Morphometric analyses

There was no significant sexual dimorphism in external measurements of Batomys from either the Central Cordillera or Mt. Isarog, although for the sample from Mt. Isarog, males averaged slightly larger than females and the difference in length of head and body approached significance (Table 1). Few cranial measurements involved samples large enough to test for sexual dimorphism, and those only among specimens of B. granti from the Central Cordillera; the only significant difference obtained was breadth of the zygomatic plate, which was greater for males (Table 2).

We conducted a principal components analysis of 18 cranial measurements using combined samples of adult male and female individuals of Batomys from Mt. Amuyao (n = 4), Mt. Data (n = 2), Mt. Isarog (n = 3), and Kalinga (n = 2). The first four components accounted for 75% of the total variance. Components 1 and 2 with eigenvalues of 4.9 and 3.9, respectively, together accounted for 49.1% of the total variation (Table 3). Components 3 and 4 together accounted for 25% of the total variation; both had eigenvalues of less than 3 and with few variables having large magnitude loadings. In a bivariate plot of individual specimen scores on components 1 and 2 (Fig. 2), specimens from Mt. Amuyao, Mt. Data, and Mt. Bali-it are clearly separated from those from Mt. Isarog; the latter have longer maxillary toothrows, broader incisors and palates, lower braincases, and shorter nasals, incisive foramina, and diastemas compared to specimens from the Central Cordillera (see the higher positive and negative loadings for these traits in Table 3).

Table 3.

Character loadings, eigenvalues, and percent variance explained on the first two components of a principal components analysis of log₁₀-transformed cranial and dental measurements of adult Batomys granti and B. uragon (Fig. 2).

Fig. 2.

Results of principal components analysis of cranial and dental measurements of Batomys (see Table 3). Projections of specimen scores on components 1 and 2 for B. granti from Mt. Amuyao (closed circles), Mt. Data (open circles), and Mt. Bali-it (open squares), and for B. uragon from Mt. Isarog (closed triangles).

Molecular analysis

The tree derived from the maximum likelihood analysis of the cytochrome b data set (Fig. 3) strongly supports (bootstrap = 81%) monophyly of Batomys granti specimens from four localities in the Central Cordillera (including Mt. Data, the type locality), monophyly of the specimens from Mt. Isarog (bootstrap = 100%), and the sister-group relationship of these two montane samples (bootstrap = 100%). The Central Cordillera and Mt. Isarog groups are separated by 4.2% net uncorrected p-distance. As found in previous genetic analyses, the genus Batomys, as currently defined, is paraphyletic because Crateromys heaneyi, as generically allocated by Gonzales & Kennedy (1996), is nested within it; however, support for this arrangement is not strong (bootstrap = 67%). As reported previously, the four species of Musseromys (Heaney et al. 2014a) form a well-supported clade that is sister to Carpomys phaeurus, and Phloeomys cumingi is recovered as the basal taxon to all other members of the cloud-rat clade.

Fig. 3.

The maximum-likelihood (lnL = −8305.40) tree resulting from analysis of cytochrome b sequences from species of the Phloeomys Division under the GTRGAMMA model in RAxML. Solid circles at nodes represent bootstrap values ≥ 90%; numbers at nodes indicate maximum-likelihood bootstrap values < 90%. Terminal taxa are identified by species names and a unique alphanumeric identifier (either museum catalog number or GenBank accession number; Table 4).

Table 4.

List of specimens used in genetic analysis.

Based on these DNA and morphometric results, we restrict Batomys granti to include only those populations in the Central Cordillera, and recognize Batomys from Mt. Isarog as a new species.

Batomys uragon, new species

Figs. 2–6, Tables 1–3

Fig. 4.

Adult female Batomys uragon (USNM 458943, paratype) from Mt. Isarog, 1550 m elevation, Camarines Sur Province, Luzon Island. Photographed by E. A. Rickart, 18 March 1988.

Fig. 5.

Dorsal, ventral, and lateral views of cranium and lateral view of mandible of adult Batomys from Luzon Island. A, B. granti (FMNH 193697); B, B. uragon (USNM 458949, holotype).

Fig. 6.

Bivariate plots of measurements comparing Batomys granti and B. uragon. A, tail length vs. ear length; B, nasal length vs. length of incisive foramina; and C, interorbital breadth vs. breadth of incisors at tip.

Holotype

USNM 458949, old adult male collected on 25 Apr 1988 by Eric A. Rickart (field number: EAR 2004); preserved in ethanol, with skull subsequently removed. The skull (Fig. 5B) is in good condition, except for the broken nasal tips, anterior edge of right premaxilla, and left pterygoid process. A small sample of muscle tissue was removed from the thigh for molecular studies and a section of the femur was taken for karyology (Rickart et al. 1991, Rickart & Musser 1993).

Type locality

Mt. Isarog; 4 km N, 22 km E Naga City, Camarines Sur Province, Luzon Island, Philippines; in primary mossy forest at 1750 m elevation; 13°40′N, 123°22′E (Fig. 1; Site 6 in Heaney et al. 1999).

Paratypes (n = 14)

Aside from the holotype, 14 additional specimens from Mt. Isarog were examined, consisting of five from the type locality (USNM 458914, 458945–458948); six from 4 km N, 21.5 km E Naga City, elev. 1550 m, 13°40′N, 123°22′E (USNM 458939–458944); and one each from the following localities: 4 km N, 21 km E Naga City, elev. 1350 m, 13°40′N, 123°22′E (USNM 458950); 4 km N, 21 km E Naga City, elev. 1650 m, 13°40′N, 123°22′E (FMNH 152033); and 8.9 km N, 0.8 km E Ocampo Municipality, Camarines Sur Province, elev. 1700–1800 m, 13°38′32″N, 123°23′30″E (FMNH 142046).

Distribution

Known only from Mt. Isarog, where it occurs in old-growth montane and mossy forest, elevation 1350–1800 m. The habitat is described in more detail in Heaney et al. (1999, Sites 4–6 and 16–18).

Etymology

From the Bicolano word uragon (no direct English word counterpart, but roughly translatable to “possessing great ability, vitality, or power”), to highlight its dispersal ability, persistence, and uniqueness to the Bicol Peninsula; adjective in the nominative singular neuter.

Diagnosis

A species of Batomys, distinguished from B. granti, its nearest phylogenetic relative based on cytochrome b sequences (Fig. 3), by the following combination of features: medium body size (combined lengths of head and body 181–206 mm), with absolutely and comparatively shorter tail; moderate body weight, 160–220 g; pale golden-brown pelage dorsally, paler ventrally; skull longer (40.64–41.71 mm) with shallower brain case (11.66–11.98 mm) and narrower interorbital region (5.25–5.91 mm); rostrum deeper (9.52–10.16 mm) but shorter (17.42–18.79 mm) with correspondingly shorter nasals (15.52–16.76 mm) and incisive foramina (7.64–8.64 mm); zygomatic plate narrower (3.65–4.18 mm); palate broader at M1 (7.43–8.17 mm) and at M3 (4.21–5.04 mm); molar toothrow longer (7.80–8.76 mm) and upper incisors broader (2.77–3.15 mm).

Karyology

2n = 52, FN = 52, entirely telocentric chromosomes. The standard karyotype is indistinguishable from those of Batomys granti from the Central Cordillera and B. salomonseni from Leyte and Mindanao islands (Rickart & Musser 1993, Rickart & Heaney 2002).

Description and comparisons

Among the Batomys from Luzon Island, B. dentatus, known only from the holotype, is the most distinctive. From both B. granti and B. uragon it differs in possessing an absolutely and relatively longer tail, which also is distinctive in having the distal third covered in white hairs that form a terminal tuft. Cranially, B. dentatus is distinguished from the other two species in having a wider zygomatic plate, shorter incisive foramina relative to the bony palate, and much more robust dentition (see Table 4 and Fig. 20 in Musser et al. 1998). The remainder of this section provides a detailed description of B. uragon in comparison to B. granti.

Batomys uragon is similar to B. granti in external appearance and length of head and body (Figs. 4, 6, Table 1). The dorsal pelage of B. uragon is medium to pale golden-brown (Fig. 4); orange-brown to dark brown among congeners. Individual color variation in B. uragon was noted in the field; most individuals had golden dorsal fur, but the holotype was paler and was noted in the field by Rickart to have “no red coloration (colored like Rattus everetti).” The predominance of the pale golden pelage is particularly evident in FMNH 142046 and 152033. The pelage is woolly and shorter than in B. granti; B. dentatus has the longest fur overall (Musser et al. 1998). The venter is grayish buff with patches of unpigmented fur in the chin, pectoral, and inguinal regions. One male specimen (FMNH 142046) has a small pale brown patch of fur on the left side of the lower abdomen. The ears are slightly shorter than in B. granti (Table 1), and are pale brown covered with short grayish-brown hairs. The eyes are surrounded by pale rings of skin covered with very short grayish-brown and unpigmented hairs, giving the impression of naked eye rings (Fig. 4). The vibrissae are comparable to those of B. granti (mystacial up to 65 mm, superciliary 17–36 mm, extending to or slightly beyond the dorsal margin of the ears, and postocular vibrissae 24–26 mm). The forefeet are broad, each with a mid-dorsal stripe of short, dark brown fur, bordered with short, pale golden-brown fur on each side (Fig. 4). Claws are opaque, with ungual tufts of unpigmented hairs extending slightly beyond the outer margin of the claws, particularly on digits III and IV. The palmar surface, including pads, is unpigmented. The hind feet are broad and about as long as in B. granti, with short, dark brown fur on the dorsal surface, and pale golden-brown fur laterally (Fig. 4). Ungual tufts, composed of unpigmented hairs, extend to the tips of the claws and up to 2 mm beyond them (in FMNH 152033). Plantar surface and the six large, fleshy plantar pads are unpigmented (Fig. 5 in Musser et al. 1998). Compared to B. granti, the tail of B. uragon is shorter on average (Fig. 6, Table 1), absolutely and relatively (128–149 mm, or 63–75% of the combined lengths of head and body), except for an adult female (USNM 458939) that has a longer tail (171 mm or 92% of combined lengths of head and body). The tails of B. uragon and B. granti are uniformly colored, in contrast to the bi-colored tail of B. dentatus (basal two-thirds brown, distal third white; Musser et al. 1998). Batomys uragon has three hairs arising in association with each caudal scale, but has a less visibly pilose tail than does B. granti due to its shorter hairs: 4–9 mm around mid-tail, 7–12 mm at the tip (7–11 mm at mid-tail, 8–30 mm at the tip in B. granti).

In Musser et al.'s (1998) redefinition of B. granti, the specimens from Mt. Data that were then available had significantly shorter tails than those of specimens from Mt. Isarog (Table 1 in Musser et al. 1998). The tails of specimens in that Mt. Data series have blunt tips and are covered with hairs, suggesting to the authors that their shorter length was natural, rather than caused by damage that occurred during life or during preparation of the specimens. Thomas (1895, 1898), however, suspected that the tail of the holotype of B. granti was incomplete. Because of the small number of specimens available from Mt. Data, Musser et al. (1998) refrained from drawing any taxonomic conclusions based on differences in tail length. Three recently obtained specimens of B. granti collected on Mt. Data (FMNH 188321–188323) have intact tails that are as long as, or longer than, those of all specimens of B. uragon, except USNM 458939 as mentioned above (Table 1). Furthermore, examination of specimens of B. granti from Kalinga Province and Mt. Amuyao with blunt tail-tips, which at first appeared natural, revealed that the tips are missing; in several, one can feel the sharp edges of the broken vertebra that are barely covered with healing skin. Among adult specimens of B. granti (n = 11), only one specimen from Mt. Amuyao (FMNH 193691) had a distinctly bobbed tail (115 mm or 62% of the combined lengths of head and body). This specimen has a tail that is slightly longer than those of specimens previously available from Mt. Data (75–110 mm; FMNH 62504, BMNH 95.8.2.15 [holotype], BMNH 97.5.2.31; Thomas 1895, 1898; Musser et al. 1998). In summary, we regard the smaller tail length measurements recorded for older specimens of B. granti as artifactual; in fact, specimens from the Central Cordillera, including Mt. Data, have significantly longer tails than those of specimens from Mt. Isarog, herein described as B. uragon (Table 1).

Cranial features are similar among congeners, as noted earlier by Musser et al. (1998), but distinct and consistent differences in size and proportion are evident among the specimens of B. granti and B. uragon (Figs. 5, 6, Table 2). Batomys uragon has shorter nasals and a narrower interorbital region. The shorter rostrum (associated with short nasals) is nonetheless more robust in terms of rostral depth. Dorsally, the inflated cranium contrasts sharply with its narrow interorbital region, and because the anterior frontal region is narrowly flared relative to its posterior section, the interorbital region has a longer and more slender appearance than in B. granti. In contrast to the narrow rostrum of B. granti, that of B. uragon has a pronounced lateral inflation that is formed by the swollen capsular projection on each side of the upper incisor roots (Fig. 5).

In lateral profile, the dorsal surface of the skull of B. uragon traces an evenly convex arc, extending from the tips of the curved nasals, across the arched frontals of the interorbital region, to the gently curved parietals. In contrast, the dorsal surface of the skull in B. granti is flatter along the nasals and frontals, and the parietal region is more abruptly convex (Fig. 5). The zygomatic arch of B. uragon is similar to B. granti in breadth, but it is distinctive in having the central portion, under the jugal, markedly deeper, brought about by broad postero-ventral expansion of the maxillary zygomatic process (Fig. 5). The zygomatic plate is narrower than in B. granti. Ventrally, skulls of the two species are generally similar, but B. uragon typically has shorter incisive foramina, a broader palate, and smaller auditory bullae that have a moderate to pronounced postero-ventral inflation imparting a rounder appearance (Figs. 5, 6, Table 2).

The dental features of B. granti and B. uragon are similar, differing mainly in the relative size of incisors and molars, length of the molar toothrow, and orientation of the crown and anterior rooting of the first upper molars (Figs. 5, 6, Table 2). In particular, B. uragon has a longer molar toothrow, broader upper incisor tips, and shorter diastema. Its dentition is also distinct in having a pronounced inclination of the first row of cusps of M1, further exaggerated by its similarly inclined anterior molar root. The steep inclination of the anterior root of M1 is present among most specimens of B. uragon we examined, except one female (USNM 458939) and one young adult male (USNM 458947), both of which had a more vertical orientation; the latter individual had the posterior edge of the incisive foramina farthest from the anterior root of M1. Compared with B. granti (Fig. 5), the posterior portion of the dentary of B. uragon is wider due largely to its broadly rounded angular process and more expansive base. The condyloid process is as broad as in B. granti but appears shorter due to the shallow sigmoid notch that separates it from the coronoid process. In both species, the coronoid process itself is short and slightly backswept, projecting dorsally to about the same level as the condyloid process.

Ecology

Specimens of Batomys uragon were trapped in montane and mossy forest between 1350 and 1800 m elevation, in habitats with thick leaf litter and humus layers. During the 1988 survey of Mt. Isarog, 11 out of 13 animals (85%) were captured in traps baited with roasted coconut and peanut butter, and two in traps baited with live earthworms. Stomach contents from 3 individuals contained finely chewed vegetable matter only. Two live-trapped animals were presented potential food items. One readily accepted sprouting seeds of an unknown species of dicot, gnawing off and rejecting the sprouts and consuming only the seeds. This animal rejected other plant items that were offered, including a variety of small fruits, bulbs of orchids and other plants, several types of leaves, dried beans, peanuts, and coconut-peanut butter bait. A second animal likewise readily accepted the seeds of the sprouting dicot (again, discarding the sprouts and consuming only the seeds), and also accepted a single type of small fruit as well as coconut-peanut butter bait. Both individuals rejected all animal items including earthworms and various insects. These data suggest that B. uragon has a granivorous-frugivorous diet. All specimens were trapped at night on the ground, along runways and among root tangles. We conducted limited arboreal trapping on Mt. Isarog, but some individuals were captured in places where the terrain was very steep, such as on nearly vertical sides of deep ravines, which suggests that the species is a good climber (Rickart et al. 1991, Heaney et al. 1999). During a mark and release trapping study in 1993–1994, one female was caught five times in a three-day period and moved a mean distance of 11 m (range 5–14 m) between captures; another female moved an average of 44 m (range 14–85 m) between eight captures over a period of three months (Balete & Heaney 1997). There were no signs of reproductive activity among individuals captured in March–May 1988 and December 1993–May 1994 (Rickart et al. 1991, Balete & Heaney 1997, Heaney et al. 1999). Two adult females, both taken in April 1988, had large mammae, perforate vaginas, and old placental scars; there were four scars in USNM 458945 (220 g) and three in USNM 458948 (175 g), but we could not determine the number of litters involved. Further details on reproduction of this species are discussed in Heaney et al. (1999).

Phylogenetic relationships

Our phylogenetic analysis using cytochrome b produced results (Fig. 3) that are consistent with those of other studies in placing Phloeomys in a basal position as the sister taxon to all other cloud rats, Carpomys and Musseromys as sister genera, species of Batomys from Luzon as sister-group to Crateromys heaneyi, and species of Batomys from Greater Mindanao as sister-group to that clade. Clearly, the phylogenetic relationships within the cloud-rat clade require further study to determine whether species not yet represented in molecular analyses conform to the monophyletic groups currently apparent. Future work should involve additional taxa and sequence data from nuclear genes designed to clarify the relationship of Batomys and Crateromys from Luzon, the taxonomic status of those species of “Batomys” from Mindanao as a separate phylogenetic radiation, and the relationship of the enigmatic Dinagat Island endemics B. russatus and Crateromys australis to the other cloud rats.

Our phylogenetic analysis (Fig. 3) also provides support for reciprocal monophyly of the Mt. Isarog populations with respect to those from the Central Cordillera. Furthermore, the net uncorrected p-distance of 4.2% between these groups is comparable to that seen between recognized sister species of rodents (Baker & Bradley 2006). This result, coupled with the morphological distinctiveness of the two groups (Fig. 2), provides solid evidence for recognizing the Mt. Isarog population as a species distinct from B. granti. This recognition raises the number of native rodent species documented on Luzon to 44, and the number of known members of the endemic Philippine cloud-rat clade to 18 species, 12 of which occur on Luzon (Heaney et al. 2014b).

Biogeography

The extant species of Batomys on Luzon are narrowly distributed, with two species in the Central Cordillera and one on Mt. Isarog (Fig. 1). Mt. Isarog (1966 m elevation) remains the only location outside of the Central Cordillera where the genus is currently known to exist, in spite of extensive surveys of small mammals on isolated mountains and mountain ranges elsewhere on Luzon with elevations of at least 1500 m and seemingly suitable habitat for Batomys (Balete et al. 2009, 2011, 2013a, 2013b; Alviola et al. 2011, Duya et al. 2011, Rickart et al. 2011a, 2011b, 2013; Heaney et al. 2013a, 2013b). The recent discovery of late Pleistocene fossil remains of Batomys in a lowland area in the Cagayan Valley between the central Cordillera and Sierra Madre mountain ranges reveals a much wider geographic and elevational distribution in the past, and perhaps a wider habitat association as well (Heaney et al. 2011b). The absence of Batomys from other high mountains on Luzon seems enigmatic, yet Batomys uragon now joins Archboldomys luzonensis, Chrotomys gonzalesi, and Rhynchomys isarogensis as species that are endemic either entirely to Mt. Isarog or to Mt. Isarog and nearby mountains on the Bicol Peninsula (Musser & Freeman 1981, Rickart & Heaney 1991, Balete et al. 2012, 2013b). We note that the sister species (and only congener) of Archboldomys luzonensis (A. maximus) also occurs in the Central Cordillera, whereas the geographically closest congener of R. isarogensis (R. banahao) is endemic to Mt. Banahaw, which is geographically intermediate between the Central Cordillera and Mt. Isarog (Fig. 1). In contrast to the cognate relationships derived for these montane species, Chrotomys gonzalesi is most closely related to C. mindorensis, a species that is widespread in the lowlands of central Luzon (Figs. 7, 8 in Balete et al. 2012). Perhaps when additional fossils are found, the current montane distributions of Batomys, Archboldomys, and Rhynchomys will be shown to be relictual patterns derived from more widespread distributions during the Pleistocene. In the meantime, recognition of Batomys uragon further reinforces the importance of speciation within Luzon as a primary process generating the island's existing mammalian diversity (e.g., Heaney & Rickart 1990, Heaney et al. 2011a, 2014a, 2014b; Balete et al. 2012, Justiniano et al. 2014).

Conservation

Batomys uragon is restricted to primary montane and mossy forest on Mt. Isarog, and appeared to be moderately common during our studies (Rickart et al. 1991, Balete & Heaney 1997, Heaney et al. 1999). These high elevation habitats are within Mount Isarog National Park and populations of B. uragon under the present management regime appear to be stable and under no threat, though limited in area. In contrast, Batomys dentatus remains to be rediscovered since it was first described by Miller in 1910, and thus is of uncertain status. Batomys granti is broadly distributed in the Central Cordillera and was present but relatively uncommon at our study areas on Mts. Amuyao, Bali-it, and Data (Rickart et al. 2011a, 2011b). However, the proliferation of commercial vegetable farming on Mt. Data, the type locality for this species, has reduced the forest cover to a very small (ca. 80 ha) remnant patch of disturbed mossy forest (Heaney et al. 2006a). Furthermore, this species was not encountered during our survey of Mt. Pulag National Park where commercial agriculture also has reduced forest. Although Batomys granti is currently widespread and has some tolerance for habitat disturbance (Rickart et al. 2011a), the extensive destruction of mossy forest associated with vegetable farming in some the Cordillera poses a risk to the habitat of the species. The continued survival of Batomys and other endemic Philippine rodents will require active protection and careful management of their forest habitats, many of which form crucial watershed areas.