AMERICAN MUSEUM Novttates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET NEW YORK, N.Y. 10024 U.S.A. NUMBER 2690 NOVEMBER 2, 1979

KARL F. KOOPMAN Zoogeography of from Islands Off the Northeastern Coast of New Guinea

AMERICAN MUSEUM Novta,tes PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2690, pp. 1-17, fig. 1, tables 1-3 November 2, 1979

Zoogeography of Mammals from Islands Off the Northeastern Coast of New Guinea

KARL F. KOOPMAN1

ABSTRACT The faunas of seven islands lying off the made to explain the presence or absence of various northeast coast of New Guinea and adjacent to the Megachiroptera on these and other islands in the Bismarck Archipelago have been studied. While a Bismarck area in terms of "incidence functions," but number of characteristic New Guinea elements are it is evident that the data are inadequate to go very lacking, the faunas show a mixture of New Guinea far in this direction. and New Britain (Bismarck) elements. An attempt is INTRODUCTION The American Museum of Natural History of the University of California at Los Angeles, received two rather large collections of mam- and I thank him for donating them to the mals collected by Jared Diamond from several American Museum of Natural History and for islands off the northern coast of Papua New critically reading the manuscript. The late Mr. Guinea (Karkar, Bagabag, Crown, Long, Hobart Van Deusen initially identified much of Tolokiwa, Umboi, Sakar) which has focused the marsupial material and as usual was ex- my attention on the zoogeography of the tremely encouraging in the early stages of the Bismarck Archipelago in relation to the main- investigation. Dr. Guy Musser of the Depart- land of New Guinea. The above-named seven ment of Mammalogy identified the rodents and islands (fig. 1) are all off the New Guinea shelf critically read the manuscript; both he and Dr. and the quite separate New Britain shelf. They Mary Taylor of the University of British Co- therefore form a link between New Guinea lumbia were very helpful in discussing the dis- (which is actually an extension of the Aus- tribution of Rattus in and around the tralian shelf) with a rich continental fauna, and Bismarcks. Mr. John E. Hill of the British the Bismarck Archipelago with an insular Museum (Natural History), Drs. Hans fauna. Hackethal and Renate Angermann of the Zoo- logisches Museum der Humboldt Universitat ACKNOWLEDGMENTS (Berlin), and Dr. Patricia Freeman of the Field The collections on which this report is Museum of Natural History (Chicago) made chiefly based were made by Dr. Jared Diamond collections available and extended numerous

'Curator, Department of Mammalogy, American Museum of Natural History.

Copyright ©) American Museum of Natural History 1979 ISSN 0003-0082 / Price $1.50 2 AMERICAN MUSEUM NOVITATES NO. 2690 courtesies. Dr. Francois Vuilleumier of the De- imens of rodents. Most of these specimens are partment of Ornithology, American Museum of represented by skulls only, some by skins, but Natural History, also read the manuscript and only a few by matched skins and skulls. In the made many helpful comments, most of which I following sections, I will mention all have followed. known from New Britain, but will particularly emphasize species known from the seven small SYSTEMATIC SECTION islands, whether from Diamond's or from ear- lier collections. Nomenclature follows Laurie Of the six orders of mammals native to the and Hill (1954) unless otherwise indicated. Dis- continental Australasian fauna (including New tributions of Megachiroptera are given in table Guinea), the Monotremata (see Van Deusen 2 and those of other species in table 1. and George, 1969, for distribution) and Mar- supicamivora do not extend to the northeast beyond the New Guinea mainland. The other ORDER PERAMELINA four orders (Peramelina, Diprotodonta, Chirop- Only one species of bandicoot is known tera, Rodentia) all occur in the Bismarcks and, from the Bismarcks (Laurie and Hill, 1954), with the exception of the Peramelina, go on to Echymipera kalubu (at least from New Britain the Solomons (Laurie and Hill, 1954). All four and New Ireland). Interestingly enough, speci- of these orders are represented by Jared Dia- mens of Echymipera in the American Museum mond's specimens from the seven small islands of Natural History from the D'Entrecasteaux listed above, though he obtained very few spec- and Trobriand islands off the eastern end of

TABLE 1 Distribution of mammals (other than Megachiroptera) on New Guinea, New Britain, and the Seven Small Islands off the Northeastern Coast of New Guinea (All New Britain species listed but for New Guinea only those that occur on one or another of the other islands listed.)

New New Guinea Karkar Bagabag Crown Long Tolokiwa Umboi Sakar Britain Echymipera kalubu + + + + + + + Phalanger orientalis + + + + + + Petaurus breviceps + + + Dendrolagus matschei + - + Thylogale bruijni + + + + Emballonura nigrescens + - + + Rhinolophus megaphyllus + + Rhinolophus euryotis + + Hipposideros cupidus + - + Hipposideros galeritus + + + + Hipposideros diadema + + Aselliscus tricuspidatus + - + Pipistrellus tenuis + - - - + Miniopterus tristis + + Murina florium + + Kerivoula myrella - + + Pogonomys macrourus + + Rattus exulans + + + + Rattus praetor + + + Melomys rufescens + - + Uromys neobritannicus - - + Hydromys neobritannicus + 1979 KOOPMAN: ZOOGEOGRAPHY OF MAMMALS 3

New Guinea are not kalubu, but another spe- to by local residents. This is a very widespread cies, E. rufescens. Both species occur on the species ranging from Australia, Timor, and the New Guinea mainland. Moluccas to the Solomons (where it is the only Echymipera kalubu-Diamond obtained marsupial). specimens from Karkar (32 sets of lower jaws), Petaurus breviceps-This species was ob- Bagabag (three skins and skulls, one skin only, tained by Diamond only from Bagabag (three 21 sets of lower jaws), Tolokiwa (one skin skins and skulls). This is another widespread only), Umboi (one skin and skull, four skulls species that ranges from Australia to New Brit- only), and Sakar (one set of lower jaws). The ain, but does not reach the Solomons. species is probably widespread throughout the Dendrolagus matschiei-Diamond obtained New Guinea-Bismarck area. this species only on Umboi (one skin and skull, one skull only). The late Hobart Van Deusen ORDER DIPROTODONTA doubted that tree kangaroos were native to Um- boi but thought that they had been introduced, Three species of diprotodont marsupials are known from New Britain (Phalanger orientalis, since they are often kept captive in New Guinea villages. Their absence from any is- Petaurus breviceps, Thylogale bruijni), though lands off the northern coast of New Guinea only Phalanger orientalis goes on to the Sol- except Umboi would support this possibility. omons (Laurie and Hill, 1954). In addition, P. maculatus is known from the Admiralty and Thylogale bruijni-Diamond obtained speci- Ninigo islands, but apparently not from the mens from Bagabag (one lower jaw) and Um- main Bismarck Islands. Diamond also obtained boi (one skin and skull, two skulls only, five mandibles). This is the only wallaby in the a species of tree kangaroo (Dendrolagus matschiei) from Umboi, though this species is Bismarcks and even it does not reach the Sol- not known from nearby New Britain. omons. Phalanger orientalis-Diamond obtained specimens from Karkar (one skull, 363 sets of ORDER CHIROPTERA lower jaws), Bagabag (one skin and skull, one Of the five families of known from skin only, three skulls only, 152 sets of lower New Guinea, only the Molossidae have not jaws), Long (10 skins only, 14 skulls only, one been recorded from the Bismarcks. Even that set of mandibles), and Umboi (one skin and family may occur there since Tadarida skull, three skins only, five skulls only, one (Chaerephon) jobensis is known from both mandible). The American Museum of Natural New Guinea and the Solomons (Hill, 1961). History also has eight skins and skulls, two Confirmation of its occurrence in the Solomons skulls only from Long, and one skin and skull is provided by two specimens in the American from Umboi all collected by W.F. Coultas on Museum of Natural History from Choiseul. the Whitney Expedition in 1933. Two sub- (Tadarida jobensis solomonis was previously species are currently recognized in the northeast known only from Ysabel.) There are, however, New Guinea-Bismarck area (see Laurie and no records of molossids on any islands off the Hill, 1954), P. o. vulpecula on the New Guinea northeast coast of New Guinea. mainland and P. o. ducatoris on New Britain. I have been unable to distinguish skulls of these FAMILY PTEROPODIDAE two subspecies, but find that skins of vulpecula tend to be redder with the dorsal stripe better Eight genera of pteropodids are definitely defined than in ducatoris. On the basis of this known from the mainland of New Guinea. Of color difference, I identify specimens from these, only two, the newly described Aproteles Karkar and Bagabag as P. o. vulpecula and (Menzies, 1977) and , both those from Long and Umboi as P. o. ducatoris. monotypic, seem to be confined to the main- There is, however, a great deal of individual land. The other six (, , Dob- variation. According to Diamond (in litt.), the sonia, , , ) Phalanger of Long was introduced from are all known from the Bismarcks, and with the Tolokiwa in the nineteenth century as attested exception of Syconycteris, go on at least as far 4 AMERICAN MUSEUM NOVITATES NO. 2690 as the Solomons. Besides these genera, Melo- the remaining five (hypomelanus and neohiber- nycteris (including Nesonycteris, see Phillips, nicus) are known from New Britain, the former 1966) occurs widely in the Bismarcks, but ex- on the basis of three immature specimens cept for one old unconfirmed record, has never (AMNH 194302-04) from Kandrian in the west- been collected on the New Guinea mainland. ern part of the island (see Van Deusen, 1969, occurs in the Solomons, but has not p. 2). Two other species have been recorded been found on the Bismarcks or New Guinea. from New Britain, but do not occur on the All the Bismarck Archipelago genera are New Guinea mainland. These are temmincki known from one or another of the islands in and the newly described gailliardi (Van Deu- question off the northeast coast of New Guinea. sen, 1969). Two additional species should also Rousettus amplexicaudatus-This is the only be mentioned. Pteropus admiralitatum was species of Rousettus in the New Guinea- originally described from the Admiralty Is- Bismarck area. Laurie and Hill (1954) recog- lands, but is also known from the Tabar Islands nized a second species, R. stresemanni, but (AMNH 99841) to the northeast of New Ire- after looking over all the Indo-Australian Rou- land. Furthermore, there are specimens from settus in the American Museum of Natural His- the eastern end of New Britain in the Berlin tory, I cannot distinguish these two alleged Museum. Pteropus tonganus, most of whose species on New Guinea and therefore regard range is far out in the Pacific, east of the stresemanni as a subspecies of R. amplexicau- Solomons, occurs also on Karkar. There are datus. (Rousettus celebensis of Celebes and the still other species of Pteropus that occur in the Sanghir Islands is, however, distinct.) Rouset- Solomon Islands. tus amplexicaudatus extends from southeastern Pteropus hypomelanus-As indicated above, Asia through the Malay Archipelago, New this species is known from both New Guinea Guinea, and the Bismarcks to the Solomons. and New Britain. Diamond collected 340 man- Bagabag is the only one of the seven islands dibles from Karkar, 11 mandibles from here discussed from which Rousettus has been Bagabag, a skull from Long, and a mandible obtained (eight skulls collected by Diamond). It from Umboi. The American Museum of Natu- should be mentioned that the Bagabag speci- ral History also has the skin and skull of an mens are clearly referable to the larger R. a. immature from Long collected by Coultas in stresemanni of mainland New Guinea rather 1933. This species appears to be rare on the than to the smaller R. a. brachyotis of the mainland of New Guinea as is indicated by its Bismarcks (including New Britain). absence from any of the extensive mainland Genus Pteropus-This genus has a very wide New Guinea Pteropus collections at the Ameri- range in the Old World tropics from the west- can Museum of Natural History. In the New ern Indian Ocean east at least as far as the Guinea region, as elsewhere in its extensive Cook Islands. There are many species currently Indo-Australian range, it seems to be chiefly a recognized, some dubiously distinct. Seven small island species. Two subspecies are cur- species (hypomelanus, melanopogon, alecto, rently recognized in the New Guinea region, h. conspicillatus, neohibernicus, macrotis, pohlei) luteus, described from Kiriwina Island in the have been recorded from the New Guinea Trobriands to the east of New Guinea, and h. mainland. The occurrence of P. melanopogon vulcanius from Vulcan (=Manam) to the north- in New Guinea is, however, as will be shown west of Karkar. The latter subspecies was dis- below, based on misidentified P. neohiber- tinguished from luteus (Thomas, 1915) solely nicus. Two of the remaining species have lim- on the basis of color, a specimen from Dampier ited ranges distant from northeastern New (=Karkar) being referred to luteus. This char- Guinea. These are alecto, which is known only acter is obviously of no help in allocating the from a very limited area in southwestern Papua Diamond mandibles to subspecies and I there- (Tate 1942, p. 337) and pohlei which is still fore make no attempt to do so. recorded only from Japen Island in Geelvink Pteropus tonganus-As indicated above, the Bay, northwestern New Guinea. Only two of Karkar subspecies (P. t. basiliscus) is far re- 1979 KOOPMAN: ZOOGEOGRAPHY OF MAMMALS 5 moved geographically from the nearest island, ord of P. melanopogon from the mainland of Rennell in the southeastern Solomons, where New Guinea, the species otherwise being con- another subspecies (P. t. geddei) occurs. Dia- fined to the Moluccas, except for a subspecies mond obtained 68 mandibles from Karkar (the from the Aru Islands. I have compared the type only island from which the species had previ- and second specimen of sepikensis with the ously been obtained), but none from any other very limited material of P. neohibernicus avail- island. This supports the idea that in the New able at the Field Museum and have been able Guinea region P. tonganus is restricted to Kar- to compare the second skull of sepikensis with kar. abundant material in the American Museum of Pteropus temmincki-This species is well- Natural History of P. neohibernicus from the known and widely distributed in the Bismarcks, same part of New Guinea from which but has not been recorded from any of the the type was obtained (Sepik and Madang dis- small islands here discussed. It was not ob- tricts). I am unable to find any differences be- tained by Diamond on any of them, but the tween them when tooth wear is taken in to British Museum has a specimen from Ruk account. When compared with limited material (= Umboi). of P. melanopogon from Buru and the Kei Pteropus neohibernicus-This species is com- Islands, the two skulls of sepikensis are not mon and widespread in New Guinea and the only larger but show marked dental differences. Bismarcks. Diamond obtained it from Karkar The upper canines are longer and more slender, (six mandibles), Umboi (28 skulls, 25 man- the larger cheek teeth are about as wide as they dibles) and Sakar (one skull, three mandibles). are long and are practically featureless. On the The British Museum also has a specimen from other hand, P. melanopogon has these teeth Ruk (= Umboi). At present, specimens from much longer than wide, and with well-devel- the mainland of New Guinea are allocated to P. oped cuspidate labial and lingual ridges. I am n. papuanus, whereas those from New Britain therefore inclined to regard both papuanus and are identified as P. n. neohibernicus. Andersen sepikensis as synonyms of P. n. neohibernicus. (1912, pp. 389-90) distinguished these two Genus -This genus is widely dis- forms solely on the basis of color and later in tributed from the Philippines, Celebes, and the the same work (pp. 823-24) strongly suggested Lesser Sundas to Australia and the Solomons. that the color character did not hold and that Two species (moluccensis and minor) occur on papuanus should be synonymized with neo- the mainland of New Guinea. Dobsonia moluc- hibernicus. I have compared skulls of speci- censis also occurs on the Bismarcks along with mens from northeast New Guinea and New another species, D. praedatrix, which does not Britain and am unable to find any consistent occur on the mainland. Bergmans (1975, p. 6) differences. I am therefore allocating the speci- regards the Bismarck representative of D. mo- mens from Karkar, Umboi, and Sakar only to luccensis (anderseni) as a separate species, but, species. The subspecies from the Admiralty Is- as explained below, I do not agree with him. lands, P. n. hilli (Felten, 1961), however, is All three species are found on one or another clearly distinct as shown by the smaller size of of the seven islands discussed in this paper. three adult specimens in the American Museum Dobsonia praedatrix-Although widely dis- of Natural History. Sanborn (1931, p. 14) de- tributed in the Bismarck Archipelago, this spe- scribed a new species, Pteropus sepikensis, cies has not been recorded outside it. Diamond from the Sepik River in northeastern New obtained two skulls from Umboi, however, and Guinea, based on a single skin and skull. (A the British Museum also has an Umboi speci- second skull was later obtained.) Pteropus men. sepikensis was stated to be a member of the Dobsonia moluccensis-This species was ob- melanopogon group and was compared only tained by Diamond from Karkar (one skull with P. melanopogon. Laurie and Hill (1954, only, four mandibles only), Bagabag (15 skulls p. 35) tentatively listed it as a subspecies of P. only, two mandibles only), Tolokiwa (one skull melanopogon. This has remained the only rec- only), Umboi (10 skulls only, 13 mandibles 6 AMERICAN MUSEUM NOVITATES NO. 2690 only), and Sakar (one skull only). The Karkar and Gulf district (Papua). None of the localities specimens are referable to the New Guinea sub- is quite as far east as Bagabag. species, D. m. magna, but those from the other Genus Nyctimene-Four species (albiventer, islands belong to the New Britain form, D. m. draconilla, cyclotis, and aello) are definitely anderseni. Bergmans (1975, p. 6) considered known from the mainland of New Guinea. Nyc- anderseni to be separate species on the basis of timene draconilla was listed as a subspecies of its clearly smaller size, but I believe that some N. albiventer by Laurie and Hill (1954), but intergradation is evident between magna and Greig-Smith (1975) has reported the two forms anderseni among the islands here discussed. sympatrically from the Sepik district of north- (By intergradation, I mean simply that there are eastern New Guinea. Another sympatric occur- intermediate populations between the two taxa, rence is in the Western District of Papua irrespective of how this intermediacy is main- (southeast New Guinea), where, among the nu- tained.) Due to the nature of the material, I merous specimens reported by Tate (1942) as have settled on a modification of condylocanine Nyctimene papuanus (a subspecies of N. albi- length, namely, from the posterior edge of the venter) from the upper Fly River, seven are condyle on one side to the anterior edge of the actually N. draconilla. The type locality of N. upper canine on the same side. Since many major geminus is given as "South of Huon specimens are either immature or badly broken, Gulf, Papua," which would put it on the main- sample sizes unfortunately are small but I think land of New Guinea, but there are no other do show a trend. All specimens measured are mainland records for this species and it is in the American Museum of Natural History doubtful that N. major really occurs there. and all New Guinea mainland specimens are Mayr (1941, p. 168) has indicated that a similar from Morobe and Madang districts, imme- locality for a bird (Manucodia comrii) collected diately opposite the islands in question. The by the same man (Dr. P. Comrie) is an error modified condylocanine measurements (in mm.) for Fergusson Island. The N. major record (the are as follows: New Guinea mainland, seven type locality for N. m. geminus) is probably a specimens (56.6-58.9); Karkar, one specimen similar error for Fergusson Island, where N. m. (55.9); Bagabag, three specimens (48.6-49.4); geminus is known to occur. Of the species Umboi, seven specimens (49.1-53.9); mentioned above, only N. albiventer and N. Bismarcks (New Britain, Tabar Islands, Lihir major are known from the Bismarck Archipel- Islands), five specimens (47.9-50.6). To me ago. Both appear to be widespread there since, this indicates some intergradation on Karkar on the basis of specimens in the British and and Umboi and therefore conspecificity of an- Berlin museums, N. albiventer is known from derseni with Dobsonia moluccensis. New Britain and the Admiralties, N. major Dobsonia minor-Diamond obtained three from New Britain and New Ireland. One addi- skulls from Bagabag. This is the only record of tional species in the New Guinea region (N. D. minor known to me from off the mainland cephalotes) does not occur on the mainland or of New Guinea and is also unusual in being the in the Bismarcks, but does occur on several easternmost record for the species. Andersen small islands off the northern coast of New (1912, pp. 461, 825) recorded it from only two Guinea. localities, both in the western half of New Nyctimene albiventer-Diamond obtained Guinea. Specimens in the American Museum skulls of this species from all of the islands in of Natural History (from the Second and Third question but Karkar; one from Bagabag, 15 Archbold expeditions to New Guinea) are either from Crown, seven from Long, seven from from Indonesian New Guinea (Irian Jaya) or in Tolokiwa, 80 from Umboi, and 14 from Sakar. extreme western Papua (upper Fly River). None Nyctimene albiventer is a wide-ranging com- of the extensive collections from farther east in mon from the northern Moluccas through New Guinea include D. minor. McKean (1972) New Guinea and the Bismarcks to the Sol- has, however, recorded D. minor from both omons. East Sepik district (northeastern New Guinea) Nyctimene cephalotes-This species has a 1979 KOOPMAN: ZOOGEOGRAPHY OF MAMMALS 7 fairly extensive westem range from Celebes since 1912) and these agree very well with the and Timor to the Moluccas, but in the New type of lullulae from Woodlark and also, in Guinea region is known from only two islands, skull measurements, fall within the variation of Numfoor (in Geelvink Bay, western New Diamond's Karkar series. Skulls of the two Guinea) and Umboi (from which the subspecies Woodlark specimens in the American Museum N. c. vizcaccia was described). Aside from the of Natural History, although a trifle larger type specimen of vizcaccia, the only specimen (measurements in mm.) than the Karkar and known to me from the islands in question is Bagabag skulls (condylobasal length, 32.1, one skull collected by Diamond on Umboi. 32.3 vs. 30.2-31.7; width of rostrum just out- This is somewhat smaller than the type (7.9 side molars, 9.6, 9.8 vs. 8.9-9.5) are more like versus 8.7 mm. in maxillary tooth row length) them than to anything else. It is clear, there- but clearly falls outside the variation in con- fore, that populations on Karkar and Bagabag dylobasal length (28.3 vs. 24.8-27.2 mm.) of (to the west of New Britain and Sakar) are the large series of N. albiventer papuanus from much more similar to the population on Wood- Umboi and the rostrum is narrower in relation lark, and perhaps Kiriwina (to the south of to its length. There seems to be no doubt, New Britain and Sakar) than either are to the therefore, that there are two species of Nyc- populations of New Britain and Sakar. It is timene represented in Diamond's collection possible that they can be separated, but I am from Umboi. unable to do so and I therefore allocate the Nyctimene major-Diamond collected two Karkar and Bagabag populations to lullulae. quite distinct forms of N. major (sensu Laurie The presence of this form on islands on the two and Hill, 1954) on the islands in question. sides of New Britain raises the question as to From Sakar, a single skull was collected, whether lullulae should be treated as a separate which agrees very well with N. m. major from species from major. My initial reaction was to the Bismarck Archipelago. On the other hand, do so, since the size difference between the 17 skulls from Karkar and two from Bagabag Karkar and Bagabag specimens on one hand are much smaller than N. m. major. Since N. and the Sakar specimen on the other are ex- m. major is the only subspecies which occurs treme. However, I have decided to retain these at all close to Karkar and Bagabag, I have as subspecies for the following reasons. Firstly, investigated the only other two subspecies of there is no evidence of sympatry except on N. major in the New Guinea-Bismarck region, Kiriwina. However, it is probable that the two namely geminus and lullulae, both probably known specimens from this island are simply confined to islands off the eastern end of New large and small individuals of a single variable Guinea. Fortunately the American Museum of population, which could therefore show inter- Natural History has good series of N. major gradation between geminus and lullulae. (I find from Sudest, Rossel, and Misima in the it very difficult to believe that there could actu- Louisiade Archipelago and from Normanby, ally be two closely related species of Nyctimene Fergusson, and Goodenough in the on such a small low island as Kiriwina.) Sec- D'Entrecasteaux group. These are all referable ondly, the differences between lullulae and to the large N. m. geminus. Two specimens smaller individuals of geminus are not great from Woodlark and one from Kiriwina, and suggest subspecific rather than specific dif- however, are smaller and at least the Woodlark ferences. Finally, since N. major is an un- specimens are referable to N. m. lullulae. usually large species of Nyctimene, it is These allocations are all in agreement with probable that the small size of lullulae is primi- those of Andersen (1912, pp. 709-13, who, tive relative to the larger major and geminus however, called major, geminus, and lullulae and also possible that it formerly occupied the separate species) except that Andersen allocated adjacent mainland of New Guinea and at that the single Kiriwina specimen in the British Mu- time independently colonized Karkar and seum to geminus. The British Museum also has Bagabag to the north and Woodlark to the East. two skins and skulls from Karkar (obtained All three of these islands were almost certainly 8 AMERICAN MUSEUM NOVITATES NO. 2690 colonized overwater since they are separated Britain, totaling 12 specimens (all in the Ameri- from the mainland by water with depths over can Museum of Natural History) show clear 200 meters. My present tentative conclusion is differences in both condylobasal length of skull to call the populations on Karkar, Bagabag, (23.2-25.7 vs. 22.3-23.9 mm.) and braincase Woodlark, and possibly Kiriwina Nyctimene width (10.1-11.7 vs. 9.8-10.8 mm.). Although major lullulae. Obviously more material from there is some overlap, this is rather small when Woodlark, Kiriwina, and the islands east of the two measurements are plotted against each Bagabag would be very useful for determining other. Only six out of 32 specimens from the the true range of lullulae-like populations. mainland fall within the New Britain range, Macroglossus lagochilus-This is the only and five out of 12 New Britain specimens fall species of its genus in the Australian region within the mainland range. Diamond collected and is widely distributed in both New Guinea skulls of Syconycteris from six of the seven and the Bismarcks. Diamond obtained speci- islands under discussion: Bagabag (eight), mens from Karkar (three skulls), Crown (12 Crown (43), Long (97), Tolokiwa (47), Umboi skulls), Tolokiwa (two skulls), Umboi (20 (five), Sakar (seven). I have taken condylobasal skulls), and Sakar (one skull). lengths and braincase widths on usable adult Syconycteris australis-The genus Syconyc- specimens from all these islands and plotted teris has a rather wide range from Australia them in the same way as I have the similar through New Guinea (and on many nearby is- measurements from the New Guinea mainland lands) to the Bismarck Archipelago, also west and New Britain. In the following, the first to the Moluccas. Two species have been recog- range is of condylobasal lengths, the second is nized from the mainland of New Guinea (Tate, of braincase widths, both in mm. Bagabag 1942, pp. 346-47), S. australis (Papua as well (23.3-23.9; 10.6-10.8). Crown (23.5-25.0; as northeastern Australia) and S. crassa (New 10.0-11.2); Long (23.3-25.4; 10.0-11.2); Guinea as a whole as well as the Bismarcks Tolokiwa (23.4-25.1; 10.2-11.4); Umboi (23.9; and most of the other smaller islands). These 10.3-10.8); Sakar (23.8-25.1; 10.5-11.2). With two species were distinguished by Andersen the exception of two braincase width measure- (1912, pp. 775-81) solely on the different pro- ments from Crown Island, all these measure- portions of the cheek teeth. After comparing ments fall within the variation of the New series of Syconycteris from a number of areas Guinea mainland specimens. Since all of the from northern Queensland through southeastern larger series (Crown, Long, Tolokiwa, Sakar) New Guinea to northeastern New Guinea (in- fall mostly outside the range of the New Britain cluding specimens allocated by Tate to both measurements, I am inclined to refer all these species), I find myself in full agreement with small island populations to S. a. papuana (the Hill (in Greig-Smith, 1975, p. 119). There is in New Guinea mainland subspecies) rather than fact a great deal of individual variation in the to S. a. finschi (the New Britain subspecies). proportions of the cheek teeth, but no clearcut There is perhaps some indication of intergrada- dichotomy into broad and narrow toothed tion in that the maximum condylobasal length forms. Since the few Australian specimens of skulls from the Crown-Long group of islands available to me tend to be smaller than those (25.4) is less than that from the New Guinea from New Guinea, I would tentatively retain mainland (25.7) and that from the Tolokiwa- those from the two areas as subspecies, S. a. Umboi-Sakar group is still less (25.1) thus australis of Australia, and S. a. papuana of the showing a slight approach to that of New Brit- New Guinea mainland. McKean (1972, p. 9) ain (23.9). has synonymized the New Britain subspecies S. Meloncyteris melanops-With the exception a. finschi with S. a. papuana, but I am inclined of an old unconfirmed record from the New to retain it since comparison of a large amount Guinea mainland this species has been consid- of material (32 specimens) from northeastern ered endemic to the Bismarck Archipelago. New Guinea (Morobe and Madang districts) Phillips (1966), however, referred two species with two fairly good series from western New from the Solomon Islands to the genus Melo- 1979 KOOPMAN: ZOOGEOGRAPHY OF MAMMALS 9 nycteris. New Britain specimens in both the phyllus and R. euryotis from New Britain and British Museum and the Berlin Museum are R. megaphyllus also from New Ireland. Laurie from the Gazelle peninsula (Keravat and and Hill (1954) do record the sole New Guinea Ralum, respectively) on the northeastern part of species of Aselliscus (A. tricuspidatus) from the the island. The American Museum of Natural Bismarcks. None of these three species are History, however, has specimens from both known from any of the seven islands off the eastern (Wide Bay) and western (Whiteman northeast New Guinea coast. Of the nine spe- range and near Kandrian) New Guinea. Dia- cies of New Guinea Hipposideros recognized mond also obtained M. melanops from Long by Hill (1963) in his revision, five (ater, pa- (three skulls) and Tolokiwa (one skin and pua, muscinus, wollastoni, semoni) are un- skull). known in the Bismarcks. Hill does record H. calcaratus and H. cupidus from Duke of York FAMILY EMBALLONURIDAE Island in the Bismarcks and H. galeritus from New Ireland. He does not record H. diadema Three genera of this family have been re- from the Bismarcks proper but does list it from corded from the New Guinea mainland since I Manus. The American Museum of Natural His- agree with Barghorn (1977) that Saccolaimus is tory has no specimens of H. calcaratus from distinct from Taphozous. However, only Em- off the New Guinea mainland but does have ballonura reaches the Bismarcks or the islands specimens of the other three species from the off the northeast coast of New Guinea, though Bismarcks: H. cupidus (New Britain, Duke of Saccolaimus is known from the Solomons. Al- York, Tabar Islands); H. galeritus (New Brit- though four species of Emballonura are known ain); H. diadema (New Britain; Lihir Islands). from the New Guinea mainland and two of It seems probable, therefore, that all four spe- these plus another species (E. dianae) occur on cies occur on New Britain. Although some or the Solomons, only E. nigrescens is known all of these would be expected on one or an- from either the Bismarcks (except for the other- other of the seven islands off northeastern New wise Solomon Island E. raffrayana cor on the Guinea, only H. galeritus is actually known. Tabar Islands, represented by specimens in the Hipposideros galeritus-Diamond obtained American Museum of Natural History) or the two skulls on Karkar and two on Crown. This islands off the northeast New Guinea coast. is a widely distributed Indo-Australian species Emballonura nigrescens-Laurie and Hill according to Hill (1963), occurring on both the (1954) recorded this species from the Bismarcks New Guinea mainland and the Bismarcks. including New Britain. They also recorded it from Ruk (=Umboi). Diamond did not obtain this species but the American Museum of Natu- FAMILY VESPERTILIONIDAE ral History does have a series from the White- Ten genera of vespertilionids are known man range in western New Britain. Two quite from the New Guinea mainland but only three distinct subspecies are represented in the area, of these are known from the Bismarcks. Chal- New Guinea being occupied by the small E. n. inolobus (see Van Deusen and Koopman, papuana, whereas the larger E. n. solomonis 1971), Philetor (see Hill, 1966, 1971), Nyc- occurs on New Britain. Laurie and Hill (1954) ticeius (see Koopman, 1978), Nyctophilus (in- referred Umboi material to E. n. solomonis.. cluding Lamingtona), and Pharotis are apparently unknown to the northwest of New FAMILY RHINOLOPHIDAE Guinea. Myotis (including Anamygdon, see Of the three New Guinea genera of this Phillips and Birney, 1968) is known from both family, all are known from the Bismarcks. New Guinea and the Solomon Islands, but not Though Laurie and Hill (1954) do not record as yet from the Bismarcks. As discussed be- either New Guinea species of Rhinolophus in low, Murina is known from New Guinea and the Bismarcks, the American Museum of Natu- one of the islands off the northeastern coast but ral History has specimens of both R. mega- has not been recorded from the Bismarcks. 10 AMERICAN MUSEUM NOVITATES NO. 2690

Pipistrellus tenuis (including angulatus and pa- genera are known from the New Guinea main- puanus, see Koopman, 1973) is known from land, including Xenuromys (which Laurie and several islands in the Bismarcks (including a Hill, 1954, inadvertently omitted), but not series in the American Museum of Natural His- Baiyankamys (which Mahoney, 1968, syn- tory from New Britain) and the mainland of onymized with Hydromys). However, only five New Guinea, but has not been recorded from of these (Pogonomys, Rattus, Melomys, any of the seven islands off the northeastern Uromys, Hydromys) are known from the coast. The genus Miniopterus is currently in a Bismarcks. There are also several species of state of taxonomic confusion. Dr. R.L. Peter- uncertain generic allocation (related to Melomys son is revising the genus, but at present, and Uromys) from the Solomons. Pogonomys, though three species have been recorded from Uromys, and Hydromys were all recorded from the general Bismarck area, none are known New Britain by Laurie and Hill (1954). The from the seven islands off the northeast coast American Museum of Natural History has spec- of New Guinea. In fact, only two species of imens of Rattus, Melomys, Uromys, and Hy- vespertilionids are at present known from these dromys from New Britain. Uromys islands. neobritainnicus and Hydromys neobritannicus Murina florium-Only two specimens of are both endemic to New Britain and each is Murina are known from east of the Moluccas the only species of its genus in the Bismarcks. (Van Deusen, 1961). One of these is from the Except for Rattus, none of these genera is New Guinea mainland, the other from Ruk known from the seven islands off the north- (= Umboi). Diamond did not obtain this eastern coast of New Guinea. Taylor, Calaby, spe.ies, which is otherwise known from Lesser and Van Deusen are currently revising the spe- Sundas, the Celebes area, and the Moluccas. cies of New Guinea Rattus, but it is not at Kerivoula myrella-Hill (1965) has revised present clear how many of these there are. the Indo-Australian species of Kerivoula and Only two (exulans and praetor), however, are recognizes the Bismarck species (K. myrella) as known from the small islands which are the distinct from the New Guinea mainland K. subject of this paper. muscina. In fact, each species is considered to Rattus exulans-This widespread species, oc- be more closely related to certain Indo-Malayan curring in both New Guinea and the Bismarcks, ones than to the other species in the New is a well-known commensal of man. It proba- Guinea-Bismarck area. (K. hardwickei and K. bly, therefore, at present occurs on all islands papillosa in the case of myrella and K. white- in this area having human populations. Dia- headi and K. picta in the case of muscina.) Hill mond, however, obtained only one skull from (1965) recorded K. myrella from Manus and Long, two skulls from Tolokiwa, and none Duke of York Islands in the Bismarck Archi- from the other islands. pelago and also from Ruk (= Umboi). The Rattus praetor-This is the species that Lau- American Museum of Natural History has a rie and Hill (1954) called R. ruber, but Calaby series from the Whiteman Range in western and Taylor (in press) have shown that the type New Britain. Diamond did not obtain any spec- of ruber belongs with another species, making imens of Kerivoula. Hill (1965) recognized a praetor the oldest valid name. In any case, R. second genus related to Kerivoula, Phoniscus praetor ranges from New Guinea through the with one New Guinea species (papuensis), but Bismarcks to the Solomons. The American Mu- I am inclined to treat Phoniscus as only a seum of Natural History has specimens from subgenus of Kerivoula. In any case, Phoniscus New Britain and Diamond obtained two mandi- is not known from either the Bismarcks or the bles from Karkar. seven islands off northeastern New Guinea. GENERAL ZOOGEOGRAPHY ORDER RODENTIA The seven islands treated in this paper are All rodents in the Australia-New Guinea re- relatively small and lie more or less between gion belong to the family Muridae. Some 20 two much larger islands, New Guinea and New 1979 KOOPMAN: ZOOGEOGRAPHY OF MAMMALS 11

Britain (fig. 1). As would be expected, some biota of Long Island has certainly colonized the species are shared with one, some with the island during the past 200 years, since the pre- other, and some with both. There are no spe- vious biota had been destroyed by a volcanic cies endemic to these small islands, but there eruption (Diamond, 1974). are a few that have their closest affinities on If we exclude Dendrolagus matschiei, Rattus neither large land mass. In cases where the exulans, and the Long Island record of Pha- same species occurs on both New Guinea and langer orientalis as probable introductions on New Britain, there may be a marked mor- the small islands in question, there are 23 na- phological difference between specimens from tive species of mammals which can be consid- the two islands. In such cases, it is possible to ered at present. In nine of these, (Echymipera determine the affinities of the small island pop- kalubu, Petaurus breviceps, Thylogale bruijni, ulations by identifying the subspecies involved. Pteropus hypomelanus, P. neohibernicus, Nyc- Since these seven islands are not on either timene albiventer, Macroglossus lagochilus, the New Guinea or the New Britain shelf, they Hipposideros galeritus, Rattus praetor) the were almost certainly colonized overwater be- populations on New Guinea vs. New Britain cause they would not have been connected to are either very similar or as yet unanalyzed. either island during low water levels at various Dobsonia minor and Murina florium occur on times in the Pleistocene (see Mayr, 1944, p. 5 New Guinea, but are not known from New for a discussion of this question). The present Britain. The former occurs on Bagabag, the

FIG. 1. Map of the eastern half of New Guinea and the Bismarck Archipelago showing most of the islands mentioned in the text. St. Mathias and Emirau are off the map to the northwest of New Ireland and the Solomons are off the map to the east. 12 AMERICAN MUSEUM NOVITATES NO. 2690 latter on Umboi, and both have presumably lands, being referable to the small D. m. an- dispersed overwater from New Guinea. The derseni. There has been therefore overwater distribution of Murina florium is so poorly dispersal from both sides with evidence that known, however, that its apparent absence some islands have received colonists from both. from New Britain may be an artifact. Pteropus Nyctimene cephalotes-This is another spe- temmincki, Dobsonia praedatrix, cies that does not occur on the mainland of melanops, and Kerivoula myrella occur on New New Guinea or in the Bismarcks but is known Britain but probably not on New Guinea. The from one of the seven small islands in ques- Pteropus, Dobsonia, and Kerivoula do not get tion. The main range of N. cephalotes is in beyond Sakar or Umboi, but Melonycteris oc- Celebes and the southem Moluccas. It is also curs on both Tolokiwa and Long. For the re- known from Timor and Numfoor Island (in maining species a more detailed treatment is Geelvink Bay, northwestern New Guinea). Its necessary. occurrence on Umboi is therefore most surpris- Phalanger orientalis-Of the two subspecies, ing. Actually, in this case, there are two signi- P. o. vulpecula of mainland New Guinea is ficant gaps since it is unknown between Ceram found on Karkar and Bagabag, whereas P. o. and Numfoor and between Numfoor and Um- ducatoris occurs on Umboi. Overwater disper- boi. In this case, a relict distribution is most sal is clearly indicated from both ends. probable. Possibly, the presence of two other Rousettus amplexicaudatus-The New Guinea large and medium-sized species of Nyctimene subspecies, R. a. stresemanni, is found on (cyclotis and aello) has displaced it from west- Bagabag. The quite distinct R. a. brachyotis ern and northern New Guinea. Neither species occurs on New Britain. occurs within its present range. Pteropus tonganus-This is the most puzzling Nyctimene major-This is another species mammal element on these seven islands. with an unusual distribution pattern. Nyctimene Pteropus tonganus has an extensive distribution major almost certainly does not occur on the in the Pacific from the Cook Islands in the east mainland of New Guinea, but the Sakar speci- to Rennell Island in the west. Although it oc- men, as expected, agrees with the large N. m. curs on most islands within this vast area, it is major of New Britain. The specimens from unknown from the Solomon Islands (except Karkar and Bagabag do not, however, but in- Rennell), the Bismarcks, or the New Guinea stead agree with the small N. m. lullulae of the mainland. The occurrence of a population of Trobriands. Again, either long-distance disper- this species on Karkar is completely unex- sal or a relict distribution (occurring formerly pected. Whether this represents long distance on the New Guinea mainland) may be the ex- overwater dispersal or a relict distribution is planation. Another possibility is that lullulae uncertain. At one time, I considered the pos- may represent the primitive condition of N. sibility that the Karkar form (basiliscus) actu- major (a reasonable hypothesis since N. major ally represented the New Guinea mainland P. is one of the two largest species of Nyctimene). conspicillatus rather than the Pacific P. If so, then perhaps the Bismarcks were origi- tonganus. However, J. E. Hill very kindly nally inhabited by a lullulae-like form which compared the type (and only complete speci- subsequently became larger, evolving into N. men) of basiliscus with specimens of both spe- m. major. cies in the British Museum and assured me (in Syconycteris australis-This is another case litt.) that basiliscus is indeed a form of P. in which the small island populations show tonganus. some intergradation between S. a. papuana of Dobsonia moluccensis-Although, as I have New Guinea and S. a. finschi of New Britain. shown above, there is evidence of intergrada- All the island populations, however, are more tion on at least two of the seven intermediate like the New Guinea subspecies, even those islands, only the population on Karkar is refer- from the three islands nearest New Britain. able to the large mainland D. m. magna, those Overwater dispersal from New Guinea has evi- on Bagabag, as well as the three eastern is- dently been more active than the reverse, 1979 KOOPMAN: ZOOGEOGRAPHY OF MAMMALS 13 though the intergradation seen on Long, myrella). It should be emphasized again, Crown, and Bagabag suggests some minor dis- however, that the distribution of Murina in the persal in the opposite direction. New Guinea region is so poorly known that its Emballonura nigrescens-In this species apparent absence from New Britain may not be there is again one subspecies on New Guinea significant. Finally, Sakar has four relevant (E. n. papuana) and another on New Britain taxa of which one (Syconycteris australis pa- (E. n. solomonis). As would be expected from puanus) is of New Guinea origin, whereas the its position near New Britain, the Umboi popu- other three (Pteropus temmincki, Dobsonia mo- lation is referable to E. n. solomonis. luccensis anderseni, and Nyctimene major ma- If we combine the species and subspecies jor) are shared with New Britain. distributions, ignore intergradation, and include All in all, it is clear that dispersal has taken only cases where the taxon occurs on either place from both ends of the small island chain. New Guinea or New Britain but not both, then With the exception of Karkar (at the western six New Guinea taxa have dispersed westward end) all islands with more than one relevant along the island chain (Phalanger orientalis element have been colonized from both ends. vulpecula, Rousettus amplexicaudatus strese- As would be expected, the threee eastern is- manni, Dobsonia moluccensis magna, D. lands have predominantly New Britain elements minor, Syconycteris australis, and Murina flo- (8 out of ten taxa). Even here, however, Syco- rium). Likewise, eight New Britain taxa are nycteris australis papuanus and Murina florium found on one or another of the small interme- have evidently dispersed from New Guinea to diate islands (Phalanger orientalis ducatoris, these islands but have not reached New Britain. Pteropus temmincki, Dobsonia praedatrix, D. moluccensis anderseni, Nyctimene m. major, "INCIDENCE FUNCTIONS" Melonycteris melonaps, Emballonura nigrescens solomonis, and Kerivoula myrella). Diamond (1975) has used this term to distin- Of the two species on Karkar, which are rele- guish various patterns of distribution based on vant to this problem, both (Phalanger orientalis numbers of small islands occupied by various vulpecula, Dobsonia moluccensis magna) are species. He distinguishes a series of arbitrarily New Guinea elements. Bagabag, on the other defined classes going from the "high-S" spe- hand, has five relevant elements of which four cies, which occur only on the largest and spe- have come from New Guinea (Phalanger orien- cies-rich islands to "supertramps," which are talis vulpecula, Rousettus amplexicaudatus absent from the largest islands but occur on stresemanni, Dobsonia minor, and Syconycteris many small species-poor ones. In between are australis), but one is a New Britain element "A-, B-, C-, and D-tramps" which occur on (Dobsonia moluccensis anderseni). Crown has large islands and also successively larger num- only one relevant element, Syconycteris aus- bers of small ones. These classes, he believes, tralis papuanus, from New Guinea. Long has reflect both dispersal abilities and (in the case two relevant taxa, one from New Guinea (Syc- of supertramps) inability to invade large stable onycteris australis) and one from New Britain communities. (Melonycteris melanops). Tolokiwa also has Diamond (1975) based his analysis on the three relevant elements, one from New Guinea birds of the Bismarck Archipelago area where (Syconycteris australis papuana) and two from he had data (unfortunately as yet unpublished) New Britain (Dobsonia moluccensis anderseni, for a great many species on virtually every Melonycteris melanops). Umboi has seven rele- large and small Bismarck island. Data for vant taxa, of which two have come from New mammal distribution is much less adequate, but Guinea (Syconycteris australis papuanus, I will try to make an analysis for the Mega- Murina fiorium) and five are shared with New chiroptera, for which information is best, and Britain (Phalanger orientalis ducatoris, Dobso- all of which are either fruit or flower feeders. nia praedatrix, D. moluccensis anderseni, Em- This information is based on museum speci- ballonura nigrescens solomonis, and Kerivoula mens and other records I have been able to cull 14 AMERICAN MUSEUM NOVITATES NO. 2690 from the literature, chiefly Laurie and Hill that occur on Bagabag and Sakar which have (1954). The species and their distributions are roughly the same area. The largest number of given in tables 2 and 3. For New Guinea, I species from any one island is 15 from New have only included the species which occur in Guinea (northeast). The species from the the northeastern part of the island. New Britain largest number of islands are Pteropus neo- is the only other island which is here consid- hibernicus and Dobsonia moluccensis, each ered "large." Three or more species are known known from 12 islands. for only 14 islands, and these species are the Eight species are known only from New only ones listed in the body of the table, but in Guinea or New Britain; all but one are New the footnote I have indicated additional islands Guinea elements, only Pteropus gailliardi from which certain species are known. I have being a New Britain endemic. The latter spe- combined Pteropus hypomelanus and P. admir- cies and the New Guinea Aproteles bulmerae alitatum as the P. hypomelanus group because are both poorly known so it is possible that they are closely related and probably allopatric these two species are not confined to their re- (apparently forming part of a superspecies, hy- spective islands. However, these eight should pomelanus being known only from western and almost certainly be considered "high-S" spe- admiralitatum only from eastern New Britain. I cies. One additional species almost qualifies as am aware that some of the material recorded a high-S species-Dobsonia minor-which is from Duke of York may have actually been confined to the New Guinea mainland (as far as collected on adjacent New Britain or New Ire- is known) except for its occurrence on land. However, I have decided to include all Bagabag. It might be classed as an "A-tramp." seven species since this is the same number Eight species fall in a middle ground, occur-

TABLE 2 Distribution of Megachiroptera on the Larger Islands of the Bismarck Area, including Northeast New Guinea (Largest islands to the left, smaller to the right.)

New New New Guinea Britain Ireland Manus Umboi Karkar Long Rousettus amplexicaudatus + + + Pteropus hypomelanus gp. + + - + + + + Pteropus tonganus + Pteropus temmincki + + Pteropus conspicillatus + Pteropus neohibernicus + + + + + + Pteropus macrotis + Pteropus gailliardi + Aproteles bulmerae + Dobsonia predatrix + + + Dobsonia moluccensis + + + + + Dobsonia minor + Nyctimene albiventer + + + + + Nyctimene draconilla + Nyctimene cephalotes + Nyctimene major - + + + Nyctimene cyclotis + Nyctimene aello + Paranyctimene raptor + Macroglossus lagochilus + + + + + + Syconycteris australis + + + - + Melonycteris melanops + + + 1979 KOOPMAN: ZOOGEOGRAPHY OF MAMMALS 15

TABLE 3 Distribution of Megachiroptera on the Smaller Islands of the Bismarck Area (Largest islands to the left, smallest to the right.)

Duke of Lihir Tabar Tolokiwa York Bagabag Sakar Crown Rousettus amplexicaudatus' + - + + Pteropus hypomelanus gp.2 + + Pteropus tonganus - - Pteropus temmincki + + Pteropus conspicillatus Pteropus neohibernicus 3 + + + - + Pteropus macrotis Pteropus gailliardi Aproteles bulmerae Dobsonia praedatrix 4 - + Dobsonia moluccensis S + + + + + Dobsonia minor - - + Nyctimene albiventer + + + + + + Nyctimene draconilla Nyctimene cephalotes Nyctimene major + + + Nyctimene cyclotis Nyctimene aello Paranyctimene raptor Macroglossus lagochilus + + + + Syconycteris australis + + + + Melonycteris melanops 6 - - + + 1 Also Emirau; 2 Also Manam; 3 Also Mioko and Rambutyo; 4 Also St. Mathias; 5 Also Boang and Emirau; 6 Also Mioko. ring on New Britain and in four cases also on iar supertramps. The species of birds that Dia- New Guinea, and also being known from three mond placed in this category occur on many to seven of the smaller islands. These may small islands. The two megachiropterans I have correspond to the "B- and C-tramp" categories. tentatively and perhaps erroneously assigned Pteropus temmincki occurs on three small is- here are known from a single small island in lands; Dobsonia praedatrix on four; Rousettus the Bismarck region (though each occurs amplexicaudatus, Nyctimene major, and Melo- widely elsewhere). I have no explanation for nycteris melanops on five; the Pteropus hypo- this peculiar pattern. It is certainly quite differ- melanus group and Syconycteris australis on ent from that seen in the three species I have six; and Macroglossus lagochilus on seven. placed in the D-tramp category. Three species occur on both New Guinea Diamond (1975) has stated that for birds in and New Britain and also are known from nine the Bismarcks, the smaller the island, the fewer (Nyctimene albiventer) or 10 (Pteropus neo- A-tramps and the more (relatively) D-tramps hibernicus, Dobsonia moluccensis) smaller is- and supertramps it supports. Due probably to lands. These would seem to fall into the "D- the fact that collections on many islands (ex- tramp" category. Finally, two species cept for New Guinea and New Britain) have (Pteropus tonganus, Nyctimene cephalotes) oc- been inadequate, no such relationship can be cur only on small islands, in this region being seen among the Megachiroptera. As can be absent from both New Guinea and New Brit- seen from tables 2 and 3, there is a rough ain. This would seem to place them in the tendency for larger islands to have more spe- supertramp category. However, they are pecul- cies recorded from them. However, even here 16 AMERICAN MUSEUM NOVITATES NO. 2690

New Ireland (the largest after New Britain) and pua New Guinea. Sci. New Guinea, vol. Bagabag (the next to the smallest) have the 3, pp. 117-122. same number of species (7) recorded from Hill, J. E. them. Furthermore, New Ireland has six of the 1961. Indo-Australian bats of the genus Tad- medium range (B-C-tramps) and one D-tramp; arida. Mammalia, vol. 25, pp. 29-56. 1963. A revision of the genus Hipposideros. Bagabag has one A-tramp, four in the medium Bull. Brit. Mus. (Nat. Hist.), Zool., vol. range, and two D-tramps. Quite obviously, 11, pp. 1-129. much more distributional data for Bismarck 1965. Asiatic bats of the genera Kerivoula and Megachiroptera is necessary before much can Phoniscus (Vespertilionidae), with a note be said about incidence functions of this group on Kerivoula aerosa Tomes. Mammalia, in this area. This in no way implies that I do vol. 29, pp. 524-556. not think that interspecific competition is very 1966. A review of the genus Philetor (Chirop- important, only that I do not think it can be tera: Vespertilionidae). Bull. Brit. Mus. demonstrated with the present data. (Nat. Hist.), Zool., vol. 14, pp. 371-387. 1971. The status of Vespertilio brachypterus LITERATURE CITED Temminck, 1840 (Chiroptera: Vesperti- lionidae). Zool. Mededelingen, vol. 45, Andersen, Knud pp. 139-145. 1912. Catalogue of the Chioptera in the collec- Koopman, Karl F. tion of the British Museum, 2nd ed. Brit. 1973. Systematics of Indo-Australian Pipistrel- Mus. (Nat. Hist.), vol. 1, pp. i-ci, 1-854. lus. Period. Biol., vol. 75, pp. 113-116. Barghom, Steven F. Laurie, Eleanor M. O., and J. E. Hill 1977. New Material of Vespertiliavus Schlosser 1954. List of land mammals of New Guinea, (Mammalia, Chiroptera) and suggested re- Celebes, and adjacent islands, 1758-1952. lationships of emballonurid bats based on London, Brit. Mus. (Nat. Hist.), pp. cranial morphology. Amer. Mus. Novi- 1-175. tates, no. 2618, p. 1-29. McKean, John L. Bergmans, W. 1972. Notes on some collections of bats (Order 1975. A new species of Dobsonia Palmer, 1898 Chiroptera) from Papua-New Guinea and (Mammalia, Megachiroptera) from Bougainville Island. C.S.I.R.O., Aus- Waigeo, with notes on other members of tralia, Div. Wildlife Res. Tech. Paper., the genus. Beaufortia, vol. 23, no. 295, no. 26, pp. 1-35. pp. 1-13. Mahoney, J. A. Calaby, J. H., and J. Mary Taylor 1968. Baiyankamys Hinton (Muridae, Hydro- [In press] Re-evaluation of the holotype of Mus myinae), a New Guinea rodent genus ruber Jentinck, 1880 (Rodentia: Muridae) named for an incorrectly associated skin from western New Guinea (Irian Jaya). and skull (Hydromyinae, Hydromys) and Zool. Mededelingen. mandible (Murinae, Rattus). Mammalia, Diamond, Jared M. vol. 32, pp. 64-71. 1974. Colonization of exploded volcanic islands Mayr, Ernst by birds: the supertramp strategy. Science, 1941. List of New Guinea Birds. New York, vol. 184, pp. 803-806. American Museum of Natural History, pp. 1975. Assembly of Species Communities, pp. i-xi, 1-260. 342-444. In Cody, Martin L. and Jared 1944. Wallace's Line in the light of recent zoo- M. Diamond (eds.), Ecology and evolu- geographic studies. Quart. Rev. Biol., tion. Harvard University, Belknap Press, vol. 19, pp. 1-14. pp. i-ix, 1-545. Menzies, J. J. Felten, Heinz 1977. Fossil and subfossil fruit bats from the 1961. Eine neue Unterart von Pteropus neo- mountains of New Guinea. Aust. Jour. hibernicus. Senck. Biol., vol. 42, pp. Zool., vol. 25, pp. 329-336. 417-418. Phillips, Carleton J. Greig-Smith, Peter W. 1966. A new species of bat of the genus Melo- 1975. Notes on a collection of bats and their nycteris from the Solomon Islands. Jour. ectoparasites from the Sepik district, Pa- Mammal., vol. 47, pp. 23-27. 1979 KOOPMAN: ZOOGEOGRAPHY OF MAMMALS 17

Phillips, Carleton J., and Elmer C. Bimey sectivorous bat, Murina. Jour. Mammal., 1968. Taxonomic status of the vespertilionid vol. 42, pp. 531-533. genus Anamygdon (Mammalia; Chirop- 1969. Results of the 1958-1959 Gailliard New tera). Proc. Biol. Soc. Washington, vol. Britain Expedition 5. A New Species of 81, pp. 491-498. Pteropus (Mammalia, Pteropodidae) from Sanbom, Colin Campbell New Britain, Bismarck Archipelago. 1931. Bats from Polynesia, Melanesia, and Ma- Amer. Mus. Novitates, no. 2371, pp. laysia. Field Mus. Nat. Hist., Zool. Ser., 1-16. vol. 18, pp. 7-29. Van Deusen, Hobart M., and Graeme G. George Tate, G. H. H. 1969. Results of the Archbold Expeditions. No. 1942. Results of the Archbold Expeditions. No. 90. Notes on the Echidnas (Mammalia, 48. Pteropodidae (Chiroptera) of the Arch- Tachyglossidae) of New Guinea. Amer. bold collections. Bull. Amer. Mus. Nat. Mus. Novitates, no. 2383, pp. 1-23. Hist., vol. 80, pp. 331-347. Van Deusen, Hobart M., and Karl F. Koopman Thomas, Oldfield 1971. Results of the Archbold Expeditions. No. 1915. On some Pteropine Bats from Vulcan and 95. The Genus Chalinolobus (Chiroptera, Dampier Islands, off the N.E. Coast of Vespertilionidae) Taxonomic Review of New Guinea. Ann. Mag. Nat. Hist., ser. Chalinolobus picatus, C. nigrogriseus, 8, vol. 15, pp. 387-389. and C. rogersi. Amer. Mus. Novitates, Van Deusen, Hobart M. no. 2468, pp. 1-30. 1961. New Guinea record of the tube-nosed in-