Phalanger mimicus (Thomas, 1922)

Phalanger mimicus (Thomas, 1922) View in CoL

Phalanger intercastellanus Thomas, 1895 .

Phalanger orientalis mimicus Thomas, 1922 .

Phalanger microdon Tate and Archbold, 1935 .

Phalanger orientalis peninsulae Tate, 1945 .

HOLOTYPE AND TYPE LOCALITY: BMNH 11.11.11.93, puppet skin with skull and dentaries of an adult male. Collected by G. C. Shortridge, Parimau, Mimika River , Dutch New Guinea (= Irian Jaya), 4°31'S, 136°36'E, 250 ft, 4 October 1910. GoogleMaps

DIAGNOSIS: Skull and teeth small, molars especially so. Lachrymal contribution to the face is short. Zygomatic arches run parallel or diverge from midline of skull rostrally Well­developed postparacrista on M1, with a pronounced buccal kink. Distinguished from P. orientalis and P. intercastellanus by its smaller size, lack of convergence of zygomatic arches rostrally, and by well­developed postparacrista. Distinguished from P. orientalis by absence of white tail tip in females. Distinguished from P. intercastellanus by absence of lingual cingulum on protocone of M2.

.

DESCRIPTION: Critical skull dimensions are listed in table 2. The most notable feature of the skull is the shape of the zygomatic arches. In all other species of Phalanger the arches reach their widest point at the caudal end, near the suture between the jugal and the squamosal, and converge toward the midline of the skull rostrally. In P. mimicus the arches run parallel to the midline, or even diverge rostrally, giving the skull a different shape ( fig. 3 View Fig ). In general the skull shows less development of the supraorbital, parietal, and sagittal crests than is seen in other species of the genus; indeed, it is not unusual for the parietal crests to remain unfused, even in very old individuals. The facial extent of the lachrymal is comparatively small, as in P. intercastellanus . The frontals are deeply penetrated by the nasal bones; the nasals do not taper abruptly at their caudal extremity, but are rounded ( fig. 4 View Fig ). The teeth are remarkably small, even given the overall small size of the animals. In this respect P. mimicus is similar to the rare montane Telefomin cuscus, P. matanim (Flannery, 1987) . The welldeveloped postparacrista on M1, with its pronounced buccal kink ( fig. 5 View Fig ), is seen in several species of Phalanger , but not in either P. orientalis or P. intercastellanus , the two species within which P. mimicus was formerly reduced to synonymy.

The tail wedge of P. mimicus is short, as in P. intercastellanus . Adult males of the Australian population have upper parts that are gray, with a well­defined dark dorsal stripe finishing well short of the haunches, and silver­tipped guard hairs. Three specimens, including the holotypes of P. orientalis mimicus (BMNH 11.11.11.93) and P. orientalis peninsulae (AMNH 108905), have an off­white pelage with dark guard hairs. This coloration is often seen in old males of both P. orientalis and P. intercastellanus . With the exception of aged specimens, however, P. mimicus appears to show little or no sex­based color dimorphism, at least in its Australian populations. The situation is less clear for the New Guinea population, because only one adult male, the aforementioned BMNH 11.11.11.93, is known. Females of the New Guinea population have a striking reddish­brown pelage, which was one of the characteristics used by Tate to describe P. microdon . Juvenile males also show this coloration, but given the general tendency toward a reddish pelage in juveniles of both P. orientalis and P. intercastellanus , this may be of limited significance in distinguishing between the species. It is worth noting that, unlike the condition seen in Australian P. mimicus , the dorsal stripe continues almost to the base of the tail, albeit becoming more diffuse caudally. The question of whether the differences in pelage between the Trans­Fly and Cape York populations warrant taxonomic recognition cannot be answered without a larger sample from southwest New Guinea.

HABITAT: Rand and Brass (1940) provide descriptions of the southern New Guinea region that contains the collecting localities for P. mimicus . In the area along the coast, south of the estuary of the Fly River (which includes the Oriomo River: fig. 1 View Fig , locality 62), the predominant vegetation type is scrub savanna. Rainforest occurs chiefly as fringing strips of no more than 100–200 m in depth along creeks and streams, and in isolated patches on ridge tops ( fig. 6 View Fig ). The rainforest is described by Rand and Brass as ‘‘poor and light’’, with some areas that ‘‘could only be described as dense brush’’. Savanna trees, including Acacia and Tristania , intrude into the rainforest. The savannas are dominated by Melaleuca ; on ridges this genus mixes with Tristania, Eucaplyptus , and Acacia to form savanna forests with a thin, high canopy and a ground cover of grasses. During the rainy season the flat savannas become boggy, and can flood up to a foot (Brass and Rand, 1940). In contrast, the areas around the Palmer River Camp ( fig. 1 View Fig , locality 63) and the Black River Junction ( fig. 1 View Fig , locality 64) in the northern part of Western Province are typified by steep ridges rising up to 100 m above sea level (Brass and Rand, 1940). The vegetation type is classic lowland rainforest rich in epiphytes in both the lower layers and the canopy ( fig. 7 View Fig ).

The Australian localities for P. mimicus are described by Brass (1953), and bear striking similarities in vegetation to those seen in southern New Guinea. At Shepheard’s Battery site, on the Upper Peach River ( fig. 1 View Fig , locality 60), Brass reported that specimens of P. mimicus were ‘‘found in numbers’’ in the rainforest fringing Bonanza Creek. Rainforest was confined to the margins of the creek, while the remainder of the flats and surrounding ridges were covered in open savanna forests of box, bloodwood, and ironbark trees. In the Iron Range ( fig. 1 View Fig , locality 59) lowland rainforests predominated ( fig. 8 View Fig ), though there were still patches of open forest in which rainforest was limited to creek margins and gullies. As with the Fly River basin, there was seasonal inundation of the rainforests on the floodplains, in some areas to depths of 6–10 ft (Brass, 1953).

COMMENTS: A number of authors have recognized the distinctive nature of the lowland gray cuscuses from southern New Guinea (Thomas, 1922; Tate and Archbold, 1935; Menzies and Pernetta, 1986; Flannery et al., 1987; George, 1987). Particularly noteworthy among these earlier studies is that of Menzies and Pernetta (1986), who used a discriminant function analysis of metric characters (largely drawn from the palatal region) to distinguish different populations of P. orientalis sensu lato. Their analysis revealed first, the clear distinction of the southern New Guinea population of P. orientalis from the other New Guinea populations and, second, its affinities with the Cape York population. On this basis, Menzies and Pernetta recognized the two populations as a separate subspecies, P. orientalis mimicus . Our study supports the results of their analysis, the only difference being their retention of mimicus within P. orientalis . Except for its consistently smaller size, Menzies and Pernetta could find no distinctive characters to separate the taxa. In fact, as we have demonstrated, the qualitative differences in cranial morphology between mimicus and the other populations of P. orientalis sensu lato are at least as great as those that define a number of well­established species of phalangerids, and are certainly sufficient to justify separation at the species level.

The systematic placement of P. mimicus is problematic. The characteristics that were originally proposed by Flannery et al. (1987) as the basis for inclusion of the species in Strigocuscus have proved unreliable, being for the most part either age related, homoplastic, or just poorly defined (Springer et al., 1990; Norris, 1992). Despite its original placement within P. orientalis sensu lato, there is little in the way of morphological evidence to link P. mimicus to either P. orientalis or P. intercastellanus , although sexbased color dimorphism in the New Guinea population of P. mimicus may be an indicator that the three taxa form a coherent group. Links with other species of Phalanger are equally tenuous. The small molars of P. mimicus are similar to those of the apparently rare and enigmatic Telefomin cuscus, P. matanim (Flannery, 1987), which also shows deep penetration of the frontals by the nasal bones. However, P. matanim lacks the distinctive zygomatic morphology of P. mimicus , and has long, lax fur and a broad dorsal stripe, more like the pelage texture and pattern of the midmontane P. vestitus . The welldeveloped postparacrista of M1, although absent in both P. orientalis and P. intercastellanus , is seen in such a wide range of phalangerids as to make it virtually useless as a phylogenetic character. In contrast, the distinctive alignment of the zygoma is not seen in any other species of phalangerid, although it does occur in some species of pseudocheirid possums (for example Pseudochirops cupreus and Pseudochirulus canescens ).

One of the distinctive features of the Trans­Fly biota is the heavy presence of Australian elements in the fauna (Flannery, 1995). (These are listed in table 3.) The underlying biogeographic reasons for this relate to the tectonic history of New Guinea. In contrast to the central cordillera and Huon Peninsula, which were formed as an arc of volcanic islands upthrust by the northward drift of the Austral Plate, the southern plains have always formed part of the Austral Plate. At many times in the past 2 million years, most recently during the late Pleistocene (ca. 20,000 years ago), the Trans­Fly has been continuous with Australia (Flannery, 1995). From this point of view, the existence of a distinctive species of Phalanger spanning the Trans­Fly plains and Cape York is unsurprising. Nonetheless, the distribution of P. mimicus raises some interesting questions of zoogeography, which are worth considering in more detail.

Based on the distribution of mammals, Flannery (1995) recognizes three zoogeographic provinces in New Guinea, which correspond closely to the island’s underlying geological structure and history. The Oceanic Province includes the Vogelkop Peninsula; Weyland Mountains; Japen Island; and the Cyclops, Bewani, and Torricelli Ranges. The Austral Province is defined by Flannery as the Trans­Fly plains, the Moresby Region, and the coastal grasslands that run southeast from there toward Milne Bay, and then along the northern coast as far west as Popondetta. Finally, there is the Tumbunan Province, an entity originally proposed by Schodde and Calaby (1972), and encompassing rainforest and alpine grassland. The boundary between the Austral and Tumbunan Provinces ‘‘can be conveniently regarded as the interface between woodland with gallery rainforest and more dense rainforest’’ (Flannery, 1995). None of these provinces are continuous geographically; Flannery divides the Tumbunan Province into four discrete subprovinces (Eastern, Western, Central, and Huon; Flannery, 1995), and a similar case could be made for dividing the Austral Province fragments into Trans­Fly, Northeastern, and Southeastern subprovinces.

Early biogeographic models for New Guinea’s mammal fauna (for example, Ziegler, 1977) assumed that the island was the center of diversity for rainforest­adapted marsupial species, which went on to establish limited outlying populations in Australia, most notably on Cape York. Subsequent discoveries in the fossil record of Australia, particularly in the Oligo­Miocene deposits at Riversleigh (Archer et al., 1991), revealed that a diverse fauna of rainforest mammals was present in Australia for much of the Tertiary, thus generating a competing hypothesis for the center of species diversity. In this model, New Guinea’s marsupial fauna is seen as a relict distribution of the formerly widespread Australian rainforest marsupials, with New Guinea acting as a refugium from the increasing aridity of Australia during the Pleistocene and Holocene (Archer et al., 1991). The fauna of the Austral Province was seen as containing more recent arrivals, better adapted to the diverse habitat of open woodland, gallery rainforest, savanna grasslands, and swamps found in the Trans­Fly and in coastal areas of southeast New Guinea.

At first sight, P. mimicus would appear to fit conveniently into the group of Austral Province species confined to the Trans­Fly subprovince of New Guinea and to northern Australia ( table 3), an inhabitant of patchy rainforest surrounded by open woodlands and savanna. However, it is also found in lowland rainforest areas (i.e., the Palmer River, Black River, and Mount Bosavi localities), more typical of the Tumbunan Province. That being the case, what defines the boundary between the ranges of P. mimicus and P. intercastellanus? Both species are found in lowland rainforest, and P. intercastellanus has been reported in wooded savanna in the Port Moresby area (Menzies and Pernetta, 1986). Clearly, habitat specialization alone cannot explain the distribution of the two species.

Discussing the boundary between P. orientalis and P. intercastellanus in the Markham­Ramu area, Colgan et al. (1993) noted that a number of other lowland mammal species and subspecies pairs abut in this vicinity, and postulated an underlying zoogeographic barrier. Similar examples exist for the P. mimicus / P. intercastellanus boundary; these include the closely related Melomys species M. levipes (SE New Guinea) and M. lorentzii (Trans­Fly) ; two subspecies of Rattus leucopus , R. l. dobodurae (SE New Guinea) and R. l. ringens (Trans­Fly); two subspecies of the gray dorcopsis, Dorcopsis luctosa luctosa (Trans­Fly) and D. l. beccarii (SE New Guinea); and two subspecies of spotted cuscus, Spilocuscus maculatus chrysorrhous (SW New Guinea) and S. m. goldei (SE New Guinea) (Taylor et al. 1982; Flannery, 1994, 1995; Menzies, 1996). All these pairings involve species that occur in lowland rainforest, and cannot be explained by fragmentation of the savanna grasslands and open woodlands of the Austral Province. Like the boundary between P. orientalis and P. intercastellanus , it is possible that the range of P. mimicus is limited at its eastern extent by some as yet undiscovered zoogeographic barrier. Unfortunately, and in contrast to the Markham­Ramu area, the mammal fauna of the western portion of Papua New Guinea’s Gulf Province (the likely zone of contact between P. mimicus and P. intercastellanus ) has not been extensively sampled (Flannery, 1995, Map 11).