Coccymys Menzies, 1990

Voss, Musser Scientific Publications Of The American Museum Of Natural History American Museum Novitates Bulletin Of The American Museum Of Natural History Anthropological Papers Of The American Museum Of Natural History Publications Committee Robert S., Meng, Chair Board Of Editors Jin, Prendini, Paleontology Lorenzo, Voss, Invertebrate Zoology Robert S., Whiteley, Vertebrate Zoology Peter M., Knight, Anthropology Managing Editor Mary, Lunde, And, New Guinea Coccymys, Melomys & Bulletin, Albidens Amnh, 2009, Systematic Reviews Of New Guinea Coccymys And ‘‘ Melomys’ ’ Albidens (Muridae, Murinae) With Descriptions Of New Taxa, Bulletin of the American Museum of Natural History 2009 (329), pp. 1-139 : 1-139

Coccymys Menzies, 1990

TYPE SPECIES: Pogonomelomys ruemmleri

Tate and Archbold, 1941: 6.

EMENDED DIAGNOSIS: A genus in the Pogonomys Division ( Musser and Carleton, 2005) of subfamily Murinae , family Muridae (as delimited by Carleton and Musser, 1984, and Musser and Carleton, 2005) that is distinguished from all other described murine genera by the following combination of traits: (1) dorsal pelage covering head and body thick and soft, very dark brownish with blackish infusion through brownish gray to bright tawny-russet, ventral coat soft and thick, whitish gray to dark grayish white, some individuals washed with buff; (2) short muzzle, gray face, and prominent blackish region around each eye and extending along each side of the muzzle; (3) tail slender and much longer than head and body (LT/LHB ranges from 140% to 181%), scales small and slightly swollen, their annuli abutting each other (not overlapping), three short hairs associated with each scale, dorsal surface near tip with thickened epidermis and lacking scales and hairs, indicating some degree of prehensility, entire tail brownish gray, some individuals with a white tip; (4) dorsal surfaces of front and hind feet whitish or pale tan, hallux with claw, hind foot elongate with full complement of plantar tubercles; (5) three pairs of teats, one postaxillary and two inguinal; (6) rostrum narrow and moderately long, interorbital and postorbital margins smooth, interorbit hourglass-shaped in dorsal view, zygomatic arches flare from sides of skull, braincase smooth and globular, interparietal wide, occiput deep, evident cranial flexion; (7) zygomatic plate narrow, its anterior margin straight and barely projecting beyond dorsal maxillary root of zygomatic arch, its posterior edge set just in front of the first molar, superficial masseter inserts on a rugose bump at base of ventral zygomatic root; (8) vertical ridge marking frontal-squamosal suture at back of postorbital region, section of parietal projects ventrad beyond temporal beading to form part of lateral braincase wall, squamosal intact except for large subsquamosal foramen; (9) wide, bony alisphenoid struts; (10) moderately wide and long incisive foramina, their posterior margins ending slightly anterior to front faces of first molars, even with them or, projecting slightly between; (11) molar rows diverge slightly, bony palate short with its posterior margin even with back faces of third molars or extending slightly beyond them, palatal surface with deep palatine grooves, posterior palatine foramina level with middle of second molar; (12) elongate and spacious sphenopalatine vacuities; (13) wide pterygoid plates with moderately deep pterygoid fossa, small to moderate sphenopterygoid openings; (14) small ectotympanic (auditory) bulla relative to skull size, capsule incompletely covering periotic, posterodorsal wall of carotid canal formed by periotic and not bullar capsule; (15) large stapedial foramen, no sphenofrontal foramen or squamosal-alisphenoid groove, indicating a carotid arterial pattern widespread within Murinae (character state 2 of Carleton, 1980; pattern 2 described by Voss, 1988); (16) dentary stocky, high ramus between incisor and molar row, high ascending ramus, large coronoid and condyloid processes, end of alveolar capsule forming prominent labial projection at level of sigmoid notch between coronoid and condyloid processes; (17) upper incisor enamel orange, lowers slightly paler; (18) each upper molar (maxillary) with three roots, each lower (mandibular) with two; (19) molars brachydont, cusp rows forming moderately complicated cuspidate occlusal patterns, third molar large relative to others in toothrow; (20) first and second upper molars of most specimens with some form of an enamel ridge projecting from anterolingual surface of cusp t8 anteriorly to posterior margin of lingual cusp t4, a labial enamel ridge connects anterolabial margin of cusp t9 with posterolabial margin of cusp t6, a comparable but shorter ridge projects from the anterior surface of cusp t5 to meet the posterior margin of cusp t3 near the cingulum, and with conspicuous posterior cingulum on first molars, also present on second molar in a few specimens; (21) no cusp t7; (22) anteroconid formed of chunky anterolingual and anterolabial cusps, anterocentral cusp absent, anterolabial remnant on second lower molar, no anterior labial cusplets on first and second lower molars, but posterior labial cusplet present on each tooth, posterior cingulum elongate; (23) phallic morphology highly divergent from other endemic New Guinea murines surveyed ( Lidicker, 1968); (24) sperm head falciform in shape, with single and long apical hook and lacking ventral processes, spermatozoal tail moderately long ( Breed and Aplin, 1994).

THE SPECIES OF COCCYMYS

Most lists, faunal studies, and taxonomic revisions published since Tate’s 1951 treatise recognized only ruemmleri , whether as a species of Pogonomelomys ( Tate, 1951; Laurie and Hill, 1954; Menzies and Dennis, 1979; Willett et al., 1989; Flannery, 1990) or Coccymys ( Menzies, 1990; Musser and Carleton, 1993; Flannery, 1995). This view of a single species was championed in Menzies’s (1990: 133) revision in which C. ruemmleri was recorded as occurring ‘‘throughout the central mountain ranges of New Guinea from Lake Habbema in the west to the Wau-Bulolo area in the east, possibly further east, usually over 1500 asl and up to 2500 or higher.’’ By 2005, Musser and Carleton (2005) had accepted the Central Cordilleran distribution of ruemmleri previously outlined by Menzies but also announced the presence of a second undescribed species in the Maneau Range of New Guinea’s eastern peninsula.

The present report on Coccymys was sparked by our intention to describe that eastern peninsular species, and was not meant to comprise a revision of the genus. But this simple goal could not be achieved without comparing the sample from the Maneau Range with material collected from the Central Cordillera in western and central New Guinea, an undertaking that has result- ed in a revision of species diversity in Coccymys . Through those comparisons, we could demonstrate the distinctiveness of the new form and in the process reject the notion that a single species occurred from the Lake Habbema region in the west to the Wau-Bulolo area in the east. The reality is different—two species are found throughout this Cordilleran backbone, generally replacing one another in a linear pattern but with regional sympatry in western Papua New Guinea. True C. ruemmleri is documented by specimens from the Snow Mountains and Star Mountains in the western half of New Guinea. Coccymys shawmayeri replaces C. ruemmleri in the montane forests and alpine grasslands of the Central Cordillera from the Telefomin area in the west eastward beyond the Wau-Bulolo region to Mt. St. Mary in the east, the western region of the Owen Stanley Ranges. From the Wau-Bulolo area eastward to the Maneau Range is the territory of the new species, C. kirrhos , which is sympatric with C. shawmayeri in the Wau-Bulolo area and farther east in the western portion of the Owen Stanley Ranges.

We arrived at this view of three species occurring along the mountainous backbone of New Guinea by first analyzing phenetic variation in external, cranial, and dental traits among AMNH samples of Coccymys only. The results were later tested against data derived from AM, BBM, and BMNH samples. We organized the AMNH specimens into six population samples, three from the Snow Mountains in western New Guinea, one of which contains the holotype and type series of ruemmleri ; two from the eastern Cordillera, Mt. Hagen, and Mt. Wilhelm; and the three examples from Mt. Dayman in the Maneau Range at the far end of the Owen Stanley Ranges (table 1). Variation among these samples in external traits (lengths of head and body, tail, and hind feet) is summarized by univariate means (tables 4–6). Multivariate approaches in the form of discriminant function analyses allowed us to assess the significance of morphometric variation in cranial and dental variables among the six AMNH samples.

Morphometric differences among the six population samples are summarized in the patterns of covariation in cranial and dental variables reflected in graphs of specimen scores projected onto the first and second canonical variates extracted from discriminant-function analysis (fig. 4). Three patterns in multivariate space are revealed. First, two slightly overlapping clusters of scores are aligned along the first axis. The left cluster identifies specimens from the Snow Mountains (Mt. Wilhelmina, Lake Habbema, and Bele River valley), the right encompasses specimen scores from eastern New Guinea samples (Mt. Hagen and Mt. Wilhelm). Covariation in breadth of mesopterygoid fossa; height of braincase; and lengths of bony palate, rostrum, and maxillary molar row are principally responsible for this scatter of scores along the first axis, partic-

TABLE 1 Population Samples of the Species of Coccymys and Brassomys Employed in Univariate and Multivariate Analyses of Cranial and Dental Variables (Number in parentheses after each locality keys to numbered collection locality in gazetteer and distribution maps in fig. 2 for Coccymys and fig. 3 for Brassomys . Number in brackets indicates total number of specimens for each species. Specimens measured are identified by museum initials and catalog numbers in footnotes.)

ularly rostral length—breadth of first upper molar also contributes but with less force (see table 2, and diagram of factor loadings in fig. 4). Compared with the three series of C. ruemmleri from the Snow Mountains, samples from Mt. Hagen and Mt. Wilhelm have a relatively wider mesopterygoid fossa; shallower braincase (less bulbous); noticeably shorter bony palate, rostrum, and maxillary toothrow; and somewhat narrower first upper molar.

These proportional distinctions are also mirrored in the univariate summaries listed in tables 4–6, and are summarized in a phenetic clustering pattern of population samples based on Mahalanobis distance squared (fig. 5). This dichotomy, based on cranial and dental variables, is accompanied by contrasts in length and coloration of fur, relative length of tail, and frequency of white tail tips in the samples. Specimens in the three samples from the Snow Mountains are typically characterized by a thicker dorsal coat along with darker upperparts, underparts, and tail than specimens from Mt. Wilhelm and Mt. Otto (we could not use most of the specimens from Mt. Hagen to assess coloration of pelage or most dimensions of appendages because in the field they were all dumped into fluid preservative and not measured beforehand; the bodies are now contorted and stiff, the fur and surfaces of appendages discolored). Relative to length of head and body, the tail is short in the samples from the Snow Mountains (LT/LHB 5 140%–146%), the frequency of a white tail tip is very low (15% of 41 specimens), and when present the white portion is short (5– 30 mm). By contrast, samples from Mt. Wilhelm and Mt. Otto in Papua New Guinea tend to have a relatively much longer tail (LT/LHB 5 162%–170%), nearly all the tails have white tips (77% of 31 specimens), and the white segment averages longer (10– 42 mm) ; see table 8.

These qualitative and quantitative contrasts reflect distinctions between two species: the AMNH samples from the Snow Mountains represent Coccymys ruemmleri ; attributes of the holotype of Hinton’s (1943) shawmayeri fit with those exhibited by samples from Mt. Hagen, Mt. Wilhelm, and Mt. Otto, and that name applies to those samples.

The second pattern in the canonical ordination in figure 4 is revealed by specimen scores aligned along the second canonical variate. There the scores representing the three AMNH specimens from Mt. Dayman

TABLE 2 Results of Discriminant Function Analysis Derived from 76 Adults in Six Population Samples of Coccymys ruemmleri , Coccymys shawmayeri , and Coccymys kirrhos (Correlations, eigenvalues, and proportions of variance are explained for two canonical roots; see fig. 22. Localities constituting each sample, and sample sizes, are listed in table 1. Mean values of measurements in each sample are listed in tables 4–6.)

are isolated from the two constellations corresponding to the samples of C. ruemmleri (Snow Mountains) and C. shawmayeri (Mt. Hagen and Mt. Wilhelm) . The distribution of scores along this axis is influenced most strongly by breadths of the incisive foramina, mesopterygoid fossa, interorbit, and zygomatic plate: compared with C. ruemmleri and C. shawmayeri , the incisive foramina and mesopterygoid fossae are relatively broader in the Mt. Dayman sample, and its interorbit and zygomatic plates are relatively narrower (proportions also reflected in ratio diagrams, figs. 30, 31). Other variables responsible for the spread of scores along the second axis, but with less influence, are breadths of the bony palate and braincase, zygomatic breadth, and lengths of incisive foramina, diastema, bullae, and postpalatal region. The Mt. Dayman population has a relatively narrower bony palate, braincase, and zygomatic expanse, but shorter incisive foramina, diastema, bullae, and postpalatal region (see table 2 and the factor loadings in fig. 4). Phenetic clustering of population samples based on Mahalanobis distances squared also reflects the broad morphometric gulf betwen the sample from Mt. Dayman and those representing C. ruemmleri and C. shawmayeri (fig. 5).

The AMNH sample from Mt. Dayman forms the type series of the new species, C. kirrhos , to be described in the accounts of species following this section. Coupled with the multivariate contrasts outlined here, the new species is typically smaller in some external dimensions and in most cranial and dental measurements than the other two species, and exhibits different fur thickness and coloration, and tail proportion and color pattern.

A final pattern of covariation among variables is evident in the clustering of projected scores along the first canonical axis in figure 4. Scores representing C. kirrhos , n. sp., are aligned with those for the samples of C. shawmayeri from Mt. Hagen and Mt. Wilhelm in the Central Cordillera west of the Owen Stanley Ranges, and not with scores for the three population samples of C. ruemmleri from the Snow Mountains. Covariation in breadth of mesopterygoid fossa, height of braincase, and lengths of bony palate, rostrum, and maxillary molar row are principally responsible for the dispersion of scores along the first axis, particularly rostral length and breadth of mesopterygoid fossa; breadth of first upper molar also contributes but with less impact (see table 2, and diagram of factor loadings in fig. 4). Compared with the three series of C. ruemmleri from the Snow Mountains, samples of C. shawmayeri from Mt. Hagen and Mt. Wilhelm, along with C. kirrhos , n. sp., from Mt. Dayman, have a relatively wider mesopterygoid fossa, narrower braincase and first upper molar, but shorter bony palate, rostrum, and maxillary molar row. This pattern supports the assumption that the population of C. kirrhos , n. sp., in the Maneau Range (and, as we will demonstrate below, populations elsewhere in the montane reaches of the Owen Stanley Ranges) is more closely related to populations of C. shawmayeri in the eastern half of the Central Cordillera than to those of C. ruemmleri in the western region, suggesting past geographic isolation and subsequent morphological divergence from C. shawmayeri that eventually resulted in the distinctive phenetic traits now characterizing C. kirrhos , n. sp.

We also uncovered patterns in covariation of cranial and dental variables by employing a range of principal components analyses designed to test the significance of covariation among different sets of geographic samples. For this multivariate approach, various AMNH samples were combined in different analyses with BBM population samples from the Star Mountains, the Telefomin Valley, Bulldog Road in the Wau area, and Mt. St. Mary (table 1). Single specimens from Smith’s Gap (BBM), Dumae Creek and Mt. Simpson (BBM) in the Manau Range, and the holotype of shawmayeri (BMNH) were also included. The patterns revealed in the principal components analyses, combined with our observations of qualitative external traits and univariate summaries of cranial and dental variables, identify the sample from the Star Mountains as C. ruemmleri ; the material from the Telefomin area, Mt. St. Mary, and six of the seven from Bulldog Road as C. shawmayeri ; and one individual from Bulldog Road, a specimen from Smith’s Gap, one from Dumae Creek, and another from Mt. Simpson as C. kirrhos , n. sp. Results of these analyses are presented in the appropriate accounts of species.

The first account consists of a detailed description of C. ruemmleri , the type species of the genus, and covers its morphology based on evidence primarily from skins and skulls, known geographic range, habitats, and whatever other biological information that is available in the literature. The characteristics of C. ruemmleri form the standard to which the other two species, C. shawmayeri and C. kirrhos , n. sp., will be compared.