorum was recorded in the scientific literature by most taxonomists as a distinct species and first placed in Mus ( Trouessart, 1897: 485; Matschie, 1900: 286), then either Rattus ( Tate, 1936: 551; Ellerman, 1941: 190) or Frateromys ( Sody, 1941: 316) .

Two reports during this period provided different allocations of fratrorum . Meyer (1899: 24) listed Mus fratrorum , as well as Mus chrysocomus , as synonyms of Mus callitrichus , and Miller and Hollister (1921a: 72) ‘‘included’’ fratrorum in ‘‘ Rattus chrysocomus .’’ The reasoning behind Meyer’s action is explained in the first paragraph of his account of ‘‘ Mus callitrichus ’’ ( Meyer, 1899: 24; translated from the German by E.M. Brothers; the German text is reproduced in appendix 2):

As Hoffmann described for M. chrysocomus , the Dresden Museum possesses no examples of callitrichus , as had been indicated in Jentink’s account. However, four specimens were uncovered in 1894, which the good Doctor Jentink was kind enough to compare with his types and thus one can determine their identity with certainty, even though the description of the species (NLM 1879 I, 12) is not comprehensive enough for one using it alone to know how to identify them with complete confidence. Between the sole available chrysocomus (the type) and the examples of callitrichus now before me, I can in some fashion diagnose essential differences, which are as few as exist between M. fratrorum Thos. (two in the Dresden Museum which Thomas identified) and callitrichus . [The two identified by Thomas were collected by Charles and Ernest Hose in 1895 from Rurukan—see locality 7 in gazetteer—and formed part of the series upon which Thomas based his original description of Mus fratrorum . Thomas noted that in addition to the holotype, which had been obtained at Rurukan, he had studied several other specimens from the same place. Two of these had been sent to Dresden and are now cataloged as B.3384 and B.3386. Written on the skin label of each is ‘‘ fratrorum Thos. ’’ in Thomas’s handwriting.] Of these he said (AMNH XVIII, 247) that fratrorum [and] M. chrysocomus are very closely allied, but [ fratrorum ] is distinctive on the basis of overall size, a beaded supraborbital rim, and enormous molars. However, the two skulls available to me do not exhibit these beadings, and they have rather sharp rims as is customary. Overall size and huge molars can not be regarded as species differences, these notable differences arise with regard to age and sex.

I borrowed eight specimens that Meyer had identified as ‘‘ Mus callitrichus ,’’ which are stored in the Staatliches Naturhistorische Sammlungen Dresden, Museum für Tierkunde. Two are from Rurukan (locality 7 in the gazetteer) and were part of Thomas’s original series of fratrorum ; the rest were collected at Lota (locality 4 in the gazetteer). All are examples of Bunomys fratrorum . Meyer was able to contrast the holotype of chrysocomus with specimens that Thomas had identified as fratrorum , but the distinctive size differences in cranial and dental dimensions between specimens in the two samples that had stimulated Thomas to name the larger animal as a new species (remember that Thomas had also studied the holotype of chrysocomus ) were interpreted by Meyer as the kind of differences related to age and sex found in a sample of a single species. Furthermore, Meyer considered chrysocomus and fratrorum to be synonyms of Jentink’s (1879) Mus callitrichus .

Meyer had accepted Jentink’s identification of the four specimens sent to him (presumably four of the examples identified here as fratrorum ). Jentink compared Meyer’s material with the type series of Mus callitrichus , which was stored in the museum at Leiden, and found no noteworthy differences. This is not surprising. The original types associated with M. callitrichus consisted of 12 specimens and represented not one species but five: three specimens of Taeromys callitrichus , one example of Paruromys dominator , three specimens of Bunomys chrysocomus , four Bunomys fratrorum , and a single example of Rattus hoffmanni ( Musser, 1970) . To Jentink, the characteristics of Meyer’s specimens fit within the range of variation exhibited by the type material of M. callitrichus .

In 1921, Thomas (1921: 111) rebutted Meyer’s (1899) synonymy: ‘‘In his large paper on the mammals of Celebes … Dr. A.B. Meyer has placed both chrysocomus and fratrorum as synonyms of R. callitrichus, Jent. , but all these are perfectly distinct differing considerably in size and having quite appreciable diagnostic skull characters.’’ The statement was quoted 20 years later by Sody (1941: 316), who also noted that ‘‘This certainly is correct (as I was able to control when studying the type of Mus chrysocomus at the Dresden Museum).’’ In addition to Thomas (1896) and Meyer (1899), Sody was the only other researcher who had actually studied the holotype of Mus chrysocomus , contrasted it with examples of fratrorum , and published their observations.

In the same year that Thomas (1921) reasserted the distinctness of fratrorum as compared with chrysocomus, Miller and Hollister (1921a) ‘‘included’’ (their word) fratrorum within chrysocomus , but did not explain why. I suspect their view derived from study of the mammals collected in northeastern Sulawesi by H.C. Raven during the early 1900s. Raven obtained large samples of a species from Teteamoet, Kuala Prang, Gunung Klabat, and Temboan (localities 1, 5, 6, and 14 in the gazetteer and on the map in fig. 50) that were idenfitifed by Miller and Hollister as ‘‘ Rattus chrysocomus .’’ These are the only examples of a species in what has been called the Rattus chrysocomus group (see Tate, 1936, for example, and table 4), now recognized as Bunomys , that were collected by Raven during his survey in northeastern Sulawesi. Neither Miller nor Hollister had ever seen an actual chrysocomus from the northeastern end of the northern peninsula. At the time none were present in collections of museums in the United States, and the species was represented only by specimens collected during the late 1800s in institutions outside that country: the holotype in Dresden ( Hoffmann, 1887) and three of the 12 specimens at Leiden that had been designated cotypes of Mus callitrichus by Jentink in 1879 ( Musser, 1970). It would not be until 1985 that chrysocomus would again be encountered by collectors from northeastern Sulawesi (at two sites in Bogani Nani Wartabone National Park, localities 3 and 4 in gazetteer for Bunomys chrysocomus and on the map in fig. 22). Because Raven obtained only one kind of Bunomys in the northeastern peninsula, a species which seemed to be common, Miller and Hollister apparently assumed it was chrysocomus (the older name) and the same animal that Thomas had described as Mus fratrorum . Thomas’s description of fratrorum does match all of Raven’s specimens, and most external, cranial, and dental dimensions of the holotype of fratrorum , but not the holotype of chrysocomus , fall within the range of variation shown by Raven’s samples (tables 19, 40–42). Raven had collected examples of the large-bodied fratrorum and surprisingly had not encountered the smaller-bodied chrysocomus .

Influential checklists published after 1941, and originating primarily from the Natural History Museum in London, treated fratrorum as a synonym of Rattus chrysocomus ( Ellerman, 1947: 265, 1949: 52; Laurie and Hill, 1954: 118). But by 1970 its separate status as a distinct species, and reaffirmation of Thomas’s observations, was promulgated by Musser (1970, 1981 b, 1991), Musser and Newcomb (1983: 393), Musser and Holden (1991: 406), and accepted by Corbet and Hill (1991, 1992) in their world list of mammalian species and systematic review of Indomalayan mammals, and cemented in place by Musser and Carleton (1993: 582; 2005: 1300).

Bunomys fratrorum and B. chrysocomus are easily distinguished by body size alone and that this dimensional contrast could be interpreted as variation within a single species as Meyer did is remarkable. Bunomys fratrorum is physically larger than B. chrysocomus with longer head and body, tail, and hind foot (values for B. chrysocomus in table 19 and those for B. fratrorum in table 41). The tail in B. fratrorum is about equal to length of head and body (LT/LHB 5 99 % –101 %), while B. chrysocomus has a relatively shorter tail (LT/LHB 5 84 % –98 %). Color pattern of the tails also differ. In both species, tails are brown or brownish gray over their dorsal surfaces. Tails end in a moderately long white tip relative to tail length (mean 5 18.0 %, range 5 4 % –75 %) in most specimens of B. fratrorum (86 % of total specimens surveyed), and the ventral surface of the tail is white or mottled with brown (the pigment is contained in the scales and tail hairs). By contrast, a short white tail tip occurs in only 20 % of all specimens of B. chrysocomus surveyed and is short relative to tail length (mean 5 5.5 %, range 5 1 % –25 %). The undersurface of the tail ranges from brown to heavily mottled brown in most samples—a white ventral surface is uncommon.

Testes size provides a strong contrast between the species (table 9). Relative to length of head and body, the testes of B. chrysocomus are longer (22 %), those of B. fratrorum much shorter (8 %), a difference also indicated by relative testes mass. Testes mass relative to body mass was measured by Breed and Taylor (2000) and reported as 3.02 % in B. chrysocomus (N 5 2) and 0.28 % in B. fratrorum (N 5 2).

The two species also differ in spermatozo- an morphology. In B. chrysocomus , the falciform sperm head is long and thin with a long tail attaching near the end of the head, but in B. fratrorum , the sperm head is shorter and gently curved, with the tail (which is much shorter than in B. chrysocomus ) attached near the middle of its concave surface ( Breed and Musser, 1991; Breed and Taylor, 2000).

With the exception of three variables, all cranial and dental dimensions measured are absolutely much greater in B. fratrorum than in B. chrysocomus , a distinction qualitatively marked in illustrations of skulls (figs. 16, 17) and quantitatively expressed by univariate means for 232 B. chrysocomus and 100 B. fratrorum (table 42). Among the three dimensional exceptions, the interorbit is absolutely wider in B. chrysocomus (M 5 6.4 mm) than in B. fratrorum (M 5 6.2 mm), and breadth of the bony palate (M 5 3.8 mm in both species) and length of bulla (M 5 6.4 mm for B. chrysocomus , 6.5 mm for B. fratrorum ) are statistically identical in the two species. Set against the absolute size contrast between the two, B. chrysocomus has a conspicuously wider interobital region and bony palate, and a longer bulla relative to all other cranial and dental dimensions measured, proportional distinctions diagrammed previously (fig. 19).

Lower molars of B. fratrorum have a somewhat simpler occlusal topography than do those of B. chrysocomus judged by frequencies of certain cusps and cusplets (table 11). An anterolabial cusp is present on the second molar in only 12 % of the sample and all the specimens lack the cusp on the third molar; this cusp is part of the occlusal surface of the second molar in 90 % of the sample of B. chrysocomus and occurs on the third molar in 65 % of the sample. An anterior labial cusplet is not found on the first molar of all examples of B. fratrorum surveyed, but is usual in about half the sample of B. chrysocomus . The posterior labial cusplet of the first molar occurs in about half the sample of B. fratrorum , but is present in 97 % of the series of B. chrysocomus .

Thomas was impressed by the prominent (‘‘beaded’’) supraorbital ridges in B. fratrorum compared to B. chrysocomus . Interorbital and postorbital margins are delimited by low ridges on the skulls of both species, and those in B. chrysocomus are not as prominent as in B. fratrorum (figs. 16, 17). The absolute difference is related to size, the much larger skull of B. fratrorum carries with it heavier ridging than does the smaller skull of B. chrysocomus ; relative to size of skull, however, prominence of the interorbital and postorbital borders is similar.

The conspicuous contrasts between B. fratrorum and B. chrysocomus in body size, relative length of tail, length of the white tail tip and its frequency in populations, cranial and dental dimensions in absolute and proportional terms, frequencies of cusp and cusplets on lower molars, structure of spermatozoa, relative testes size, and sympatry should dispel any notion that the samples of each represent a single species.

Bunomys fratrorum and B. andrewsi : B. andrewsi has never been collected on the northern peninsula of Sulawesi. It ranges from the southern end of the northern peninsula over the lowlands and highlands of Sulawesi’s core, through the southeastern peninsula and on Pulau Peleng at the end of that arm, and into the southwestern peninsula (see the map in fig. 50). Bunomys andrewsi and B. fratrorum are similar in some phenetic characteristics. Both have brownishgray upperparts lightly speckled with buff and black and grayish-white underparts tinted or washed with a range of buffy tones. Bunomys fratrorum is comparable in lengths of head and body and hind foot to B. andrewsi (table 41). The most conspicuous external difference is expressed in length of tail and its color pattern. The tail is long (means 5 167.5–179.0 mm in six population samples), coequal to length of head and body in samples of B. fratrorum (LT/LHB 5 99 % – 101 %), and most individuals (86 % of the sample) show a long and white terminal portion (mean 5 18.0 %, range 5 4 % –75 % of tail length). By contrast, the tail is absolutely shorter in B. andrewsi (mean 5 111.5–156.8 for all population samples; table 41), much shorter relative to length of head and body (LT/LHB 5 75 % –92 %), a white tip occurs infrequently in the samples (20 % of 133 specimens), and when present is short relative to length of tail (mean 5 7.7 %, range 5 1 % – 13 %; table 8).

Relative testes size and spermatozoal morphology differ (table 9). Size of testes relative to body size is less in B. fratrorum (8 %) compared with the range in B. andrewsi (8 % –15 %). Spermatozoa of B. fratrorum have a short and gently curved sperm head to which a short tail is attached near the middle of its concave surface ( Breed and Taylor, 2000) while B. andrewsi resembles B. chrysocomus is in having a falciform sperm head that is long and thin with a moderately long tail attaching near the base of the head ( Breed and Musser, 1991).

Most of the cranial and dental measurements derived from the samples of B. fratrorum average greater than in the combined population samples of B. andrewsi (table 42). Interorbital breadth and expanse of the incisive foramina are exceptions. Bunomys fratrorum has a narrower interorbit (mean 5 6.2 mm) compared with B. andrewsi (mean 5 6.7 mm) and shorter incisive foramina (mean 5 7.3 mm for B. fratrorum , 8.0 mm for B. andrewsi ), contrasts evident when skulls of comparable adult age classes are compared (figs. 52–54).

A scatter plot of specimen scores representing B. fratrorum and B. andrewsi project- ed on first and second principal components summarizes the morphometric distinctions between the two species (fig. 55, upper graph). The major axes of the two elliptical clouds of scores are phenetically discrete: the regression lines of the second principal component on the first are parallel, their Yintercepts are significantly different between the two species (+0.951 versus ‾ 0.951; F 5 73.62, P 5 0.000), but their slopes are statis- tically equivalent (‾ 0.406 versus ‾ 0.407; F 5 0.00, P 5 0.998). Although there is marginal overlap between the two ellipsoidal clusters of scores along the second axis, there is enough separation to indicate that B. andrewsi has a relatively wider interorbit, narrower rostrum, narrower braincase, shorter but wider bony palate, narrower mesopterydoid fossa, markedly longer and wider incisive foramina, larger bulla and smaller molars compared with B. fratrorum (see the correlations in table 43). The ratio diagram in figure 56 presents most of these same proportional contrasts in graphic form but also shows the braincase of B. andrewsi to be wider relative to its height compared with B. fratrorum , and the molars smaller relative to expanse of the palatal bridge.

Frequencies of two structures on the mandibular molars differs betwen samples of B. fratrorum and those representing B. andrewsi (table 11). An anterolabial cusp is present on the second molar in only 12 % of the sample of B. fratrorum , but occurs in 88 % –100 % of the combined samples of B. andrewsi . About half of the sample of B. fratrorum has a posterior labial cusplet on the first molar, but nearly all specimens of B. andrewsi bear this structure (97 % –100 % of all specimens surveyed).

The morphometric distinctiveness of B. fratrorum relative to B. andrewsi is reinforced by a phenogram based on squared Mahalanobis distances among centroids for all population samples representing Bunomys (fig. 21). The cluster of samples identifying B. fratrorum unites with that of B. karokophilus , n. sp., and those samples link with the clusters of population samples for B. andrewsi and B. penitus . There is no evidence from morphology of skin and skull suggesting that B. fratrorum is a fragment of B. andrewsi isolated in the northeastern segment of the northern peninsula.

Bunomys fratrorum and Bunomys penitus: Within the forests mantling the northeastern end of the northern peninsula, the altitudinal distribution of B. fratrorum stretches from lowlands into mountains. Bunomys penitus inhabits only mountain forests clothing highlands in the west-central mountain block of the island and on the southeastern peninsula. The two species are smililar in body size, but the tail averages slightly longer and is generally longer relative to head and body length in B. fratrorum compared with B. penitus (table 41).

The dorsal pelage of B. penitus is dense and long (up to 25 mm), soft to the touch, silky in appearance, and brownish gray tending toward grayish tones (coat is not as long in B. fratrorum , not longer than 12 mm in most specimens, somewhat courser in texture and not silky, and brownish gray tending toward darker tones). Underparts are dark grayish white (underparts are also grayish white in B. fratrorum but tinged or washed with buffy tones in many specimens). The tail ends in a long white segment, behind it the dorsal

TABLE 43 Results of Principal-Components Analysis Comparing All Population Samples of Bunomys fratrorum with Those of B. andrewsi and B. penitus Correlations (loadings) of 16 cranial and two dental log-transformed variables are based on 100 B. fratrorum , 98 B. andrewsi , and 185 B. penitus ; see figure 55.

surface is glossy grayish brown or brownish gray, and the ventral surface is pure white (dorsal surface is brown in B. fratrorum except for a white terminal segment, the ventral surface ranges from pure white to white speckled with brown to mottled white and brown). Although a white-tipped tail is characteristic of both species (table 8), the white segment is longer relative to tail length in B. penitus (means 5 12.3 % –30.7 % for six samples of B. penitus ; 15.0 % –22.0 % for four samples of B. fratrorum ) and never speckled—some portion of specimens in most samples of B. fratrorum exhibit a speckled distal tail segment. Dorsal surfaces of the feet are white or white suffused with gray; the claws are delicate and nearly concealed by long, dense, and silvery ungual tufts (metatarsal and metacarpal surfaces range from white to brownish white in B. fratrorum , the claws are sturdy and not concealed by ungual tufts.

Testes size relative to head and body is similar in the two species (table 9). Sperm morphology is different. Spermatozoa of B. fratrorum have a short and gently curved sperm head to which a short tail is attached near the middle of its concave surface ( Breed and Taylor, 2000) while B. penitus has a falciform sperm head that is long and thin with a moderately long tail attaching near the base of the head ( Breed and Musser, 1991).

The skull in the sample of Bunomys penitus averages overall somewhat smaller (indicated by means for occipitonasal length and zygomatic breadth) than that of B. fratrorum , the braincase is not as deep, the zygomatic plate startlingly narrower, the bony palate narrower, and the postpalatal region (basicranium) shorter, as reflected by absolute differences in univariate means (table 42). Means for three internal dimensions are not significantly different between the two species (breadths of rostrum and braincase, and length of diastema), but the interorbit of B. penitus is wider, the rostrum and bony palate longer, the mesopterygoid fossa more spacious, the incisive foramina longer and broader, the ectotympanic bulla larger, and the molars heavier (longer molar row, wider molars).

Separate clouds of specimen scores for B. penitus and B. fratrorum projected on first and second principal components provide a multivariate summary of the contrast between the two species in cranial and dental variables (fig. 55, lower graph). Covariation in interorbital breadth, along with lengths of rostrum, bony palate, incisive foramina, bullar capsule, and molar size (large positive loadings for these variables, r 5 0.33–0.65; table 43) strongly influence the spread and separation of scores along the first axis, and highlights the greater magnitude of these dimensions in B. penitus as compared to B. fratrorum , measurement contrasts reflected by univariate means (table 42).

There are striking proportional differences between B. penitus and B. fratrorum in cranial and dental variables, which are revealed in a ratio diagram (fig. 56). The most conspicuous contrasts involve magni- tudes of certain cranial dimensions relative to size of skull (indexed by occipitonasal length): B. penitus is relatively narrower across the zygomatic arches but has a much broader interorbit, the facial region is relatively longer (longer rostrum, incisive foramina, and diastema), the zygomatic plate strikingly narrower, the bony palate longer, the mesopterygoid fossa wider, the bullar capsule larger, and the molars more massive. The elongate facial region is also proportioned differently in that B. penitus has a narrower rostrum, incisive foramina, and bony palate relative to their respective lengths.

Frequencies of certain cusps and cusplets forming parts of the coronal surfaces of molars differ between the two species. No specimen of B. fratrorum I examined shows a large labial cusplet adjacent to or merging with cusp t6 on the first upper molar (such a cusplet is present in 30 % of the sample of B.

penitus; fig. 75). Cusp t3 is absent from the third upper molar on most examples of both species, is part of the second molar in only 11 % of the sample of B. fratrorum , but is present in 62 % of the sample of B. penitus (table 10). A posterior labial cusplet is present on the first lower molar in nearly half of the sample of B. fratrorum but in every specimen of B. penitus , and an anterolabial cusp occurs infrequently on the second lower molar in the sample of B. fratrorum (12 %) but is present in about half of the sample of B. penitus (55 %); see table 11.

Summary of contrasts between population samples of Bunomys fratrorum and those of B. andrewsi and B. penitus : Bunomys fratrorum is found only on the northern peninsula east of the Gorontalo region and inhabits tropical lowland evergreen and montane rain forests; B. andrewsi occurs primarily in tropical

lowland evergreen rain forest in Sulawesi’s core and the southeastern and southwestern peninsulas; B. penitus is strictly montane, found so far only in the west-central mountain block and Pegunungan Mekongga on the southeastern peninsula. Body size is similar among the species.

Bunomys andrewsi has a shorter tail (both absolutely and relative to combined length of head and body), and a short, white tail tip is present in only a small portion of the sample; tails are longer and often coequal with length of head and body in B. fratrorum and B. penitus , and a long, white tip is usual in both species (tables 8, 41). Bunomys penitus has a longer dorsal coat and grayer underparts; dorsal surfaces of the front and hind feet are typically white rather than white- to brownish gray, the range in the other species; front claws are more robust and ungual tufts are sparse at the bases of front and hind claws in B. andrewsi as well as B. fratrorum compared with the smaller, gracile front claws of B. penitus and its longer ungual tufts.

Absolute and proportional similarities and contrasts in cranial and dental variables between B. fratrorum and the other two species have been presented in the images of skulls (figs 52–54), univariate summary statistics (table 42) results from principal-components analyses (fig. 55, table 43) and ratio diagrams (fig. 56) previously described. A graphic summary of the morphometric distinctions among the species is portrayed by individual specimen scores projected onto first and second canonical variates (fig. 57). Scores for the three species fall into three discrete clusters. Along the first axis, the group of points representing B. fratrorum is isolated from the other two aggregations by covariation in several cranial and dental variables (moderate to high loadings; table 44) that, compared with B. andrewsi and B. penitus , reflect its more greatly flared zygomatic arches, narrower interorbit, short- er rostrum (compared with B. penitus ), higher braincase, much wider zygomatic plate, longer basicranium (measured by postpalatal length) shorter bony palate and incisive foramina, smaller bulla, and smaller molars (compared with B. penitus ).

Along the first canonical axis, scores for B. penitus form a cluster opposite the clump representing B. fratrorum , which emphasizes the wider interorbit of B. penitus compared with B. fratrorum , its longer rostrum, wider but shallower braincase, much narrower zygomatic plate (which is diagnostic for B. penitus ), longer bony palate and incisive foramina, larger bulla, and heavier molars (see the loadings in table 44). The separation of scores denoting B. andrewsi from the aggregations for B. fratrorum and B. penitus along the second axis reflects proportional contrasts and is influenced primarily by the relatively smaller skull of B. andrewsi compared with the other two species, its relatively shorter and narrower rostrum, smaller braincase, wider zygomatic plate, shorter diastema, shorter but wider bony palate, narrower mesopterygoid fossae, and less robust molars (table 44).

The isolation of B. fratrorum from B. andrewsi and B. penitus by morphometric attributes is reinforced by a phenogram derived from squared Mahalanobis distances among centroids for population samples (fig. 21). Population samples for B. andrewsi and B. penitus join in a monophyletic group before linking with those representing B. fratrorum . This picture, along with the phenetic distinctions among the sets of samples evident from inspection of skins and skulls, absolute and proportional contrasts in dimensions, results of other multivariate analyses, and different geographic and altitudinal relationships support the hypothesis that B. fratrorum is a distinct entity with a different genetic history (species) from B. andrewsi and B. penitus . Neither the montane B. penitus nor the lowland and middle-elevation B. andrewsi can be viewed as simply geographic or altitudinal variants (subspecies) of the northern peninsular B. fratrorum . Its closest phenetic relative, judged by the pattern seen in the Mahalanobis distance cluster, may be the central Sulawesian B. karokophilus , n. sp. (see that account).

GEOGRAPHIC VARIATION: Variation in external traits among samples of B. fratrorum does not conform to any obvious geographic pattern. The range of variation of coat color and thickness, tail pattern, and external dimensions within any one of the three largest samples of B. fratrorum I studied (those from Teteamoet, Kuala Prang, and Temboan; see table 41) matches the extent of variation in these variables present among all the population samples except one. The three specimens collected above 1800 m in montane forest on Gunung Klabat have slightly longer and darker fur covering the upperparts of head and body.

Morphometric variation in cranial and dental variables does exist among population samples as results of principal-components and discriminant-function analyses illustratrate. A scatter plot defined by first and second principal components encloses specimen scores scattered along the two axes generally without forming perceptible patterns (fig. 58). Scores for the Teteamoet and Temboan samples fill the multivariate space, overlapping points for nearly all the other samples. The spread along the first component is influenced primarily by size, as signified by the moderate to high positive correlation coefficients (r 5 0.32–0.78) of most cranial variables (table 46) and likely mostly reflects age and individual variation within the range of adult cohorts from young to old. The filled square at the top of the scatter plot along the second axis denotes an adult from Gunung Mogogonipa with an exceptionally long bony palate and short incisive foramina (r 5 0.72 and ‾ 0.53, respectively; table 46).

A different view of covariation in cranial and dental variables among population samples is provided by individual specimen scores projected on first and second canonical variates in which scores for the larger samples form a pattern related to geographic provenance (fig. 59; table 46). Points identifying specimens from Teteamoet and Kuala Prang, located at the northeastern tip of the northern peninsula (localities 1 and 5 on the map in fig. 50), form a cluster in the right half of the ordination along the first axis that marginally overlaps the cloud of scores in the

TABLE 44 Results of Discriminant-Function Analysis Comparing All Population Samples of Bunomys fratrorum , B. andrewsi , and B. penitus Correlations (loadings) of 16 cranial and two dental log-transformed variables are based on 383 specimens; see figure 57.

left half of the scatter plot designating specimens from Temboan (locality 17 on the map in fig. 50), which lies far to the southwest of Kuala Prang and Teteamoet, and Maujat + Mogogonipa, highlands west of Temboan (localities 18 and 19 on the map in fig. 50). Scores for the holotype from Rurukan clusters with the Teteamoet sample, the two representing Gunung Klabat are contained in the Temboan cluster. Pattern of the scores reflects the slightly larger skulls (indexed by occipitonasal length and zygomatic breadth) in the Kuala Prang and Teteamoet samples, their deeper braincase, longer postpalatal region and incisive foramina but shorter bony palate compared with the Temboan and Maujat + Mogogonipa samples as signified by the larger positive and negative loadings for these variables on the first axis listed in table 46 (and reflected by univariate means in table 45). Intersample variation summarized in the canonical variates ordination suggests a gradient in skull size from east to west within the range of B. fratrorum .

Cluster analysis derived from average Mahalanobis distances (squared) among centroids for population samples also illustrate a close relationship between samples from Teteamoet and Kuala Prang, the localities close to one another, and these link to Temboan. Samples from Gunung Klabat and Rurukan cluster, and these places are geographically near one another (see the map in fig. 50). The westernmost sample from Gunung Maujat and Gunung Mogogonipa is distant from the other five population samples in the cluster diagram, an arrangement suggested by the position of scores in the canonical variate scatter plot where scores identifying the four specimens from Maujat and Mogogonipa lay at the far left margin of the total scatter of scores.

The population samples for B. fratrorum come from two regions of the northern peninsula that are deemed areas of endemism as illustrated by macaques ( Fooden, 1969), toads ( Evans et al., 2003a), and fanged frogs ( Evans et al., 2003a, 2003b). The northeastern end of the peninsula from the tip west to approximately the drainages of the Sungai Onggak Mongondaw and Sungai Onggak Dumoga describes one endemic region (‘‘northeast’’) and it is where the specimens of B. fratrorum from Teteamoet, Kuala Prang, Gunung Klabat, Rurukan, and Temboan were obtained. Those samples are closely linked (in Mahalanobis distance units) compared with the distant position of the samples from Gunung Maujat and Gunung Mogogonipa, which were collected to the west in the ‘‘north central’’ endemic region, approximately between Sungai Onggak Dumoga and the Gorontalo area (see fig. 1). But the Maujat and Mogogonipa samples are small, consisting of two adults, a young adult, and very young adult. Study of larger samples from the ‘‘north central’’ segment along with collections from the region between Gunung Maujat-Gunung Mogogonipa and the eastern site of Temboan will be required to determine whether the populations living between Sungai Onggak Dumoga and the Gorontalo area are a separate species, a morphometrically

TABLE 45

Descriptive Statistics for Cranial and Dental Measurements (mm) Derived from Population Samples of Bunomys fratrorum

Mean ± 1 SD and observed range (in parentheses) are listed.

definable geographic variant of B. fratrorum inferring some degree of genetic isolation, or simply another geographic sample of that species genetically connected to populations occurring to the east. In the context of such an inquiry the use of DNA sequences would be most welcome.

Most examples of B. fratrorum come from altitudes below 1000 m in tropical lowland evergreen rainforest habitats. A few were collected at higher elevations in lower montane forest on Gunung Maujat (at 1780 m) and Gunung Klabat (at 1829 and 1982 m). It is significant that study of skin and skulls along with results of multivariate analyses place these specimens with other population samples of B. fratrorum and not with any other species of Bunomys (see particularly the cluster diagram in fig. 21, showing phenetic relationships among all species of Bunomys based on squared Mahalanobis distances among centroids for population samples). Within its range, B. fratrorum is not replaced in montane forest formations by another species of Bunomys . This elevational distribution contrasts with the pattern demonstrated by other species in the B. fratrorum group occurring elsewhere on Sulawesi. In the northern part of the west-central region of Sulawesi’s core, for example, tropical lowland evergreen rainforest habitats are occupied by B. andrewsi and B. karokophilus , n. sp., and these two species are replaced by B. penitus at higher elevations in lower and upper montane forests. Bunomys andrewsi also inhabits lowland habitats on the southeastern peninsula and is replaced by B. penitus at higher elevations there in montane forest habitats. That B. fratrorum is the only large-bodied species of Bunomys found in both lowland and montane evergreen rain forests on the northern peninsula of Sulawesi may be related to past periodic fluctuations in sea level and tectonic activity that isolated the region east of Gorontalo as one or more islands ( Fooden, 1969). One of those islands may have supported a population of ancestral Bunomys from which B. fratrorum evolved in isolation from populations on mainland Sulawesi. Modern samples of B. chrysocomus represent the only other species of Bunomys currently recorded from the northeastern tip east of Gorontalo and may be a relatively recent immigrant to that region.

NATURAL HISTORY: I have no first-hand experience with B. fratrorum and no information from the collector’s field journals. The species occupies habitats in tropical lowland evergreen and montane rain forests and, as is usual with the other species of Bunomys , is likely terrestrial and nocturnal.

I can provide a bit of data about diet. Fortunately, H.C. Raven preserved an adult male in fluid (USNM 217084) that he caught at Teteamoet on the coastal plain (locality 1 in the gazetteer and on the map in fig. 50). The stomach contains remains of up to five small earthworms, fragments of a small adult beetle, head of a termite, and pieces of a small unidentified fruit. The composition of items is similar to that constituting the diet of B. chrysocomus and B. andrewsi (see those accounts and table 13).

Spermatozoa of B. fratrorum have been used in several inquiries. Body and testes weights (mass) along with sperm size was derived from two B. fratrorum by Breed and Taylor (2000) and employed in an investigation of murines designed ‘‘to test the hypothesis that differences in relative testes mass, and perhaps sperm size, relate to interspecific differences in the amount of intermale sperm competition and in breeding systems.’’ Related studies in which spermatozoa of B. fratrorum was included investigate morphometry, competition, cooperation, and sociality in rodent sperm ( Immler et al., 2007; Pizzari and Foster, 2008).

ECTOPARASITES: Bunomys fratrorum is parasitized by sucking lice, fleas, ticks, and chiggers (table 14). Of the sucking lice, Hoplopleura sembeli has also been recorded from the endemic Sulawesi rats Maxomys hellwaldii and Rattus hoffmanni ( Durden, 1990; Durden and Musser, 1991), and Polyplax wallacei also parasitizes Bunomys chrysocomus and a species of Taeromys ( Durden, 1990; Durden and Musser, 1991).

Of the species in the four genera of fleas ( Siphonaptera ) that parasitize B. fratrorum , Sigmactenus sulawesiensis (Leptopsyllidae) also parasitizes four other endemic Sulawesi murines ( Bunomys fratrorum , Eropeplus canus , Maxomys musschenbroekii , and Paruromys dominator ) and the endemic tree squirrel Prosciurillus topapuensis ( Durden and Beaucournu, 2000) . Sigmactenus alticola crassinavis also infests a native shrew ( Crocidura sp. ) and the endemic murines, Maxomys musschenbroekii , Paruromys dominator , Rattus hoffmanni , and R. xanthurus ( Durden and Beaucournu, 2000) . Musserella, n. gen. and species # 4 ( Pygiopsyllidae ), infests not only Bunomys fratrorum , but also five other Sulawesi endemic murids ( Bunomys chrysocomus , Rattus hoffmanni , Paruromys dominator , Maxomys hellwaldii , and M. musschenbroekii ) and the nonnative Rattus tanezumi (Durden, in litt., 2008). Nestivalius sulawesiensis (Pygiopsyllidae) is recorded from Bunomys fratrorum and B. chrysocomus ; Maxomys hellwaldii and M. musschenbroekii ; and Rattus facetus (recorded as R. marmosurus ) and R. hoffmanni ( Mardon and Durden, 2003) . Finally, Macrostylophora theresae also infests Paruromys dominator and Rattus xanthurus ( Durden and Beaucournu, 2006) .

Bunomys fratrorum is host to the immature stages of species in three genera of ticks ( Acari : Ixodoidea): Amblyomma sp. , Derma- centor sp., Haemaphysalis hystricis , and Haemaphysalis sp. ( Durden et al., 2008). In addition to Bunomys fratrorum , immatures of members of these three tick genera have been collected from a suite of other mammal hosts living in Sulawesi: shrews (the endemic Crocidura sp. and Crocidura elongata , and the commensal Suncus murinus ), pigs (Sus celebensis , endemic, and the domestic Sus scrofa), rusa ( Rusa timorensis , nonnative), water buffalo ( Bubalus bubalis , nonnative), domestic dog (nonnative), three endemic squirrels ( Rubrisciurus rubriventer , Hyosciurus heinrichi and H. ileile ), 12 species of endemic murid rodents ( Bunomys chrysocomus and B. andrewsi ; Margaretamys beccarii ; Echiothrix centrosa ; Maxomys hellwaldii , M. musschenbroekii , and M. wattsi ; Paruromys dominator ; Taeromys sp. ; Rattus hoffmanni , R. xanthurus , and R. facetus [recorded as R. marmosurus ]), and four nonnative murines ( Mus musculus ; Rattus tanezumi [recorded as R. rattus ], R. argentiventer , and R. exulans ); whereas adults have been recorded from pigs (Sus celebensis, Sus scrofa), rusa ( Rusa timorensis ), water buffalo ( Bubalus bubalis ), and domestic dog ( Durden et al., 2008).

Four species of chiggers ( Acari : Trombiculidae ) are known to parasitize Bunomys fratrorum ( Goff et al., 1986; Goff and Durden, 1987; Whitaker and Durden, 1987). Schoengastia sulawesiensis is also found on Maxomys musschenbroekii and Rattus hoffmanni . Walchiella oudemansi also parasitizes Bunomys chrysocomus and the nonnative Rattus exulans ). Leptotrombidium deliense has also been recorded from the endemic rats Bunomys chrysocomus , Maxomys musschenbroekii , Paruromys dominator , Rattus hoffmanni , and Rattus xanthurus . Finally, Gahrliepia lupella infests Maxomys musschenbroekii in addition to Bunomys fratrorum .

TABLE 46 Results of Principal-Components and Discriminant- Function Analyses Performed on All Population Samples of Bunomys fratrorum Correlations (loadings) of 16 cranial and two dental log-transformed variables are based on 100 B. fratrorum ; see figures 58 and 59.

SYNONYMS: No other scientific names have been proposed that apply to B. fratrorum .

Bunomys andrewsi , the next species to be discussed, is not all that different from B. fratrorum in fur color and physical size, but judged by dimensions of the skull and molars is phenetically more like the montane B. penitus than B. fratrorum (see fig. 21), and has been recorded only from the southern half of Sulawesi where it occurs primarily in tropical lowland evergreen rain forest.