the generic status of Baiyankamys and description of a new species of Hydromys Author Helgen, Kristofer M. text Zootaxa 2005 913 1 20 journal article 51057 10.5281/zenodo.171028 d3f1ce59-f24d-4abf-9ebd-28af273be687 1175­5326 171028 Hydromys ziegleri new species Figures 3 , 4 ; Tables 3 , 4 Holotype : BBM­NG 101683 , adult female, skin ( Fig. 3 ) and skull ( Fig. 4 ), from Bainyik ( 03°40'S , 143°05'E ), circa 5 km south of Maprik, 700 ft (= 213 m ), East Sepik Province, Papua New Guinea , collected 29 October 1972 by A.B. Mirza. According to its label the holotype—the only known specimen of this species—was live­trapped in a local garden (see Ziegler 1984 : 101). Diagnosis : Hydromys ziegleri is the smallest species of Hydromys . It most closely resembles H. hussoni , endemic to the upland Wissel Lakes of west­central New Guinea ( Figs. 3–4 ; Table 3–4 ), but differs from that species in its smaller size, broadened rostrum, shorter incisive foramina, shorter toothrow, broader first molars, lower­domed braincase, and much less densely­furred tail. Distribution : Hydromys ziegleri is known only from the type locality, situated in the foothills of the southern slopes of the Prince Alexander Range, one of the North Coast Ranges of Papua New Guinea . The type locality lies in the vicinity of the Screw and Amuk Rivers, which drain from the highlands of the Prince Alexander Range and the Torricelli Range, respectively. I suspect that future collecting efforts may show it to occur more widely (and at higher altitudes) in the North Coast Ranges (see Discussion below). Two other amphibious murines ( Hydromys chrysogaster and Parahydromys asper ) are known from other localities in the North Coast Ranges and are thus likely to occur sympatrically with H. ziegleri , though only H. chrysogaster has been collected in the same general area to date (e.g. CSIRO 8346, adult male from Maprik). Etymology : For the late Dr. Alan C. Ziegler ( 1929­2003 ), former head of the Vertebrate Zoology Division at the Bishop Museum in Honolulu and authority on New Guinea mammals. Description : The holotype of ziegleri is smaller than most specimens of hussoni ( Figs. 3–4 ; Table 3 ). The dorsal pelage of ziegleri is dark brown (somewhat darker than the rich brown of hussoni ) tipped with paler, orange­buff hairs. The fur is considerably shorter ( 6– 7 mm on the anterior dorsum), less dense, and conspicuously more glossy than in hussoni . The fur of the venter is grey­based with ochraceous­buff tipping, similar to hussoni ( Musser and Piik 1982 ; Ziegler 1984 ), but in the holotype of ziegleri the lower cheeks and the underside of the throat and chin are dingy white, contrasting with the rest of the venter. The inguinal region and ventral base of the tail are dark brown ( Figure 3 ). As in other Hydromys , the hindfeet are long and broad and bear conspicuous webbing between the middle digits (II–IV). The dorsal surface of the pes is pigmented dark but covered with small pale hairs, and the claws are equal in size to those of hussoni , and unpigmented. The skin label gives the hindfoot length as 29, probably measured with the claw, as on the dry skin this length is closer to 27 sans unguis ( Ziegler 1984 ). FIGURE 3. Dorsal and ventral views of the skins of Hydromys ziegleri (a, c; BBM­NG 101683, adult female holotype) and Hydromys hussoni (b, d; AM M18627; adult female paratype). Scale bar = 20 mm. The tail of the holotype is black, and lacks a white tail­tip, which is present in 19 of 24 known specimens of hussoni , and always present in Hydromys chrysogaster , Baiyankamys , and Parahydromys (absent in Crossomys ). There are 13–15 tail rings per centimetre in the mid­section of the tail. Probably the most conspicuous external difference between ziegleri and hussoni is the relative hairiness of the tail. In other Hydromys , including hussoni , the tail is extensively haired, obscuring the tail rings, and the tail hairs are long, extending for the length of many tail scales. In ziegleri the tail is considerably less heavily furred, lending it a much more “naked” appearance, such that the tail rings are immediately evident to the naked eye. The tail hairs measure only about 2–3 tail rings in length and lie close to the body of the tail. FIGURE 4. Crania and mandibles of Hydromys ziegleri (below, adult female holotype) and Hydromys hussoni (above, AM M18627, adult female paratype). In many ways the cranial resemblance between hussoni and ziegleri is close ( Ziegler 1984 ). However, judging from direct comparisons between the holotype of ziegleri and a paratype of hussoni (AM M18627, formerly RMNH 29162, also adult female), my notes on the type series of hussoni at RMNH, comparisons with the verbal and mensural descriptions of hussoni by Musser and Piik (1982) and Voss (1988), and illustrations of the holotype of hussoni figured by Musser and Piik (1982) , many qualitative cranial differences between the two species are immediately apparent. The skull of ziegleri is noticeably shorter and stouter than that of hussoni ( Fig. 4 ). The rostrum is short and broad in both species, but the nasolacrimal capsules are laterally expanded to a much greater degree in ziegleri , as reflected in its relatively greater rostral breadth, and the nasals are correspondingly broader ( Table 3 ) and also more strongly retracted. The incisive foramina are more truncate in ziegleri , only 31% of the length of the diastema, versus 37% (in hussoni ) or greater in all other species of Hydromys and Baiyankamys ( Musser and Piik 1982: 164 ) . The zygomatic plate is more excavated in ziegleri than hussoni , and the braincase is markedly less expanded dorso­ventrally ( Table 3 ; braincase height is 64% of braincase breadth in ziegleri , versus 69% or greater in all other Hydromys and Baiyankamys ; Musser and Piik 1982 :164). TABLE 3. Craniodental metrics for the adult female holotype of Hydromys ziegleri and for series of adult female and mixed adult Hydromys hussoni . Measurements H. ziegleri H. hussoni ΨΨ 1 H. hussoni ΨΨ, ♂♂ 2 n = 6­9 n = 20 Length of head and body 132 150.1 ± 9.4 (138­164) 147.9 ± 15.4 (122­171) Length of tail 118 125.5 ± 14.3 (103­144) 126.4 ± 13.8 (103­152) Length of hindfoot 29 27.4 ± 1.1 (26­29) 29.8 ± 2.0 (27­33) Length of ear 13 12.4 ± 0.8 (12­14) 12.1 ± 0.8 (11­14) Greatest length of skull 29.01 31.54 ± 0.87 (30.1­32.8) — Condyloincisive length 28.50 — 30.5 ± 1.41 (27.9­32.9) Zygomatic breadth 14.98 15.60 ± 0.50 (14.8­16.3) 15.5 ± 0.87 (14.0­17.3) Breadth of rostrum 6.23 5.82 ± 0.21 (5.6 ­ 6.1) — Length of nasals 8.73 10.17 ± 0.43 (9.4­10.6) 9.8 ± 0.68 (8.6­11.2) Breadth of nasals 3.39 — 3.10 ± 0.18 (2.8­3.4) Length incisive foramina 2.48 3.08 ± 0.13 (2.9­3.3) 3.0 ± 0.33 (2.0­3.3) Maxillary toothrow 4.13 4.71 ± 0.15 (4.5­5.0) 4.6 ± 0.12 (4.4­4.8) Breadth of braincase 13.09 13.44 ± 0.28 (13.1­13.8) 13.7 ± 0.36 (13.0­14.4) Height of braincase 8.39 9.21 ± 0.19 (8.9­9.5) —

Postorbital constriction	5.09	5.28 ± 0.14 (5.1­5.5)	5.3 ± 0.14 (5.0­5.6)
Breadth of incisor tips	1.62	1.67 ± 0.13 (1.5­1.9)	1.6 ± 0.15 (1.4­1.9)
Length of diastema	7.99	8.39 ± 0.36 (7.9­9.0)	8.5 ± 0.54 (7.6­9.3)

1 From Musser and Piik (1982: 159) . 2 From Voss (1988: 467): 12ΨΨ and 8♂♂. The faces of the upper and lower incisors are pigmented yellowish­orange in ziegleri (with pigmentation more obvious in the upper incisors), while those of hussoni are very weakly pigmented or almost unpigmented. In ziegleri the first upper molars are broader relative to their length compared to hussoni , and the second upper and lower molars are more reduced (e.g. Table 4 ). The posterior palatine foramina are lengthened into a more extensive narrow groove in the hussoni paratype than in the holotype of ziegleri , in which these foramina are essentially small ovate holes. The posterior palatal spine is extremely well­developed. Relative to hussoni the auditory bullae are less flattened and less elongate. The foramen ovale is distinctly larger in ziegleri than hussoni . In both species the stapedial foramen is well­developed and the foramen ovale bears a conspicuous groove for the passage of the infraorbital branch of the stapedial artery, suggesting that both possess the primitive murine cephalic arterial configuration typical of other New Guinean water­rats of the genera Baiyankamys , Hydromys , Parahydromys , and Crossomys (and the less closely­related lowland false water­rat, Xeromys myoides , as stated by Musser and Carleton [2005] and confirmed by my examination of specimens in the Australian Museum; cf. Musser and Heaney 1992 :87). Additional measurements of the holotype (other than those presented in Table 3 ) include: palatal length 13.08; breadth of mesopterygoid fossa 2.13; greatest breadth of incisive foramina 1.67; breadth of the palatal bridge between M2–M2 2.60; breadth of the zygomatic plate 1.54; height of upper incisor 4.58; depth of upper incisor 1.55; breadth across occipital condyles 8.37; coronoid height of mandible 7.99; alveolar length of M1–2 4.35; and mandible length minus incisor 14.85 (with incisor, 18.38). TABLE 4. Comparative molar dimensions in specimens of Hydromys ziegleri (adult female holotype) and Hydromys hussoni (AM M18627, adult female paratype).

ziegleri	hussoni
M 1 length x width	2.98 x 1.54	3.24 x 1.49
M 2 length x width	1.22 x 1.07	1.52 x 1.34
M 2 length / M 1 length	0.41	0.47
M 1 width / M 1 length	0.52	0.46
M 1 length x width	2.59 x 1.28	2.84 x 1.59
M 2 length x width	1.67 x 1.18	1.81 x 1.34
Zoogeography

Several isolated mountain ranges are situated along the northern coast of New Guinea , including the Adelbert, Prince Alexander, Torricelli, Bewani, Cyclops, Foya (= Gauttier), and Van Rees Mountains, collectively referred to hereafter as the North Coast Ranges. These mountain ranges are well­separated from the Central Dividing Ranges of New Guinea by large expanses of lowland forest, have distinctive geological histories ( Flannery 1995 ), and support highly distinctive faunas. TABLE 5. North Coast endemic mammal species recorded from the various North Coast Ranges of northern New Guinea. PA = Prince Alexander Range; T = Torricelli Range; B = Bewani Range (including Mt. Menawa); C = Cyclops Range; F = Foya Range. PA T B C F Recorded elevation Monotremes Zaglossus attenboroughi Flannery & Groves, 1998 x 1600 m Bats

Marsupials
Dendrolagus pulcherrimus Flannery, 1993	? x	?	680 to 1120 m
Dendrolagus scottae Flannery & Seri, 1990	x x	900 to 2000 m
Petaurus abidi Ziegler, 1981	x	800 to 1220 m

Rodents
Hydromys ziegleri n. sp.	x	213 m
‘Microhydromys’ musseri Flannery, 1989	x	1350 m
Paraleptomys rufilatus Osgood, 1945	x x x	1200 to 1700 m

Hipposideros edwardshilli Flannery & Colgan, 1993 x 240 m Many endemic mammals occur in these mountain ranges ( Table 5 ). A recentlydescribed monotreme, the dwarf long­beaked echidna ( Zaglossus attenboroughi ), is known only by a single specimen collected in 1961 in upper montane forest ( 1600 m ) on Mt. Rara in the Cyclops Range, behind Jayapura ( Flannery and Groves, 1998 ). The treekangaroo Dendrolagus scottae (the “tenkile”) is known from the western Torricelli Range and Mt. Menawa in the Bewanis, where it is primarily restricted to narrow bands of montane forest situated at about 1200–1500 m in the Torricellis and about 1500–2000 m on Mt. Menawa ( Flannery and Seri 1990b ; Flannery et al. 1996 ). Another tree kangaroo, Dendrolagus pulcherrimus , is recorded as a living animal from the eastern Torricellis and (on the basis of a sight record) from the Foya Range (Flannery 1993; Flannery et al. 1996 ), and is rumoured to occur in the Prince Alexander Range. The latter species (originally described as a subspecies of D. goodfellowi by Flannery 1993) has disappeared from much of its former range in the Torricellis in recent decades, probably as a result of unsustainable hunting practices ( Flannery et al. 1996 ). In the past, D. pulcherrimus probably had a much more extensive distribution, as it or a very closely­related taxon is also represented in Quaternary subfossil deposits from the Vogelkop Peninsula of north­west New Guinea ( Aplin et al. 1999 ). A third marsupial found only in the North Coast Ranges is the Northern glider ( Petaurus abidi ), recorded only from mid­ and upper montane forest in the vicinity of Mts. Somoro and Sapau in the Torricelli Range ( Ziegler 1981 ; Flannery 1995 ). Among murine rodents, three endemic hydromyins are known from the North Coast Ranges. These are Paraleptomys rufilatus , a medium­sized terrestrial rat which occurs in mountaintop forests in the Cyclops, Bewani, and Torricelli Ranges ( Osgood 1945 ; Flannery 1995 ; Musser and Carleton 2005); ‘ Microhydromys’ musseri , a small moss­mouse of uncertain phylogenetic affinity, known by a single specimen collected in 1972 in upper montane forest ( 1350 m ) on Mt. Somoro in the Torricellis (Flannery 1989; Musser and Carleton, 2005); and now Hydromys ziegleri , known by a single specimen collected on the southern margin of the Prince Alexander Range. Finally, one species and one distinctive subspecies of leaf­nosed bat ( Hipposideros edwardshilli and Hipposideros wollastoni fasensis ) are known only from the northern and southern foothills of the Bewani range, respectively ( Flannery and Colgan 1993 ). During cooler climatic periods during the Late Pleistocene, montane forests extended to much lower elevations in New Guinea than they do today, and the montane­adapted species of the North Coast Ranges likely had considerably more expansive or even contiguous distributions across certain of these mountain ranges. I suspect that the seemingly haphazard distribution of endemic species recorded from the various North Coast ranges today ( Table 5 ) reflects a combination of two factors: artefacts of geographically uneven sampling effort and, at least for some of the larger­bodied species ( Zaglossus , Dendrolagus ), historical extinctions (see Flannery et al. 1996 ; Flannery and Groves 1998 ). Some of the larger endemic mammals of the North Coast Ranges, especially Zaglossus attenboroughi , Dendrolagus scottae , and Dendrolagus pulcherrimus , have minuscule global geographic ranges and are actively hunted for food ( Flannery et al. 1996 ; Flannery and Groves 1998 ), and are probably among the most endangered of all Australasian mammals (indeed, Z. attenboroughi may already be extinct; Flannery and Groves 1998 ). All North Coast Range endemic mammals, even the smallest (such as ‘ Microhydromys’ musseri and Hydromys ziegleri ), should be considered conservation priorities on account of their extremely limited geographic and altitudinal ranges, which in most cases are centred on small tracts of upper montane forest, highly susceptible to threatening processes including logging and climate change. In addition to supporting a markedly unique mammal fauna, the North Coast Ranges are also an area of considerable geographically­restricted endemism for frogs, birds, and other biotic groups ( Stattersfield et al. 1998 ; Kraus and Allison 2000; Allison and Kraus 2000, 2003), ranking these mountain ranges among the most critical priorities for conservation efforts in the Melanesian region ( Wikramanayake et al. 2002 ). Of the eight endemic mammal species currently known from the North Coast Ranges, seven were described as new to science only within the last 15 years, and all eight were both discovered and described within the last 60 years ( Table 5 ). Further, three of these remain known by a single museum specimen ( Zaglossus attenboroughi , ‘Microhydromys’ musseri, and Hydromys ziegleri ). These cursory but remarkable statistics for North Coast Range mammals aptly demonstrate how incompletely catalogued the biota of these mountains remains. While the mammal faunas of some areas in the North Coast Ranges (including the western Torricellis, Cyclops, and Bewanis) have become much better understood through recent survey efforts (e.g. Flannery 1993, 1995), others, such as the Foya and Van Rees Ranges of western New Guinea , remain almost wholly unknown mammalogically. For example, until recently the ornithologist Jared Diamond was the only westerner known to have visited the relatively expansive Foya Range, which is mostly uninhabited and largely unexplored (see Diamond 1985 ). Given the startling rate of new discoveries in North Coast mountain ranges in recent decades, there is no doubt that future survey efforts in this region will continue to encounter vertebrate species new to science. The mammal fauna of the Prince Alexander Mountains is moderately well­known. Apart from Hydromys ziegleri , at least 23 other species are recorded from these mountains (specimens at AM, AMNH, and BBM). These include the cuscuses Phalanger orientalis , P. gymnotis , and Spilocuscus maculatus , the tree­kangaroo Dendrolagus inustus , the wallabies Thylogale browni and Dorcopsis hageni , the bandicoots Echymipera clara , E. kalubu , and E. rufescens , the rodents Hydromys chrysogaster , Melomys rufescens , Rattus praetor , and R. exulans , the fruit­bats Pteropus neohibernicus, Dobsonia minor, Nyctimene aello , N. albiventer , N. cyclotis, Paranyctimene tenax, Syconycteris australis, and Macroglossus minimus , the vespertilionid bat Pipistrellus papuanus , and the molossid bat Mormopterus beccarii . However, essentially all of these species (excepting Nyctimene cyclotis , a widespread inhabitant of New Guinean montane forests) are characteristic inhabitants of lowland habitats and are widespread in northern New Guinea . Future surveys in the Prince Alexanders should thus focus on the very highest peaks in the range, which rise to 1240 m and are likely to support unique animal assemblages unknown at lower altitudes. As discussed above, rodent­collecting efforts in the Melanesian region have so far yielded five species of Hydromys— the widespread H. chrysogaster , the insular endemic H. neobrittanicus , the mid­montane Wissel Lakes endemic H. hussoni , the newly­described H. ziegleri , and an unnamed species from the Moluccan island of Obi. The three larger species ( H. chrysogaster , H. neobrittanicus , and the Obi taxon) have all been collected in lowland rainforest, including in forest at sea level, while one of the smaller species ( H. hussoni ) is known only from forest at much higher elevations. The similarly small­bodied Hydromys ziegleri is thus far known only by a single specimen taken in relatively low elevation forest ( 213 m ); given that extensive collecting efforts in other areas of lowland northern New Guinea have failed to record it elsewhere, I strongly suspect that it is a legitimate endemic of the North Coast Ranges. That the other small­bodied Hydromys ( H. hussoni ) and most other North Coast Range endemic mammals are restricted to higherelevation forests suggests to me that H. ziegleri may be an inhabitant of montane forests, and that thus far it may be known only from the very lowest altitudinal bound of its range. Future trapping regimes targeted at streams along elevational gradients in the Prince Alexander and eastern Torricelli Mountains should serve to illuminate its real geographic and altitudinal distribution. The ongoing discovery and description of new mammal species from throughout mainland New Guinea (e.g. Flannery and Groves 1998 ; Helgen and Flannery 2004 ; Helgen 2005 ), highlights how poorly­studied the biota of many areas of the island remains. Amphibious murines, which can be difficult to trap, are often especially overlooked in rapid faunal inventories. I strongly suspect that future survey efforts will identify additional small­bodied amphibious murines in outlying areas of montane New Guinea—per­haps, for example, in the mountain ranges of the Vogelkop Peninsula (the Arfaks and Tamraus), in the expansive Foya Range, or on other large mountainous islands adjacent to New Guinea , such as New Britain or Seram.