Coccymys shawmayeri ( Hinton, 1943 )

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 shawmayeri ( Hinton, 1943)

In 1943, Hinton described Rattus shawmayeri ( Hinton, 1943: 556) from a single specimen collected by Fred Shaw Mayer in the eastern portion of the Bismarck Range in Papua New Guinea, remarking that the taxon ‘‘represents evidently a species of Rattus totally unrelated to any hitherto described from New Guinea, and superficially most like Rattus eha [5 Niviventer eha ] of the eastern Himalayas.’’ It is, continued Hinton,

characterized by very small size for the genus, … strongly elongated tail, … small bullae, very weak supraorbital ridges, palatal length less than half occipito-nasal length, but palatal foramina long. … The lower incisor root of the new species forms a noticeable knob on the outer side of the jaw, just behind the coronoid process. Molars without special peculiarity. … The frontals are strongly constricted and the brain-case is broad.… The tail is mostly dark in colour, but the terminal 25 mm is pale. There is a certain growth of hair throughout its length, and the tail is different in this character from Melomys . The fur is soft; back dark brown, belly whitish grey, fifth hindtoe relatively long. Approximately fourteen rings to 10 mm on the tail.

Hinton could have been describing a specimen of ruemmleri , but this association between that taxon and shawmayeri would not be formally realized until 1993 when Musser and Carleton (1993: 585) listed the latter as a synonym of Coccymys ruemmleri . During the two and a half decades following the publication of Hinton’s description, shawmayeri was treated as a species of Rattus . Ellerman (1949: 39) placed ‘‘ Rattus shawmayeri ’’ in a group of Rattus ‘‘With small bullae, … with short palate, less than half occipitonasal length …’’ that also contained niviventer, fulvescens, coxingi, chrysocomus , cremoriventer, huang, bartelsi, eha , and inflatus, a conglomeration of taxa that are now associated with three genera, Bunomys , Niviventer , and Maxomys , none of them native to New Guinea ( Musser and Carleton, 2005). In the same volume Ellerman (1949: 52–53) formally identified this cluster as the subgenus Maxomys and separated the species into groups. ‘‘The eha Group contains two very small species with extremely elongated tail and weakly ridged skull ( eha [a distinct species of Indomalayan Niviventer ] and shawmayeri ).’’ The latter ‘‘is a remarkable species,’’ enthused Ellerman, ‘‘even smaller than eha to which it comes nearest, with shorter toothrow and narrower frontals than in that species, but very reminiscent of it in most essential characters. Its tail appears to be too hairy for shawmayeri to be referred to Melomys . It seems totally different from other New Guinea Rattus rats.’’

In his treatise on the rodents of Australia and New Guinea, Tate (1951: 341) recognized ‘‘ Rattus shawmayeri ,’’ summarized Hinton’s description, and commented that ‘‘Hinton distinguished shawmayeri from the exulans group but did not compare it with any relative of niobe . Only the type is known.’’ Tate did not comment on the proported relationship between eha and shawmayeri suggested by both Hinton and Ellerman.

The following year, Laurie’s report of the mammals collected by Shaw Mayer in New Guinea from 1932 to 1949 included an account of Rattus shawmayeri , which identified seven additional specimens from the Hagen and Bismarck ranges (see gazetteer). ‘‘These specimens,’’ noted Laurie (1952: 306), ‘‘are a useful addition to our collection in which, so far, the type of the species has been the only representative.’’

By 1954, shawmayeri was still regarded as a unique species of Rattus . In their list of the the land mammals of New Guinea, Celebes, and adjacent islands, Laurie and Hill (1954: 118) recognized the species (in the subgenus Maxomys , most certainly following Ellerman’s 1949 arrangement) and were vividly impressed by it: ‘‘A remarkable form, very different from other New Guinea Rattus . It is smaller than Rattus eha eha from Nepal and Sikkim, India, to which it is perhaps most closely related.’’

In Misonne’s (1969) treatise on evolutionary trends among African and Indo-Australian Muridae , he recognized Maxomys as a genus and excluded shawmayeri from it, disagreeing with Ellerman’s (1949) arrangement ( Maxomys as a subgenus of Rattus and containing shawmayeri ). Misonne still regarded shawmayeri as a species of Rattus , but placed it in the subgenus Stenomys , which to him was ‘‘equivalent to the assimilis-ruber-leucopus-niobe Division of Tate (1951), but not to the Stenomys group of Ellerman (1949) nor to the division adopted by Laurie & Hill (1954) ’’ ( Laurie and Hill, 1954: 136). Misonne’s action disassociated shawmayeri from Indochinese eha and brought it closer to native New Guinea murines, but still left the taxon incorrectly allocated.

Between 1969 and 1993, Rattus shawmayeri , with two exceptions, seems to have disappeared from the literature covering New Guinea mammals. Taylor et al. (1982) did not include it in their revision of Rattus from the New Guinea region, because, they noted, ‘‘A study by J.A. Mahoney of the holotype and only known specimen has shown it to be an example of ruemmleri Tate and Archbold, 1941 , a species now listed in Pogonomelomys ,’’ and Menzies (1990) apparently failed to examine the holotype of shawmayeri or any of the BMNH specimens identified as that taxon by Laurie (1952) for his revision of Coccymys ruemmleri , as none are included in his list of material examined or anywhere in his report. Neither of Flannery’s (1990, 1995) editions of ‘‘Mammals of New Guinea’’ referred to shawmayeri in any context, even as a synonym of ruemmleri . The second exception is the ‘‘Ecological check-list of New Guinea Recent mammals’’ by Ziegler (1982: 880) where in the account of ‘‘ Pogonomelomys ruemmleri ,’’ he indicated that the species included Rattus shawmayeri . Finally, by 1993 Musser and Carleton (1993: 585) formally recognized shawmayeri as a synonym of C. ruemmleri as a result of Musser’s visits to BMNH where he studied the holotype and specimens reported by Laurie (1952); this allocation was repeated in 2005 ( Musser and Carleton, 2005: 1307).

As we discovered and explain below, Hinton’s shawmayeri is not a synonym of Coccymys ruemmleri but the oldest name for a separate species of Coccymys .

HOLOTYPE AND TYPE LOCALITY: The holotype of Coccymys shawmayeri is an adult female (BMNH 1947.1155) collected by Fred Shaw Mayer (original number 636) during May 1940. It consists of a stuffed museum study skin and associated skull (fig. 24), both in good condition. Occlusal surfaces of the molars are worn; cusp patterns remain evident on first molars but are blurred on the second and third teeth. External, cranial, and dental measurements are listed in table 7.

The type locality is Baiyanka (05 ° 469S, 145 ° 109E), Purari-Ramu Divide, in the southeastern portion of the Bismarck Range at 8000 ft (2440 m), Eastern Highlands Province, Papua New Guinea (locality 14 in gazetteer and fig. 2).

EMENDED DIAGNOSIS: Coccymys shawmayeri resembles C. ruemmleri in physical appearance (see measurements listed in table 3) and most morphological traits associated with the skull and teeth (compare figs. 9 and 24). It is distinguished from C. ruemmleri by a paler and slightly shorter coat covering upperparts of head and body, longer tail relative to length of head and body, and significantly higher frequency of white tail tips in the populations. Coccymys shawmayeri has an absolutely shorter rostrum and bony palate compared to C. ruemmleri , a lower braincase (less bulbous), shorter maxillary and mandibular molar rows, and also differs in proportions of particular cranial and dental variables, which will be reported below.

GEOGRAPHIC DISTRIBUTION: The material listed in the gazetteer describes a Central Cordilleran distribution for C. shawmayeri that is confined to Papua New Guinea (fig. 2). Specimens obtained in montane forests between 2300 and 2800 m in the Telefomin region in Sandaun Province define the known western limit of the range. A sample collected in upper montane (‘‘mossy forest’’) at 3000 m on Mt. St. Mary in the western portion of the Owen Stanley Range is the easternmost record. Between these extremes, C. shawmayeri has been encountered in the high reaches of the Cordillera wherever serious mammalian inventories and ecological studies have been undertaken. The species has not been recorded from the north coastal Bewani, Torricelli, and Prince Alexander ranges in Sandaun and East Sepik provinces, in the coastal Adelbert Range of Madang Province, or in the mountains of the Huon Peninsula (Madang and Morobe provinces).

Whether the Telefomin region is the actual western limit of C. shawmayeri ’s range is unknown. It is not present in the material collected by the Archbold Expedition on the northern slopes of the Snow Mountains far to the west of the Telefomin area. The lowest records for Coccymys along that transect are from the Bele River valley at 2800 and 2200 m and those samples are definitely C. ruemmleri (see the account of that species).

The Telefomin area is interesting in another distributional aspect. The highest record for C. shawmayeri there is 2800 m; above that contour on the Star Mountains, at least by 3100 m, the upper montane forests are occupied by C. ruemmleri , which is the known easternmost occurrence of that species. To the east beyond the Star Mountains, C. shawmayeri is alone throughout most of the Central Cordillera, ranging upward beyond 2800 m into upper montane forests and alpine grasslands.

Mt. St. Mary, in the western half of the eastern Papuan peninsula, may actually delimit, or be near, the eastern margin of C. shawmayeri ’s geographic range. Mammal surveys in the mountains east of there, particularly of Mt. Albert Edward and other peaks in the Wharton Range, have not revealed the species (the 1933–1934 Archbold Expedition to the Wharton Range and nearby highlands is an example). It is also in the western portion of the eastern peninsula that another species of Coccymys is found, C. kirrhos , n. sp., with records from Bulldog Road in the Wau area, Smith’s Gap near Mt. St. Mary (thus overlapping the range of C. shawmayeri ), and the Maneau Range at the end of the Owen Stanley Ranges (see account of the new species). This peninsular species appears to be the sole representative of Coccymys beyond Mt. St. Mary. How the interplay of its present distribution and past geological and climatological events may have influenced the eastern extent of C. shawmayeri ’s range is unknown.

The recorded extremes between 1600 m in lower montane forest at Nondugl near Mt. Hagen and 3660 m in subalpine forest and alpine grassland on the eastern slopes of Mt. Wilhelm bracket the altitudinal distribution as recorded by voucher specimens along the Central Cordillera in Papua New Guinea (table 13).

DESCRIPTION AND COMPARISONS: Coccymys shawmayeri and C. ruemmleri are physically very similar: both have soft and dense fur, a small body, much longer tail, and narrow hind feet. Average differences among all samples in fur coloration and thickness, along with cranial and dental dimensions, contrast the two species. Fur covering upperparts is shorter (9–12 mm) in C. shawmayeri , compared with the thicker coat of C. ruemmleri (11–14 mm); the upperparts are paler, a warm brownish gray with slightly burnished highlights, and none of the individuals we studied show any degree of black suffusion (brownish gray through very dark brown to dark brown suffused with black in C. ruemmleri ); underparts range from whitish gray to a darker grayish white, but the paler tone is more common and no specimens examined exhibit any buffy wash (underparts of C. ruemmleri are whitish gray to dark grayish white and dark gray washed with buff in a few individuals); the tail is typically paler, grayish brown (instead of brown or dark brown), and longer relative to length of head and body (LT/LHB 5 145%–170%, as opposed to 140%–146% in C. ruemmleri ; table 3). Most specimens of C. shawmayeri (76%) have a conspicuous and moderately long white tail tip (mean 5 27.0 mm, range 5 4–42 mm); only 6 of 55 specimens of C. ruemmleri (11%) show a white tip, and the length averages shorter in those six (mean 5 19.5 mm, range 5 5–30 mm; table 8). The animal from the Telefomin area shown in Flannery’s (1995: 275) color plate demonstrates the warm brown body fur and long tail with a conspicuous white tip that is so typical of C. shawmayeri (Flannery referred to it as C. ruemmleri , but we identify the voucher as C. shawmayeri ).

Values for mass are available for 16 adults (in AMNH) from the eastern slopes of Mt. Wilhelm. All were weighed by Hobart Van Deusen who recorded weights in ounces, which we converted to grams: mean and standard deviation 5 32.9 ± 6.65 g, range 5 22.7–45.4 g. The range of variation is comparable to that reported by Flannery (1995: 274) for C. shawmayeri (as ruemmleri ): 26 g for an adult female and a range of 30–34.8 g for three adult males. A young adult C. ruemmleri caught on Mt. Capella in the Star Mountains, the only specimen of that species measured for mass, weighed 34 g, within the range of body mass recorded for C. shawmayeri .

Skulls in all samples of the two species are generally alike in size and shape, but contrast in certain dimensions, which is evident when skulls of each are studied side-by-side (com-

TABLE 13 Summary of Altitudes (m) and General Habitats where Samples of Fossil and Modern Coccymys ruemmleri , Coccymys shawmayeri , Coccymys kirrhos , and Brassomys albidens Were Collected pare figs. 9 and 24), in the graphic results of discriminant function analyses (fig. 4), and in the differences between univariate means (tables 4, 5). Generally, Coccymys shawmayeri has a conspicuously shorter rostrum and bony palate compared with C. ruemmleri , and a lower braincase (less bulbous) and smaller maxillary and mandibular molars (shorter molar rows and narrower molars). Less marked is the narrower rostrum and shorter diastema of C. shawmayeri .

The distinctions between C. shawmayeri and C. ruemmleri in chromatic aspect of the dorsal pelage and cranial and dental dimensions are most pronounced in the Star Mountains between the sample of C. shawmayeri from combined localities in the Telefomin area between 2300 and 2800 m (the westernmost recorded limit of the species) and the series of C. ruemmleri from higher in the Star Mountains at 3100 and 3200 m (the easternmost recorded extension of this western Cordilleran species). Every specimen of C. shawmayeri from those lower altitudes has warm brown upperparts and whitish gray venters, while all the C. ruemmleri have very dark brown dorsal pelage, some with a weak suffusion of black, and very dark grayish white underparts, a few suffused with buffy hues. Every cranial and dental dimension measured is less in C. shawmayeri , some markedly so, as reflected in the tables of univariate means and graphically in the principal components ordinations (fig. 14; tables 4, 5).

At the eastern recorded margin of its geographic range, C. shawmayeri overlaps the distribution of C. kirrhos , n. sp.; the two species clash in fur and tail coloration, absolute dimensions of the skull, and some cranial proportions. Compared with C. shawmayeri , the sample of the new species has bright tawny or orange brown dorsal fur, and an absolutely longer tail, on average, that is pale brown and without any white tip (tables 3, 8). Coccymys kirrhos , n. sp., has a smaller skull with a relatively much narrower interorbit, longer rostrum and bony palate, and smaller bullae (tables 5, 6). See the account of the new species for an extended discussion of similarities and differences between the two species.

GEOGRAPHIC VARIATION: Specimen scores projected on the first and second components extracted from principal components analyses in figure 25 illustrate covariation in cranial and dental variables among geographic samples of C. shawmayeri along the Papuan Central Cordillera from the Telefomin area to Mt. St. Mary. Size of braincase and breadth of first upper molar are negligible in influencing the spread of scores along the first axis, which expresses size in this case, but all other variables are to some degree responsible for the dispersion (table 14). But no obvious internal structure exists that would suggest some significant geographic pattern in covariation of cranial and dental dimensions. The large series from Mt. Wilhelm, for example, overlaps nearly every other geographic sample included in the analysis.

Scores for the five specimens from the Telefomin area form a marginal cluster along the second axis, a shape factor. Compared with other samples, the Telefomin specimens have a relatively wider zygomatic plate but narrower mesopterygoid fossa, the two variables most responsible for the dispersion of scores along the second axis (table 14). The holotype of shawmayeri from the Bismarck

TABLE 14 Results of Principal Components Analysis of Geographic Samples of Coccymys shawmayeri Samples are from the Telefomin area , Mt. Hagen region , Bismarck Range , Wau area, and Mt. St. Mary ; the holotype of shawmayeri is included. (Principal components are extracted from a covariance matrix of 18 log-transformed cranial and dental variables; see tables 5 and 7, fig. 25.)

Range near Mt. Wilhelm shares these proportional traits.

The Telefomin specimens contrast with the other samples in tail patterning. Only five of the eight specimens (63%) show a white tail tip, the lowest frequency among all the samples of C. shawmayeri (table 8). All other physical traits—lengths of hind foot and ear, relative length of tail, and color and thickness of fur—match those features in most of the other geographic samples (table 3).

It seems improbable that significant phenetic variation in the variables we examined does not exist among the populations of C. shawmayeri occurring along the high central dividing range of Papua New Guinea. Our inquiry was not really designed to resolve that aspect of morphological variation. For example, although the specimens we gathered into population samples are adults, a range of ages exists within that category, from young adults to old adults, which partly explains (along with individual and possibly secondary sexual variations) the great spread of scores for the specimens from Mt. Wilhelm, our largest sample, along the size axis in the principal components ordination (fig. 25). A more refined sorting of age classes and analyses using only members of the same class would likely be more revealing. Therefore, larger samples from regions now represented by few specimens (the Telefomin area and Bulldog Road, for example) would be useful. We have defined C. shawmayeri in comparison with C. ruemmleri to the west and C. kirrhos , n. sp., to the east, but assessment of the internal geographic variation among populations of C. shawmayeri will require further inquiry.

HABITAT: The close physical resemblance between C. ruemmleri and C. shawmayeri is mirrored in their habitat affinities. As with that species endemic to the great dividing range in western New Guinea, C. shawmayeri inhabits mountain forests and grasslands, from lower montane forest formations at 1600 m through deep mossy upper montane forest to landscapes of alpine tussock grassland margined by clumps of subalpine forest at altitudes above 3500 m.

A large sample of C. shawmayeri in AMNH comes from the east slopes of Mt. Wilhelm and south slopes of Mt. Otto in the Bismarck Range, obtained during the 1959 Sixth Archbold Expedition to New Guinea. The specimen from Mt. Otto (see gazetteer) was captured a few hundred feet above a sawmill that had been operating eight years. ‘‘Below the sawmill,’’ wrote Brass (1964: 193), ‘‘most of the original forest had been destroyed and replaced by gardens, fallow lands and Pandanus groves … and by forest regrowths and tall grass.’’ From the sawmill up to 2400 m or more, the forest had been badly damaged by logging activities. Nevertheless, Brass (1964: 194) saw that ‘‘Enough remained … of [the] forest, beginning at the sawmill, to show a complex, rich, montane, mixed rain forest … of broad leaf-gymnosperm alliance, more middle altitude in appearance and floristics.’’ It was in this zone that the C. ruemmleri was trapped, but there is no notation in Hobart Van Deusen’s field catalog as to whether the rat was taken on or above the forest floor.

Thirty-one examples of C. shawmayeri are from the eastern slopes of Mt. Wilhelm collected between 2770 and 3570 m. The specimens are simply labeled as being from ‘‘Mt. Wilhelm,’’ ‘‘East slopes,’’ ‘‘Pengagl Creek,’’ ‘‘ 2 mi E Lake Aunde, 3400 m,’’ and ‘‘Lake Aunde, 3570 m.’’ Floristic formations from the Pengagl camp at 2770 m to the Lake Piunde-Aunde camp at 3570 m ranged from lower montane forest through subalpine forest to alpine grassland, which is described in appreciative detail by Brass (1964).

The Pengagl camp at 2770 m was encircled by forest ( Brass, 1964: 190; a fine photograph of the Pengagl camp is reproduced in Brass, 1964: pl. 7, fig. 2; also see fig. 26). Cool ambient temperatures recorded during July 3 to 31 (average maximum, 17.2 ° C, extremes, 13.5–21.0 ° C; mean minimum, 6.8 ° C, extremes, 5.0–9.4 ° C) and rain nearly every day produced a cold and wet environment. Brass (1964: 191) described the forest in the vicinity of the Pengagl camp between 2600 m and 3000 m as a broadleaf-gymnosperm formation of lower montane forest,

a mixed, very moist forest of principally straight-boled trees 60 to 100 feet tall with fairly even, irregular, or broken canopy, scattered emergent taller trees, and where best developed in valleys and on moderate slopes,

having more or less distinguishable subcanopy and substage tree layers. Bryophytes abounded in a fuzzy covering or heavier growths low on the trees and on the predominantly woody undergrowth. Under thin or broken canopy, …

‘‘mossing’’ became very heavy on trunks and branches of trees and on the ground, and a slender scrambling bamboo, present everywhere in the forest, ran rampant and formed dense tangles. In occasional small seepage areas of broken canopy through fall of trees or wetness of ground, a big stilt-rooted Pandanus was especially common, numerous epiphytic orchids grew close to the ground in heavy moss,

Rhododendron maius , with big carnation-

scented white flowers, was conspicuous on logs,

and red-flowered R. vandeusenii grew low on trees.

Only three C. shawmayeri are labeled as being collected at ‘‘Pengagl Creek’’ but possibly the specimens labeled ‘‘E. slopes’’ came from the vicinity of this camp, or between there and the higher camp at Lake Aunde (3650 m). Brass (1964: 192) wrote that ‘‘Visiting natives brought specimens from probably as low as 6500 feet [2000 m] on the slopes below camp’’ and remarked that ‘‘the mammal collection from Pengagl was the largest and richest for any camp on the expedition [as opposed to the paucity of mammals collected at the higher Lake Aunde camp].’’ But Brass (1964: 189) also noted that ‘‘From levels below the subalpine [referring to the Lake Aunde camp], the Chimbu brought mammals for sale almost daily in good weather; all species so obtained were taken later at Pengagl Camp and are enumerated for that locality.’’ So exactly where the Coccymys were trapped or caught by local people is unknown. At the Pengagl camp, Hobart Van Deusen, Brass remarked, was obliged to devote his time to preparing specimens. Hobart once related to Musser a lingering frustration at being unable to check his own traplines, and his ignorance of where many of the specimens were actually caught.

The Lake Piunde-Aunde camp at 3560 m ‘‘was on a flat-topped, open, grassy rise at the lower end of the first of two closely adjacent, deep, alpine lakes’’ ( Brass, 1964: 185) surrounded by a mosaic landscape of fragment- ed subalpine forest, tree ferns and alpine tussock grassland (views beautifully preserved in the photographs published in Brass, 1964: pls. 10 and 11; also see figs. 27, 28). Mean maximum and minimum ambient temperatures recorded for June 14 to 28 were 13.3 ° C (extremes, 10–16.5 ° C) and 3.8 ° C (extremes, 2.5–5.0 ° C), respectively. During this period, Brass described days with light to heavy frost, occasional icing of wet ground and small pools, and snowfall during two days on the heights of Mt. Wilhelm.

Brass (1964: 187) wrote of the forest that,

A considerable amount of fragmented, heavily mossed, low, subalpine forest occurred on slopes in the vicinity of camp, diminishing rapidly up to tree limit at 13,000 to 13,200 feet

[3965–4026 m], and increasing downward until it almost completely covered the slopes and all but the cold, mostly ill-drained bottom of the U valley. … Down the mountain, the subalpine forest merged into a different forest type in a broad ecotone, with a rather definite change apparent at about 10,800 [3294 m]. Typically, the subalpine forest consisted of a dense canopy layer, 25 to 30 feet high, of generally crooked small trees with small stiff leaves and erect stiff branches. … Within the forest, a thin bryophyte layer constituted the only ground covered in very dense shade. Elsewhere an occasional herb entered from the borders to associate with a few true undergrowth plants. …

The grassland near the Lake Aunde camp also captured Brass’s attention ( Brass, 1964: 188):

Alpine grassland, as original primary vegetation, occupied the mountain above the forest zone, and below that ground too wet or too cold for forest. At camp level, in the valley bottom down to 10,900 feet [3325 m], and up to variously 13,000 to 13,500 feet [3965–4118 m] on open slopes, was tussock grassland. At lower to middle levels the long-grass or tussock grassland (‘‘peaty grassland’’ of Hoagland) occupied mainly poorly drained or seepage-wet peaty soils, at higher levels well-drained soils, some of which … formerly carried forest.

The grassland communities consisted not only of a great many species of grass, but a diversity of herbs, low shrubs, ground ferns, and cycadlike tree ferns resistant to frost and fire.

Of the 14 Coccymys obtained at ‘‘Lake Aunde,’’ site information (taken from notations on skin tags or Van Deusen’s field catalog) is associated with only 4: AMNH 192092 and 192105 were ‘‘snared in long grass,’’ 192100 was ‘‘snared in peaty grassland,’’ and 192107 was ‘‘snared in grass near subalpine forest.’’

Other associations between specimens and habitats come from published accounts or notations on tags attached to museum skins. One juvenile C. shawmayeri (AM 16745) was taken from the crown of a Pandanus at 2600 m from the upper Sol River valley during a mammal survey conducted in

TABLE 15

Summary of Habitat Information Associated with Specimens of Coccymys shawmayeri Stored in the Bernice P.

Bishop Museum

(Nearly all specimens were collected by A.B. Mirza. Information comes from notations on skin tags.)

Sandaun Province. In 1984 and 1986, Tim Flannery’s base camp ( Flannery and Seri, 1990: 178):

was beside the westwards-flowing headwaters of the Sol at an altitude of between 2,200 and 2,300 m. The floor of the valley at this altitude is gardened by the inhabitants of Telefolip Village. The valley walls rise sharply to the south, north and east enclosing a relatively flatlying basin, and gardening is restricted to its lower parts. … Away from the gardened zone, the valley is clothed in climax Lower Montane Forest. Even at altitudes of 2,600 m large fruited Castanopsis are common. … The forest does not become very mossy until 2,600 to 2,800 m. The lower part of the valley floor is dominated by gardens and secondary growth in various stages.

Willett et al. (1989: 11) collected a specimen in ‘‘mid-montane rain-forest’’ on the southern flanks of Mt. Missim at 2000 m. They provide several views of the forest where surveys were conducted but no details describing where the specimen was caught.

The specimens of C. shawmayeri in the Bishop Museum collected by A.B. Mirza have short habitat notes on the skin tags, which are tabulated in table 15. The geographic coverage mirrors the Cordilleran range of C. shawmayeri , and the range of habitats at the collection sites extends from lower montane forest through upper montane formations to alpine grasslands.

BIOLOGY: We have been unable to locate reproductive information for Coccymys shawmayeri . None of the specimens we examined for this report carried any notes on litter size, breeding periods, or any other reproductive aspects, and the published literature is unhelpful. (Many of the specimens were collected by A.B. Mirza, who recorded succinct habitat notes on skin labels

TABLE 16

Contents from Stomachs of Coccymys ruemmleri , Coccymys shawmayeri , and Coccymys kirrhos

Species, locality, and specimen Contents

C. ruemmleri Lake Stomach empty: fecal pellet in intestine consisted of unidentifiable dark brown mash and a small

Habbema 152740 cockroach nymph (first or second instar)

C. shawmayeri Entirely insects, probably cockroaches: very small dark brown fragments of sclerites, a few long

Nondugl 183634 filamentous antennae, partly digested tissue adhering to many of the larger sclerite fragments; translucent gelatinouslike pieces resembling partly digested soft tissues. Fragments are likely from early instar cockroach nymphs. Did not find mouth-parts, legs, or wings

183609 Mostly fruit, at least one insect larva, maybe more: very small pieces of fibrous fruit pulp; small fragments of fig with the inner lining adherent and tiny seeds embedded in it, many isolated sections of tissue filled with tiny fig seeds; a few small sclerite fragments, the larger pieces attached to partly digested tissue; and translucent sections from the body of an insect larva

183633 Mostly insects, a little fruit: many very small beetles (head fragments, short segmented antennae, legs, membranous wings), small cockroach nymphs, early instars (long, filamentous antennal segments, abdominal sclerites); intact early instar nymphs of katydids along with long thin antennal segments; and partly digested soft tissue mixed with translucent ventral abdominal pieces; a few small chunks of fruit pulp, some resembling outer surface of skin; no seeds

Mt. Wilhelm Fruit only: suspension of digested pulp and fragments of brown seed coats—one kind of fruit; similar

192734–738, to contents of 156588 and156546 from Mt. Hagen

192740, 192307

156618 Fruit only: suspension of fruit pulp

192733 Fruit only: finely digested pulp, many tiny black seeds that resemble fig, one fragment of possible fig skin; some larger brown seeds from a different fruit

156608 Fruit, vegetative fragments, insect larvae: suspension of digested pulp, fragments of brown seed coats, small wedge-shaped brown seeds, few grass seeds; plant tissue (stems, leaves, petals), and many small (5 mm long) insect larvae (probably Coleoptera )

192105 Mostly fruit: suspension of fruit pulp, tiny bits of seed coats, many small flat seeds; one intact flea (probably ingested during grooming)

192092 Fruit only: suspension of pulp, with fragments of seed coats, small pieces of skin from fruit

Mt. Hagen 156490 Fruit only: numerous tiny brownish black oblong seeds and partly digested fruit pulp; appears to be only one kind of fruit present; stomach greatly distended with these contents

156548 Mostly fruit: pulp and pieces of brown seed coat; some small (3 mm long) intact larvae (probably Diptera ), the kind found infesting fruit; no other type of insect remains

156605 Fruit only: some digested pulp, but mostly two kinds of seeds, tiny irregularly-shaped blackish seeds, and much larger, round flat seeds (same kind as in 192733)—definitely two kinds of fruit consumed

156547 Insects: sclerite fragments and mouthparts from adult beetles, suspension of digested tissue; few small plant fragments; unidentifiable suspension of brown particles, either fruit pulp or insect tissue (stomach poorly preserved)

156546, 156588 Fruit only: stomachs full of digested pulp and tiny fragments of brownish seed coat—apparently only one kind of fruit

C. kirrhos Mt. Fruit only: small amount of pulp suspension, large tan and brown pieces of seed coat

Simpson 184493

but no reproductive information. His field journals, which we have not read, are stored at the Bernice P. Bishop Museum and should be consulted by those researchers planning any reports covering reproductive information for C. shawmayeri .)

Documented dietary data for C. shawmayeri is equally elusive. Menzies and Dennis (1979: 40) noted that ‘‘Limited evidence indicates that the diet [of ‘‘ C. ruemmleri ’’] consists only of vegetable matter including leaves,’’ but they cited no source for their observations. We turned to specimens preserved in fluid and stored at AMNH to extract stomachs, and surveyed the contents from 21 specimens of C. shawmayeri collected in Papua New Guinea at Nondugl, Mt. Wilhelm, and Mt. Hagen. Their contents are summarized in table 16. They collectively reflect a diet composed of fruit, seeds, and insects, although most stomachs contained only fruit, or mostly fruit with some insect material; the remaining stomachs contained only insects, or mostly insects with some fruit. One stomach was full of fruit, small insect larvae, and bits of vegetative plant tissue. The bulk of the insects are in pieces and had been masticated, indicating active capture and chewing; some of the very small larvae are whole and may have been passively ingested with fruit.

Particular anatomical traits characteristic of Coccymys are consistent with a mixed diet of fruit and insects; dentition and stomach anatomy are examples. The morphology of the incisors is not specialized; their shape, pigmentation and extent of enamel cover, and size relative to the skull and mandible form a configuration common among murines. Occlusal patterns of the molars typical of Coccymys , especially those of the upper molars, provide rows of cuspidate and wide surfaces suitable for masticating fruit or insects. Broadly, but not in detail, they resemble the coronal patterns seen in the species of Margaretamys from Sulawesi ( Musser, 1981a: 284–285), particulary the upper molars. All three species of Margaretamys are arboreal and feed on a variety of different fruits, as well as katydids, small cicadids, camel crickets, and moths (Musser’s observations in the field). Their incisor anatomy is also generalized compared with that in such highly specialized invertebrate predators as the Sulawesian and Philippine shrew rats, for example ( Musser and Heaney, 1992; Musser, 1982b).

Coccymys shawmayeri (and the other two species in the genus) has a unilocular-hemiglandular type of morphology, which is described by Carleton (1973). The stomach consists of a single chamber (unilocular), and the gastric mucosa is bisected nearly equally into glandular epithelium to the right of the esophagus, lining the antrum, and cornified squamous epithelium to the left, lining the corpus (hemiglandular); the division between the two epithelial linings is marked by a bordering fold extending from the incisura angularis (at the right of the esophagus) directly across to the greater curvature of the stomach. This single-chambered hemiglandular morphology, in which the glandular zones are separated by a smooth bordering fold, forms the gastric conformation that, as suggested by Carleton (1980: 101), represents the primitive evolutionary state among muroid rodents. The fully distended stomach of Coccymys ruemmleri resembles this general unilocular-hemiglandular conformation, and is similar to that of the Sulawesian Rattus hoffmanni , figured in Musser and Durden (2002). The design is also common to the species of Sulawesian Margaretamys , in which the diets consist of fruit and insects.

Being able to consume fruit, seeds, and insects is probably advantageous to species such as C. shawmayeri and its congeners, which we suspect to be scansorial (see discussion in account of Brassomys albidens ). The very long tail relative to length of head and body, and moderately long and slender hind feet possessed by C. shawmayeri and the other two species are appendage proportions and shapes usually associated with scansorial activities. We view a scansorial animal as one that forages over the ground (partly terrestrial) and climbs into shrubs, pandans, tangles of woody vines, small trees, and lower branches of larger trees within the forest understory (partly arboreal); nests may be anywhere in that vertical range. Judged by collection data attached to some specimens of C. shawmayeri , individuals were trapped on the ground, on a tree trunk laying on the ground, on trees 3–4 ft above ground, and in the top of a Pandanus (see table 16 and ‘‘Habitat’’), which describes the vertical range of scansorial activities. Opportunities for food within those horizontal and vertical habitat spaces are measurably increased when the animal is able to eat fruits, seeds, and vegetative plant parts, as well as insects.

Coccymys shawmayeri is also nocturnal and one of the species preyed upon by the sooty owl, Tyto tenebricosa . These assertions are based on contents of owl pellets. Beneath a boulder on a ridge rising sharply behind the camp at Lake Aunde (3560 m, June 1959) on the eastern flanks of Mt. Wilhelm, Hobart Van Deusen uncovered a cache of owl pellets. The boulder is about 100 m above the camp; the landscape is dominated by alpine tussock grassland surrounding islands of subalpine forest (see pl. 11, fig. 2 in Brass, 1964). Feathers found at the roost, wroteVan Deusen in his notes (in Mammalogy Archives at AMNH), were collected and later identified as coming from a sooty owl; no feathers of any other kinds of birds were seen around or beneath the boulder. Back at the museum the undigested bones and teeth (no feathers were discovered) were extracted from the pellets and studied. Van Deusen identified remains from the small-bodied marsupials Microperoryctes ornata (8 individuals, initially determined as M. longicauda [see Helgen and Flannery, 2004, for the use of ornata ]) and Carcartetus caudatus (5 individuals), and made provisional determinations for a few of the small rodents found. We examined all the rodent material (mostly cranial fragments and dentaries, with or without molars and incisors) and found 56 individuals of Coccymys shawmayeri (AMNH 276640–95), 40 Rattus niobe (AMNH 276600–639), 2 Pseudohydromys fuscus (AMNH 276598, 276599), 2 Abeomelomys sevia (AMNH 276596, 276597), and 3 Pogonomys sylvestris (AMNH 276593–95). At least during 1959, C. shawmayeri formed a significant component of the sooty owl’s diet. Tyto tenebricosa ranges throughout New Guinea from coastal lowlands to treeline, nests in hollow trees and understory tangles during the day and at night preys on small marsupials and rodents in a variety of forest formations below treeline, and in subalpine forest and adjacent alpine grasslands at high altitudes. Pellets from T. tenebricosa from a cache high on Mt. Wilhelmina also contained Coccymys , in this case C. ruemmleri , but the number of individuals relative to other rodent prey was much lower (see the account of that species).

Other biological aspects of Coccymys shawmayeri have been explored, although the voucher specimens used in these studies were identified as C. ruemmleri . Sperm head morphology and its significance was discussed by Breed (1997, 2004) and Breed and Aplin (1994). Genetic studies that included C. shawmayeri are represented by Watts and Baverstock (1994, 1996) who sampled one individual for their research into intergeneric clustering in Australo-Papuan murids based on albuminological comparisons (microcomplement fixation of albumin). The species has yet to be included in any published phylogenetic studies based on sequences from mitochondrial or nuclear genes.

SYMPATRIC ASSOCIATIONS: The regional sympatric relationship between the western Cordilleran C. ruemmleri and C. shawmayeri was explored in the account of the former species. Overlap of the eastern part of the geographic range of C. shawmayeri with the western distributional segment of C. kirrhos , n. sp., the eastern peninsula endemic, will be entertained in the account of the new species.

Coccymys shawmayeri also occurs sympatrically with a core of high-altitude Cordilleran marsupials and rodents. The suite of species collected by members of the Archbold Expedition from the eastern slopes of Mt. Wilhelm provides an example of the montane community. At the Lake Piunde-Aunde locality (3560 m), wrote Brass (1964: 189),

Rattus niobe was trapped easily in grass and forest, and snared in large numbers on tussock grassland by visiting natives; [ Abeomelomys ]

sevia was trapped fairly frequently on grasslands, occasionally in forest. … Mallomys rothschildii … was trapped only on tussock grassland but probably entered the forest. This also applies to the bandicoot [ Microperoryctes ]

longicauda [5 ornata ]. The gentle little marsupial [ Carcartetus caudatus ] of the forest, caught by hand in old bird’s nests, and trapped, and the ringtail [ Pseudochirops ] cupreus doubtfully ven-

tured far from forest cover at any time. No bats were seen.

‘‘Containing 34 species, the mammal collection from Pengagl [2770 m and lower] was the largest and richest for any camp on the expedition,’’ enthused Brass (1964: 192):

The collection comprised, in marsupials, [ Murexia ] melanurus , … [ Microperoryctes ] [ ornata ],

Phalanger [ sericeus ], [Carcartetus] caudatus ,

[ Pseudochirulus ] forbesi, [ Pseudochirops ] cupreus, P. corinnae , Petaurus breviceps , Dorcop-

sulus vanheurni and a species of Dendrolagus . The larger marsupials were probably kept at rather low population levels by local hunters, ranging the forests with dogs, and climbing or cutting down trees for arboreal species. … The rodents showed great dissimilarity in abundance, Rattus niobe , for example, being extremely common, while six of a total of 19 species were represented by only one specimen in the collection. The species, besides Rattus niobe , were R. [steini] (from the lower levels), R. exulans , Anisomys imitator , Pogonomys sylvestris , P. [ loriae ], Lorentzimys nouhuysi , [ Protochromys ] fellowsi , … [ Melomys rufescens , Paramelomys rubex ], [ Abeomelomys ] sevia , and Macruromys major in murines. Trapped and snared in the forest were the very small, very rare, shrew-like hydromyines Pseudohydromys murinus , [P.] fuscus , and [P.] ellermani . Taken in traps in creek-bed habitats were the big water rats Baiyankamys shawmayeri , Parahydromys asper , and the highly specialized Crossomys moncktoni . …

In addition to these nonvolant species, four kinds of bats were collected, species of Pipistrellus and Miniopterus , Nyctophilus microdon , and Syconycteris australis . The rendition of species collected at the two camps is certainly an incomplete survey of the mammalian species occurring with C. shawmayeri , but they offer an insight into the montane community.

The ‘‘beautiful little Pogonomelomys ’’ that caught Brass’s (1956: 129) attention on Mt. Dayman is the subject of the following species account.