home1.gif (2214 bytes)

Murid Rodents of the Solomon
Islands Chain

Tim Flannery, Australian Museum and Colin Groves, Australian National University

INTRODUCTION

The existence of endemic murids on the Solomon Islands was made known as early as 1883, when Ramsay described Mus salamonis from Florida Island (originally stated to be from Ugi Island, off San Christobal, but corrected on an Erratum slip in the volume). To this Thomas in 1888 added Mus rex and Mus imperator, both from Guadalcanal, in 1902 Uromys sapientis from Santa Ysabel, and in 1904 Uromys porculus from Guadalcanal. In 1910 he erected the new genus Cyromys for his Mus imperator and M.rex, and in 1922 Solomys for Uromys sapientis.

Geographically part of the Solomons chain of islands, but politically part of Papua New Guinea, is Bougainville; the three endemic rodents of this island were described by Troughton in 1936 - Melomys bougainville, Solomys salebrosus and Unicomys (new genus) ponceleti.

Only Ruemmler (1938), Tate (1951) and Laurie & Hill (1954) have published revisions of the Solomons endemics since their original descriptions. Each of these three revisions is unique: Ruemmler and Tate both seemed to feel that too many species had been described, and proposed different ways of reducing the list, and only Laurie & Hill recognised all as valid, at least in some sense. The fate of the described genera has been almost as diverse. Of the eight species, only four were recognised by all three revisions, namely salamonis, imperator, sapientis and ponceleti; the first two were placed in genus Uromys by all three, the other two in Melomys (Solomys) by Ruemmler, in Uromys by Tate and in Solomys by Laurie & Hill; the species ponceleti was separated at subgeneric level (as Unicomys) from its congeners by both Tate and Laurie & Hill. Solomys salebrosus was considered a synonym of sapientis by both Ruemmler and Tate, but a valid species by Laurie & Hill. Mus (=Uromys=Cyromys) rex was synonymised with salamonis by Ruemmler, but with sapientis by Tate, and recognised as a valid species by Laurie & Hill. Uromys porculus was synonymised with salamonis by Tate; Ruemmler and Laurie & Hill recognised it as a valid species but referred it to Melomys. Melomys bougainville was synonymised with M.rufescens rufescens by Ruemmler, but recognised as a valid subspecies of M.rufescens by Tate and by Laurie & Hill.

We feel that there is something to be learned from the shambles left by Ruemmler and Tate: study the types before leaping to conclusions. Ruemmler had seen only those in the British Museum (i.e., those described by Thomas); Tate had seen them all, but very briefly, and had only some American Museum specimens from Choiseul island to study at his leisure (these latter, not having seen them ourselves, we will not comment on). Only Laurie & Hill would appear to have taken the attitude that, not having studied specimens of the taxa described by Troughton, they would pay him the courtesy of assuming that he knew what he was about.

MATERIAL SEEN

Our studies have encompassed both the British Museum and the Australian Museum specimens. We have therefore seen all the types, and the (somewhat inadequate!) additional material, and are convinced that all described taxa are valid - almost certainly at specific level. Their generic allocation also turns out to be fairly simple, once they have all been compared.

We have also been studied subfossil material from Buka excavated by Dr and Mrs M.Spriggs (of the Australian National University) and Dr J.Specht (of the Australian Museum). This material turns out to contain examples of the extant Bougainville species plus two others which are presumably extinct; one of us (TFF) is currently describing these new species, and they will be briefly mentioned below after the discussion of the living ones.

RESULTS

Leaving aside Melomys bougainville, which we will not treat in detail, the endemics fall very clearly into two groups. To the first group belong the species imperator, rex and porculus; to the second belong ponceleti, salamonis, sapientis and salebrosus. For the first, the genus-level name Cyromys Thomas, 1910 (type, imperator) is available; for the second the available names are Solomys Thomas, 1922 (type, sapientis) and Unicomys Troughton, 1935 (type, ponceleti). The differences are as follows:

Cyromys. The palate is long, its posterior margin behind the level of the third molars and widens from front to back. As in most murids, the palate is less than 1mm thick. The palate ridges (where complete preserved specimens are available) are numerous and complex. The molar pattern is somewhat simplified, into three little-elaborated ridges. There is a protrusion anterior to the fronto-temporal sutures, on the dorsal margin of the temporal fossa. The choana is broad as in most murids. The basioccipital is long and relatively narrow. The ascending ramus of the mandible is subvertical, the condyles large, and there is a prominent masseteric knob and anterior ridge. The bullae are not enlarged, and have a distinct anterior process, with a small antebullar foramen. The incisive foramina are short, under 5mm long. The toothrow is relatively short, about 16 to (unusually) 21.5% of condyobasal length in adults.

Solomys (including Unicomys). The palate is (primitively) short, ending anterior to the third molars; its breadth is even from front to back, and it is remarkably thick, about 1.0 to 1.5 mm in the three smaller species and over 2mm in S.ponceleti. The palate ridges are few and simple, as is usual in murids. The molar pattern is complex, as in Melomys, except in the recently described S.spriggsorum. There is no protrusion in the vicinity of the fronto-temporal suture. The choana is extremely narrow compared to most murids. The basioccipital is short and broad. The ascending ramus is more backwardly sloping, especially the coronoid process; the condyle is restricted in extent; the masseteric knob and ridge are small. The bullae are considerably enlarged, their anterior processes lost in the expanded bulla wall; the antebullar foramen is very big. The incisive foramina are elongated, 5.4 to 7.5mm long. The toothrows are enlarged, 19.6 to 22.4% of condylobasal length in adults.

Some of the character states listed under Cyromys are primitive retentions (mainly the thin palate, small bulla and small teeth); the derived conditions are similar to those of Uromys, the giant rats of northeastern Queensland, New Guinea and New Britain. We have recently (Flannery & Groves, in press) proposed to rank Cyromys as a subgenus of Uromys, more plesiomorphic in most features (such as the less extreme molar simplification or palate ridge elaboration) than nominotypical U.(Uromys), but with a few derived states such as the outbowed incisive foramina, posteriorly broadened nasals, and opisthodont incisors. Of the three species, U.(C.) porculus, the smallest, is the most plesiomorphic, while U.(C.) rex is highly specialised for an arboreal habitat.

The derived conditions of Solomys - especially the thickened palate and enlarged bullae - are unlike any closely related murid; some species of Melomys show an approach to the same palatal thickening, and it is possible that it is here that the origin of Solomys should be sought. The simplified molar pattern of S.spriggsorum, however, recalls Uromys; this is another possible derivation for the genus. But in general the overall features of Solomys are so unusual and idiosyncratic that we have no option but to recommend full generic status.

The four species of Solomys may be briefly characterised as follows:

S.sapientis (Thomas, 1902). From Isabel Island (Santa Ysabel). Condylobasal length 51.7-54.9mm (n=4). Snout breadth (at maxillary-premaxillary sutures) ) 47.8-50.0% of its length (orbit to prosthion). Braincase relatively high, rounded: basisphenoid-parieta height greater than palate-nasion height. Teeth relatively small, 19.6-21.2% of condylobasal length. Upper incisors narrow. Supraorbital ridges evenly curved, laterally concave. Zygomatic arches not anteriorly expanded. Bullar mouth not tubular. Hardly any lateral spine on periotic. The pelage is grey-brown, with underparts brownish buff; hands and feet grey-white; tails with 8-9 scales per cm of length.

S.salamonis (Ramsay, 1883). Florida Islands (not, as originally stated by the describer, Ugi Island); Tate (1936) referred specimens from Choiseul to this species, but not having seen them ourselves we cannot confirm this identification: especially as he considered S.sapientis a synonym of salamonis. Known from only a single specimen. Slightly smaller in size than S.sapientis, condylobasal length in the type 49.1mm. Snout breadth 50% of snout length. Toothrow 21.8% of condylobasal length. Supraorbital ridges nearly straight. Zygomatic arch contracted anteriorly. A large lateral periotic spine. Coronoid process of mandible curiously thin, spiky. Other characters (braincase shape, incisors, bulla mouth) as in S.sapientis.

S.salebrosus Troughton, 1936. Bougainville and Buka. Size as in S.sapientis: condylobasal length 51.5-53.0mm (n=4). Snout narrower, its breadth 40.4-48.6% of its length. Braincase flattened: posterior skull height about equal to anterior. Teeth larger: toothrow length 20.8-22.2% of condylobasal length in adult. Upper incisors broader. Supraorbital ridges nearly straight. Zygomatic arch slightly expanded anteriorly, intermediate between the two previous species. Bullar mouth tubular. Lateral spine of periotic distinct.

Solomys ponceleti (Troughton, 1936). Bougainville and Buka. Very large, condylobasal length 62.4-67.5mm (n=4). Snout narrow, its breadth 44.0-45.6% of its length. Braincase considerably flattened: posterior skull height less than anterior. Teeth large: toothrow length 20.7-21.9% of condylobasal length. Incisors narrow and labiolingually expanded, this diameter 90.4-92.0% of maxillary toothrow length (cf.83.5-89.3 in the other three species). Supraorbital ridges nearly straight, but constricted at back of interorbital region. Zygomatic arch only slightly expanded anteriorly. Bullar mouth tubular. Hardly any lateral spine to periotic. Colour dark brown, the hairs long, coarse and sparse, raised into a crest along midline of neck.

One species of Solomys stands apart from the rest. This is the enormous and grotesque S.ponceleti. While it is phenetically removed from the other three, we do not at this time recommend recognition of the subgenus Unicomys, because of uncertainty whether one of the other three species might not in fact share some derived features with it. For example the other Bougainville species, S.salebrosus, has the same snout narrowing and tubular bullar mouth, and an approach to the flattened braincase.

DISCUSSION

If the Solomons chain has an endemic murid genus and an endemic subgenus of Uromys, not mention an endemic species of Melomys, how did these spread there? And when?

Simpsonian sweepstakes methods are, of course, a possibility, but seem unsatisfactory - a last resort, if no convincing alternative suggests itself. On the other hand, the murids are the only indigenous non-volant land mammals on the island chain, so a sunken land-bridge would raise the problem of why no other mammals made it across.

Geologically, the Solomons chain forms three provinces: a Pacific province (including Malaita), a Central province (including southeastern Bougainville, Choiseul, Isabel, part of Guadalcanal, and San Christobal), and a Volcanic province (including Buka, most of Bougainville, New Georgia, and at least the northwestern tip of Guadalcanal) (Hackman, 1974). The Volcanic province would represent, with Vanuatu, the remains of the volcanic inner arc near the edge of the Indo-Australian plate, while the Central province, plus New Ireland, is the outer non-volcanic arc on the plate margin itself. These two arcs were originally parallel, but were deformed by the arrival of the Ontong Java Plateau, probably in the Miocene, which collided with the plate margin around New Britain as a fixed point; after which a new plate margin appeared, incorporating part of the old one in the Solomons region (Ravenne et al., 1982).

The earliest terrestrial deposits in the region are probably Lower Miocene, and are in the Central province - whose components (exclusive of Guadalcanal) were, indeed, a continuous land-mass at times of low sea level (called Greater Bukida by Diamond (19 )). Uplift has continued ever since, being extremely rapid in some places: on Guadalcanal, for example, Pleistocene surfaces of marine erosion are now 800 metres a.s.l. according to Hackman (1974).

We propose the following scenario: in the late (?) Miocene, primitive mosaic-tailed murids spread along the outer edge of the plate margin (which, for all we know, might have been a single land-mass at that time), having reached it via New Britain, presumably crossing what must have been very narrow channels at either end of that island. These became the genus Solomys. After the break-up of the old plate margin by the collision of it with the Ontong Java Plateau, different species began to differentiate on the resulting islands, being from time to time brought back into contact by falls in sea level. The volcanic inner arc of the new subduction zone, like all such arcs highly unstable, saw the rise and fall of islands all along it; Uromys is present even today on New Britain, and if a hypothetic primitive Uromys was present in, perhaps, the Pliocene, shifting volcanics could have isolated a population and carried it sequentially southeast as far as Guadalcanal.

Attributing Solomys to the outer arc, Cyromys to the inner, would account for their curious distribution: the former spread all along the Pacific-ward side of the chain, the latter confined to Guadalcanal.

As for speciation processes, clearly isolation (intermittent though it seems to have been) accounts for the divergence of the three small species of Solomys. Not, however, for S.ponceleti: as we have seen, its closest relative is plausibly the sympatric S.salebrosus. Bougainville is a sizeable island, and some geographic isolating mechanism cannot be ruled out, especially if we bear in mind that the nearby island of Buka also has both species. This is hardly the case, however, for Guadalcanal: the coexistence of the three species of Cyromys there cannot readily be accounted for by any allopatric mechanism. The ecological separation of the (highly autapomorphic) arboreal Uromys (Cyromys) rex from the (more plesiomorphic) terrestrial U.(C.) imperator is exactly the sort of set-up envisaged by the sympatric speciation model of Maynard Smith (1966). The ecological requirements of the still more plesiomorphic U.(C.) porculus are unknown. Chromosome studies are awaited for all these taxa.

CONCLUSIONS

The Solomon Islands chain has (at least?) eight species of murid rodents, belonging to three genera (all members of the "mosaic-tailed" group). One species belongs to the widespread genus Melomys; three to an endemic subgenus of Uromys; four to an endemic genus, Solomys. The four Solomys species are distributed along the island chain that is the remnant of the former outer (non-volcanic) arc at the northeastern margin of the former Indo-Australian plate; the three Uromys species are sympatric on the formerly volcanic and inferentially younger island of Guadalcanal, and cannot plausibly have arisen by any mechanism other than sympatric speciation.

REFERENCES CITED

Diamond, J.M.

Hackman, B.D. 1974. The Solomon Islands fractured arc. pp.179-191 in P.J.Coleman, ed., The Western Pacific: Island Arcs, Marginal Seas, Geochemistry. University of Western Australia Press, xviii+675pp.

Maynard Smith, J. 1966. Sympatric speciation. Amer.Nat., 100:637-50.

Ramsay, E.P. 1882. On a new species of Mus from the island of Ugi, Salomon Group. Proc.Linn.Soc.N.S.W., 7:43-44.

Ravenne, C., C.E.de Broin & F.Aubertin. 1982. Structure et histoire de la region Salomon-Nouvelle-Irlande. pp.327-341 in ORSTOM (Equipe Geologie-Geophysique Noumea), Contribution a l'etude geodynamique du Sud-ouest Pacifique. 649pp.

Ruemmler, H. 1938. Die Systematik und Verbreitung der Muriden Neuguineas. Mitt.Zool.Mus.Berlin, 23:1-297.

Tate, G.H.H. 1951. The Rodents of Australia and New Guinea. Bull.Amer.Mus.nat.Hist., 97:183-430.

Thomas, O. 1888. Diagnoses of six new mammals from the Solomon Islands. Ann.Mag.nat.Hist.(6) 1:155-8.

Thomas, O. 1902. Description of a new Uromys from the Solomon Islands. Ann.Mag.nat.Hist.(7) 9:446-7.

Thomas, O. 1904. On some mammals from British New Guinea presented to the National Museum by Mr.C.A.W.Monckton, with descriptions of other species from the same region. Ann.Mag.nat.Hist.(7) 14:396-403.

Thomas, O. 1910. New genera of Australian Muridae. Ann.Mag.nat.Hist.(8) 6:506-8.

Thomas, O. 1922. A subdivision of the genus Uromys. Ann.Mag.nat.Hist.(9) 9:260-1.

Troughton, E.Le G. 1936. The Mammalian fauna of Bougainville Island, Solomons Group. Rec.Aust.Mus.19:341-353.

home1.gif (2214 bytes)