Biogeography of the amphibians in the islands of the Southwest Pacific

18 Pages·1997·8.7 MB·English

by  Walter Creighton Brown

PROCEEDINGS
OF THE
CALIFORNIA ACADEMY OF SCIENCES
Vol. 50, No. 2, pp. 21-38, 3 figs., 2 tables. October 8, 1997
BIOGEOGRAPHY OF AMPHIBIANS IN THE ISLANDS OF THE
SOUTHWEST PACIFIC
W
/ -v Cln
U
Walter C. Brown ' 2 ]$Qy
DepartmentofHerpetology, CaliforniaAcademyofSciences,
San Francisco, CA 94118
Presentpatternsofdistributionandendemismoffrogsintheislandsofthesouthwest Pacificare
examined.Thesepatternsareanalyzedintermsofthegenerallyacceptedgeological historyand
sea-levelchangesfortheregion.
Theevidenceindicatesthattheislandanuranfaunaincludesthreecomponents.OneisanAsian
component. RepresentationofthiscomponentontheSundaShelf islandsincludesalmostallthe
mainland genera and four on Borneo that are not known on the mainland. There is a great
diversityandhighendemismforspecies.ThePhilippinefaunaincludeslessthanhalfofthegenera
ofthe Sunda Shelf islands and a greatly reduced number ofspecies. The islands ofwestern
Wallaceahaveaboutone-fourthoftheSundalandgeneraandaboutone-seventhasmanyspecies.
Onlytwooftheranidgenera(RanaandLimnonectes)oftheAsiaticcomponentoccurintheislands
eastofwesternWallacea; and thereareonlytwoorthreespeciesonanyoftheseislandsexcept
forNew Guinea.TherethegenusRanahasseveralendemicspecies.
The second component is one derived from the Australian anuran fauna. In New Guinea (the
primary Sahul Shelfisland), two ofthe three Australian genera of Hylidae (Pelodryadidae of
some)occur,withalargenumberofspecies,mostofthemendemic.Oneofthese(Litoria)dispersed
tothe Melanesian Arcsand theislandsofeasternWallacea,thesecond(Nyctimystes)onlytothe
latter. Only 5 of 20 genera of the Myobatrachidae occur in New Guinea, and they have not
dispersed beyondtheSahulShelf.
Thethird component includes 16 genera in 2 endemic subfamilies ofmicrohylids (centered in
NewGuinea)and4generainanendemicsubfamilyofranids.Threeofthesegeneraarecentered
in the Melanesian arcs, and one in Melanesia and the Philippines. Relationships of these
subfamilies to other microhylid and ranid lineages are not clear at this time. Also one genus
(Batracliylodes)inthesubfamilyRaninaeisknownonlyfromtheSolomons.Itisnotclearlyrelated
toanyoftheAsiaticranidgenera.
Received February28, 1995.Accepted May 19, 1997.
Thenumerous islandsofthesouthwesternand formsoflifeonthese islands havebeen thebasis
central regions ofthe Pacific Ocean, extending fornumerous biogeographic hypotheses,
from thecoastofsoutheastAsiaandthenorthern From the time of Wallace's "Island Life"
coast of Australia to the Hawaiian, Line, and (1880) to Darlington's "Zoogeography: The
Tuamota islands in the Central Pacific (Fig. 1), Geographical Distribution ofAnimals" (1957),
provide isolated land areas ofvarious ages. The zoogeographers explained the presence of ani-
[21]
22 PROCEEDINGS OFTHECALIFORNIA ACADEMY OF SCIENCES
Volume50, No. 2
mals on islands in terms ofthe animals' use of of this study. These data are converted to dis-
known or imagined previous land bridges, their tributional patterns of genera and species as-
abilities to cross waterbarriers, ortheirpossible signed to families or subfamilies native to
transportbyman.Theseexplanationswerebased southeast Asia and Australia, or subfamilies in
on the geological concept that the earth's crust MicrohylidaeorRanidae thatarecentered in the
was a surface covering ofocean and land areas islands.
occupying relatively fixed positions. Since the These distributional patterns are evaluated in
1960s,thegeological conceptofcontinentaldrift relation to: (1) the geological events that pro-
has provided for very different explanations of duced the islands and established their current
island biogeography. spatial relationships, (2) past sea-level changes,
Recently, there has been renewed interest and (3) climatological history, (4) their dispersal
anumberofpapers on thegeological historyand abilities, routes, and opportunities, and (5) evo-
biogeography of the islands of the southwest lution and extinction events. Biogeographical
Pacific, with emphasis on those islands between hypotheses concerning these island anurans are
the Sunda and Sahul Shelves, the region often reevaluated.
referred to as Wallacea (Fig. 2). Many ofthese
paperswerepublished in3 volumes: "Wallace's Geological Historyofthe Islands
Lineand PlateTectonics," 1981,"Biogeography
of the Tropical Pacific," 1984, and "Biogeo- Theislandsintheareaunderconsiderationcan
graphical Evolution ofthe Malay Archipelago," be assigned to one offive groups: (1) those on
1987, Oxford Monogr. Biogeography, Claren- the shallow Sunda Shelfoffthe Asian coast, (2)
don Press, Oxford. As to organisms: angio- those on the Sahul Shelfoffnorthern Australia,
spermsreceivethemostattention,withemphasis (3) the clusters of islands between these two
on palms in the plant kingdom; birds, mammals, shelves,usuallytermedWallacea(Fig.2),(4)the
and Lepidoptera receive most attention in the Philippines, and (5) the Melanesian arcs: Admi-
animal kingdom. Amphibians and other terres- ralties, Bismarcks, Solomons, and Fiji.
trial, vertebrate fauna are dealt with in the paper The geological history ofthis region is com-
by Cranbrook (1981). Shore-fish distribution is plex, and many uncertainties still exist. But re-
discussed in a separate paper (Springer 1982). cent studies have changed interpretations ofthe
Numerousotherpapers, limitedtothegeologyof origins and history of southeast Asia and the
the region, have also been published in various islandsaswell astheirprobableages.Geological
geological journals or special publications dur- processes are measured in terms ofmegayears;
ing this same period. Other islands ofthe region some of the important sea-level changes and
including the Greater Sundas, Philippines, New climatological events in terms of thousands or
Guinea and its satellites, Bismarcks, Solomons, hundreds ofthousands ofyears.
and Fiji, are considered in manyofthesepapers. All ofsoutheastcontinental Asia from the Hi-
Several papers have summarized the biogeogra- malayan-Tibet region and southern China, as
phy ofthe heipetofaunas on some major groups well as the Greater Sunda Islands, is believed to
of islands (Allison 1996; Brown and Alcala betheresultofaccretionofterranesovermillions
1970; Inger 1954, 1966). The present paper is of years. The origins of these terranes and the
limited to the amphibian faunaofthese islands. times of their collisions are fairly well estab-
lishedforsomebutpoorlyunderstood forothers,
Methods especially the olderones.
At leastfivesuch terraneshavebeen identified
In this study, distributional data for anuran in the Himalayan-Tibet region to the north of
speciesinthesouthwestPacificislands isprimar- India. Theoriginoftheoldest(mostnortherly) is
ily based on island localities cited in Frost thought not to be Gondwanan (Nishiwaki and
(1985), Duellman (1993), Allison (1996), Uyeda 1987),butatleastthemorerecentofthese
Menzies(1982),ZweifelandTyler(1982),Inger terranes (south ChinaandpartofIndochina-Ma-
(pers.comm.),andZweifel(pers.comm.).Deter- lay-Sundaland) are rift-blocks from eastern
miningaccuracyofspecies' assignments inthese Gondwanaalong the north rim oftheAustralian
sources, orby preceding authors, was not a part region (Murphy 1987; Nishiwaki and Uyeda
BROWN: BIOGEOGRAPHY OF AMPHIBIANS 23
1987; Audley-Charles 1987, 1988). This rift pe- along the central mountain range, became
riod may have extended from the late Carbonif- subaerial in early Miocene (Hamilton 1979;
erous to the mid-Jurassic (150-300 MaBP) and Audley-Charles 1981). These islands, like those
was followedby longperiodsofnorthwarddrift- ofthe SundaShelf, havebeenconnectedtoAus-
ing across the changing Tethys Ocean. The ear- tralia by land bridges during periods oflowered
liestterranes may haveaccretedtosouthernAsia sea levels. —
before the end of the Jurassic (Audley-Charles Wallacea. The islands of Wallacea, be-
1987). DataindicatethatIndiariftedfromGond- tween the shelves (Fig. 2), are partly Asian and
wana about 140 MaBP and collided with Asia partlyAustralian,andvary inageandtype. Their
about 50+ MaBP (Butler —1995; Hallam 1981). present positions are the result of compression
Greater Sunda Islands. The Greater Sunda between the converging plates over the past 40
Islands include Sumatra. Java, Borneo, Palawan millionyears.TheAsiatic(western)partincludes
(western Philippines), and several (isolated or two groups. In the south are the Lesser Sunda
clustered)small islands,all situatedontheSunda Islands, a volcanic arc extending from Bali
Shelfoffthe coast ofsoutheastAsia (Fig. 2). At throughFloresandcontinuingastheInnerBanda
times, these have been united as a subaerial ex- Arc through WetarIsland. Thisarc dates from at
tension ofthe Indochina-Malay region, most re- least 20 MaBP (Audley-Charles 1987). In the
cently about 20.000 years BP, a result oflower north is Sulawesi, the western part of which is
sea levels during the last ice age. The origin of Asian in origin and was joined to Borneo until
Sundaland has been open to several interpreta- the Eocene (Audley-Charles 1981, 1987;
tions. Recent geological research indicates that McCabeand Cole 1987). The northern andeast-
it, liketheMalay region, istheresultofbringing ernparts,derivedfromthemarginoftheadvanc-
togetherfragmentsofeasternGondwana,Pacific ingAustralia-NewGuineaPlate,probablybegan
arcs, and Seamounts. These are the most recent to emergeas an island areaabout 5-10 MaBP at
terranes to be added to the southeast margin of about the same time as it joined with western
Asia, beginning about 50 MaBP (Audley-Char- Sulawesi (McCabeand Cole 1987).
les 1987). However, the main geological proc- The remaining islands of Wallacea on the
esses which resulted in the present shape of south and east originated from the advancing
Sundaland werecompleted about 15 MaBP (Oi- front ofthe Australia-New Guinea Plate. Those
lier 1985). of the Outer Banda Arc from Sumba through
Palawan, at the north end ofthe Sunda Shelf, Timor, Tanimbar, Kaie, and Ceram and Buru in
is a composite island. The southern part is asso- thesouthern Moluccas (Fig. 2), arenotvolcanic,
ciated with Borneo and its development. The buttheresultofrecent,probablyPliocene(1.5-5
northern part and the Calamian Group are from MaBP), upthrustalongthemarginoftheAustra-
asegmentofacontinentalterrenethatriftedfrom lian Plate (Milsom and Audley-Charles 1986;
China, probably in the Cretaceous, and collided McCabeandCole 1987;Nunn 1994). TheBang-
withthewesternedgeofthePhilippineplateand gai-Sula islandgroup, accordingtoone interpre-
south Palawan in the mid- to late Miocene tation, became subaerial by late Miocene. 3
(McCabe and Cole 198—7). MaBP (Audley-Charles 1981, 1987). Thenorth-
Sahul Shelfislands. These islands,with the ern Moluccan Islands, centered about the large
exceptionofthenorthernhalfofNewGuinea,are island ofHalmahera, are volcanic. Some ofthe
part ofthe Australian continent that rifted from volcanoes on Halmahera are Miocene, but most
Antarctica and drifted northward between 45 ofthe small islands are young, Quaternary vol-
MaBP and 15 MaBP. At that time the collision canoes(Audley-Charles 1987;McCabeandCole
with the Pacific, Asian, and minorborderplates 1987). Areas ofWallacean islands are not cur-
occurred. The northern part ofNew Guinea and rently as great as during Pleistocene periods of
small satellites along the north coast and Vo- lower sea levels. At such times, some of them
gelkop Peninsula are the result ofthe accretion may have been united but notjoined to the ex-
of Tethys Ocean marginal arcs onto the New posedareas ofthe Sun—daorSahul shelves.
Guinea portion of the Gondwanan Australia- Philippine Islands. The Philippine Archi-
New Guinea Plate about mid-Miocene. There is pelagoisaregionwithaverycomplexgeological
evidence that the central region ofNew Guinea, history. The main islands of Luzon and Min-
24 PROCEEDINGS OFTHECALIFORNIA ACADEMY OF SCIENCES
Volume50, No. 2
-
BROWN: BIOGEOGRAPHY OF AMPHIBIANS 25
_—n o— £«
!»o.£
— u
r-
^ J §
I E S.
2
•a Sui
5
_<cu —3 cua
— —
TO
3 •>
§-. D £
jj-2
o g S
— w "3
£ § i
X 5 u
a) o -2
£3Jc"Z£g§
.^2 c£ "°3
3 ul <U
(3 <U C
.g5 S8a.<sin
nO
_c u
ZU£ ,-<a3U -<£3D
° § 8
26 PROCEEDINGS OFTHECALIFORNIA ACADEMY OF SCIENCES
Volume50, No. 2
danao have large pre-Miocene areas (McCabe hypothesisconcerningthehistoryoftheMelane-
andCole 1987;Heaney 1991). Thereisevidence sian Arcs is that they had their origin from a
tosupporttheviewthatsomeofthevolcanoesof double-arcsystem(TethysArcs)onseparatesub-
these islands were subaerial as long as 70-140 plates that began to break offfrom northeastern
MaBP in the Cretaceous (Audley-Charles 1981; and eastern Gondwanan-Australia between
McCabe and Cole 1987). Other islands are gen- 100-60 MaBP, atabout the same time thatNew
erallyviewedasmorerecentarcs: Mindoro8-10 Zealand to the south was rifted from the north-
Ma, Negros-Panay-Cebu 1-4 Ma, and some of easternAntarcticasegment.Thesearc-blocksbe-
the smaller islands such as Camiguin and came widely separated from Australia by sea
Sibuyan 0.1-1.0 Ma (Heaney 1991). One view floor spreading while drifting north and north-
also holds that the Philippine Archipelago westward during the subsequent 30^0 million
achieveditspresentpositionbyanorthwarddrift years(Hamilton 1979;Coleman 1980;Halloway
fromtheEocenetopresent(JarrandandSasajima 1984).
1980; Oilier 1985; McCabe and Cole 1987). Following compression between the Pacific
Althoughthere isgeneralagreementthatmostof andAustralia-NewGuineaPlates, 15-20 MaBP,
the Philippines are oceanic in origin, the North the central portion ofthese arcs gradually fused
Palawan continental terrane was a part ofsouth to the advance edge (southern New Guinea) of
Chinaatleast intotheCretaceous. Followingthe theAustralianPlate. Partofthecentral mountain
rift from south China, the North Palawan block region is derived from islands ofthe Inner Arc
beganasouthwarddrift intheOligocene,collid- and the Vogelkop and northcoast mountain
ing with and contributing to the formation of ranges from the Outer Arc (Hamilton 1979;
northernPalawan,westernMindoro,andwestern Audley-Charles 1981; Halloway 1984). Other
Panay. Other areas of Mindoro and Panay are parts ofthe OuterArc persist as an alignment of
derived from the Mindoro-Panay disrupted ter- archipelagos to the east, Bismarcks, Solomons,
rane and the Central Philippine Arc terrane Vanuatu, and Fiji (Coleman 1980; Halloway
(McCabe and Cole 1987; McCabe et al. 1985). 1984). TheVanuatuArc underwentasouthward
There is also evidence that sea-level changes rotation beginning 6-8 MaBP that brought it to
caused temporary land connections with north- its present position just east ofNew Caledonia
easternBorneoatvarioustimes. Theloweringof (Coleman 1980; Kroenke 1984).
sea level during the late Pleistocene glaciation, Just as lower sea levels caused by Pleistocene
18,000-20,000yearsagowouldhavecreatedthe glaciation events resulted in land connections
Greater Islands shown in Figure 3 and joined between Borneo and Palawan and possibly be-
Palawan to Borneo, but would not have closed tween Borneo and Greater Sulu at times in the
the water barriers between Borneo and Greater recentpast,theycreatedsimilarlandconnections
Sulu, Greater Sulu and Greater Mindanao, or between Australia and New Guinea. Although
GreaterMindanaoandGreaterLuzon.Ithasbeen theBismarcksandSolomonswerenotconnected
suggested (Morley and Flenley 1987) that in to New Guinea during these periods, the water
mid-Pleistocene (perhaps 100,000 years ago or channels separating them were narrowed. Also
more) even lower sea levels may have closed withintheSolomonsandBismarckarchipelagos,
thesegapsandalsothosebetweenGreaterLuzon the Pleistocene lower sea-level periods resulted
andGreaterNegrosaswellasGreaterLuzonand intheunitingofexistingislands intomuch larger
Greater Mindoro (Figs. 2 and 3). These island islandsforvaryingperiodsoftime(seeDiamond
connections would have required a sea-level andMayr 1976).
lowering of300 m or more, unless the sea floor
was higherat thattim—e. Anuran Families inthe Islands
Melanesian Arcs. The island arcs which
contributed to the formation of northern New Components I, II and III
Guinea at the collision of the Australia-New The anuran fauna ofthe southwest Pacific Is-
Guinea and Pacific Plates, as well as the more lands can be assigned to one of three compo-
eastern Admiralty, Bismarck, Solomon, Fiji nents: (1) Component I, those resulting from
Arcs, and the Palau Arc are also outsideofWal- colonization by present-day Asian stocks, (2)
lacea. Based on recent geological evidence, one ComponentII, thoseresultingfrom colonization
.
BROWN: BIOGEOGRAPHY OFAMPHIBIANS 27
#122°
BatanIslands•
KEY
Lubang
1
2.Burias
BabuanIslands
3.Tablas-Romblon
4.Sibuyan
5.Camiguin
18°-
ss than 120 meters
GreaterLuzon rent water depth
PleistoceneIsland
200
MindoroIsland
SamarIsland
/ <&
GreaterPalawan / o \ j.
Pleistocene Islan•d^.A, 4 ; <_ y-V^l "'!tt\
Mlfe
i>'^ £*v \ 3**r Leyte
X^/
10° Great*
si
Negros-Panay
PleistoceneIsland
-6° 1a «^ tl(_^•—5v^,^7^^' GMrienadtaenrao
<# PleistoceneIsland
—IP
Borneo
GreaterSulu
PleistoceneIsland
122°
Figure3.Philippines:presentislands(palestippledareas)andlatePleistocene,about20,000yearsBP(darkstippledareas)
basedonpresumedlowersealevelsofabout 120m(afterBrownandAlcala 1994).
28 PROCEEDINGS OFTHECALIFORNIA ACADEMY OF SCIENCES
Volume50, No. 2
by present-day Australian stocks, and (3) Com- Ford and Cantella 1993; and Duellman 1993.
ponent III, those genera in distinct subfamilies Several other major papers are limited to the
thatarecentered intheislands.Adler,Austinand southwest Pacific Islands (Brown 1953; Inger
Dudley(1995)dividedtheislandskinksthatthey 1954, 1966; BrownandTyler 1968; Zweifel and
were studying into comparable categories for Tyler 1982; Allison 1996).
theiranalysis ofdistribution patterns.
The first two components are easy to identify Results
based on the genus and family (sometimes sub-
family) to which the species are assigned. Six of Because this paper is concerned with the bio-
the eight families represented in the island fauna geography ofthe anurans ofthe southwest Pa-
are Asian (Component I) and nearly all of the cific Islands, butnot thoseofmainland Asiaand
genera are mainland genera (Table 1, Appendix Australia,only thenumberofgeneraandspecies
A). Theothertwo familiesareAustralian (Com- in Australia and southeast Asia are indicated in
ponent II), and the genera are mainland genera the tables. Forthe island areas, notonly the total
(Table 1, Appendix A). Component III includes number but the number that are endemic is also
two distinct subfamilies (Asterophryinae and given. The number ofgenera and species ofthe
Genyophryninae) of the family Microhylidae, AsianandAustraliananurans(ComponentsIand
withmostofthegenerarestrictedtoNewGuinea, II)andtheirdistributionpatternsaresummarized
and a distinct subfamily (Platymantinae) ofthe in Table l andare—listedbyname inAppendix B.
familyRanidaewithmostofthegenerarestricted Component I. Forthe families Bufonidae,
to the Solomons and Bismarcks but with one Megophryidae. and Rhacophoridae and the
genus (Plat\'mantis) also centered in the Philip- Asian lineagesoftheMicrohylidaeand Ranidae,
pines (Table 2, Appendix C). 27of34(82%)ofthegeneraoccurintheGreater
Sundas (Appendix B). This, along with 67 of
History ofthe Data Base southeast Asian species on the larger Sunda Is-
Dataon thespeciesofanurans inthe islandsof lands, suggests good dispersal opportunities in
the southwest Pacific have been acquired over the relatively recent past.
the past century and a half. Only a few species This is consistent with the evidence that the
hadbeen recordedpriortopublicationof"Island SundaShelfislandshavebeenconnectedtoeach
Life'1 (Wallace 1880), most notably the several other and to mainland southeast Asia at various
species described from New Guinea (Peters and times in the Pleistocene as a result oflower sea
Doria 1878) and several from the Philippines levels, mostrecently about 18,000-20,000years
(Boulenger 1882). ago. Thepresenceofoneendemicgenusforboth
During the last couple ofdecades ofthe I9th the Megophryidae and the Ranidae in Borneo
centuryandthefirstcoupleofthe20th,collecting and another Staurois in both Borneo and the
in the Solomons, New Guinea, and the Philip- Philippines may be the resultofan extinction of
pinesgreatlyincreased.Ofparticularinterestwas those generic lineages on the mainland, their
the discovery ofthe generathatwe now classify evolution in situ, or a sampling weakness for
as the platymantine ranids in the Solomons some areas ofsoutheastAsia. I treatBarbourula
(Boulenger 1886). Oneofthegenera(Platyman- in the Bombinatoridae (Ford and Cantella 1993)
tis) was also known from the Philippines. Two known from one species in Borneo and one in
monographs (Kampen van 1923 and Taylor Palawan, as a relict genus ofthat Asian family
1920) summarized the data on the frogs of the represented by the genus Bombina in China but
Indo-AustralianArchipelagoandthePhilippines notinsoutheastAsia.TheEurasian lineageofthe
respectively. family Hylidae, with one species of the genus
Arenewed interestintheherpetofaunaofthese Hyla recorded from Indochina and Thailand, is
islands began during World War II, and contin- not known from the Pacific Islands.
ues to intensify even today. The resulting litera- Fifteen (48%) of the Bornean genera, exclu-
ture includes numerous short papers describing siveoftherelictBarbourula,areknownfrom the
species. Major publications during this period Philippines. Eighteen speciesofthesegeneraare
thatsummarizedataonsystematicsanddistribu- shared with Borneo and 30 species are endemic.
tion ofanurans worldwide include: Frost 1985; Sixofthesevenfamilies insoutheasternAsiaand
BROWN: BIOGEOGRAPHY OF AMPHIBIANS 29
^
=2
b
*
O
ra
^ GO
oO^o5 .0o20
.5 o-
00
CCD ~CD
3 3
o
cco c«s
3CO <cy
eCO Cdl>b
< 2"S
co S "o
<Sco H-0=1 oC„
."5>2 "f«c -c—3o
00
^ >>
3
ca 00
.5 -n -C
-a E ^c
—« to °g
CcO •a££> C""O
0OCaC§>-.-—to32o->-.1CSSaS=O
U
,1-2nw>» coV
a. E
_c
30 PROCEEDINGS OFTHECALIFORNIA ACADEMY OF SCIENCES
Volume50,No. 2
southernChinaarerepresented. Thisdataiscon- widespread L. grunniens. Rana has a group of
sistent with the concept that very narrow water nine species, sevenclosely relatedand restricted
channels or direct land connections may have toNew Guinea,and two lessclearlyrelated. The
existed between northwestern Borneo and seven species are very similar morphologically
PalawanandnortheasternBorneoandMindanao andcytologically(Menzies 1987).Noneofthese
as recently as the lower sea-level event species are conspecific with species in the
(18,000-20,000 years ago), and during earlier Greater Sundas orwestern Wallacea.
periodswhensealevelswereevenlower(Fig.3), Only onegenus (3%) ofthe Sundaland genera
mid-Pleistocene (100,000-500,000 years ago) is known from the Melanesian Arcs, with two
andatothertimesbacktothe lateTertiary(Mor- species (one endemic) in the Bismarcks and one
ley and Flenley 1987). The effect of sea-level endemic species in the Solomons. These are the
changes on the distribution patterns of rhaco- result ofsecondary colonization from the New
phoridfrogswithinthePhilippineArchipelago is Guinearadiation. Oneofthewidespread species
discussed by Brown andAlcala (1994). There is in New Guinea, Rana daemeli, is also known
no direct evidence as to how the composition of from apopulation—on Cape York, Australia.
the current faunas have been affected by extinc- Component II. For the two families (Hyli-
tion ordispersal from—thePhilippines toBorneo. dae and Myobatrachidae) of Australia, repre-
Western Wallacea. The sharp reduction in sentation in the Sahul Shelf islands of New
the numberofAsiatic genera and species in Su- Guinea and its satellites differs in several re-
lawesiandtheLesserSundas(Table 1,Appendix spects from that of the representation of the
B) indicates much more limited dispersal oppor- southeast Asian fauna in the Greater Sunda Is-
tunities between eastern Sundaland and the is- lands. Ofthe 23 genera in Australia, only seven
lands ofwestern Wallacea than existed between (30%) are presently known from New Guinea
Sundaland and the mainland or the Philippines (Table 1, Appendix A). However, the two fami-
and Borneo. Eight (30%) oftheAsia-Sundaland lies are very different. Two (67%) ofthe three
generaoccurin Sulawesi andseven (26%) inthe genera of the Hylidae are in New Guinea,
lesser Sundas. But at the species level, the whereas only five (25%) ofthe 20 genera ofthe
number recorded from Sulawesi is almost twice Myobatrachidae are in New Guinea.
that ofthe Lesser Sundas and endemism is five Atthespeciesleveltheyalsodiffergreatly.For
times greater. This is consistent with the lack of the Myobatrachidae there are only 7 species in
evidence of land connections between Borneo the 5 genera, and none are endemic. For the
and Sulawesi or between Java and the Lesser Hylidae thereare 80 species, 71 (89%) endemic.
Sundasatlow sea-level timesduringthePleisto- Thenumberofspeciesisactuallygreaterthanthe
cene.Thenarrow,butdeepMakassarStraitsepa- 69 recordedfrom Australia(Table 1). Foroneof
rates Sulawesi from Borneo and the Lesser thegenera,Nyctimystes,thereisonlyonespecies
SundasfromJava.Also,Sulawesiisgeologically known from Australia but 22 in New Guinea.
much more complex than the Lesser Sundas, These differences in colonization and radiation
withwesternSulawesi olderthantheeasternpart success of the two families suggest an earlier
or the Lesser Sundas (Audley-Charles 1987; colonization for Hylidae than for Myobatrachi-
McCabe andCole 1987). dae.
Only three (10%) ofthe genera ofthe Greater Because land bridges between Australia and
Sundas(Table 1,AppendixB)arerecorded from New Guinea would have been created by the
eastern Wallacea. Ofthe five species ofranids, same lower sea-level events as were the Sun-
one species Limnonectes verruculosa is treated daland bridges, the dispersal opportunities for
as endemic, and L. modesta is otherwise known anurans during the Pleistocene were probably
only from Sulawesi. These should be validated. similar between Australia and New Guinea to
Two species areconspecific orclosely relatedto those proposed between southeast Asia and the
a group ofspecies ofRana that evolved in New SundaShelfislands.Twofactorsaremostimpor-
Guinea. tant: the large land mass ofAustralia compared
Only two (6%) Sundaic genera have success- to that ofsoutheast Asia adjacent to the Sunda
fully colonized the Sahul Shelf island of New Shelf and variation in climate for different re-
Guinea. Limnonectes has only one species, the gions of Australia. A comparison limiting the