Wallace's Line in the Light of Recent Zoogeographic Studies Author(s): Source: The Quarterly Review of Biology, Vol. 19, No. 1 (Mar., 1944), pp. 1-14 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/2808563 . Accessed: 03/02/2015 17:45

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to The Quarterly Review of Biology.

http://www.jstor.org

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions VOL. 19, NO. I March,I 944 THE QUARTERLY REVIEW of BIOLOGY

WALLACE'S LINE IN THE LIGHT OF RECENT ZOOGEOGRAPHIC STUDIES BY ERNST MAYR AmericanMuseum of Natural History, New York Z OOGEOGRAPHY has had a fate very by the Atlantic,do not differso widelyas muchlike taxonomy. It was flourishing and ." There is much truth,in this duringthe descriptive period of biological statement. Except for bats and a few rodents, sciences. Its prestige,however, declined the onlynative mammalsof Australiaare marsu- rapidlywhen experimental biology began pials and monotremes. These same two groups to come to the foreground.Again as with taxon- are entirelylacking in Asia and are replacedby a omy, a new interestin zoogeographyhas been wide varietyof placentalmammals, such as mon- noticeablein recentyears. It seems to me that keys, shrews,squirrels, ungulates, and so forth. this revival has had two causes. One is the An equally pronouncedfaunal differenceexists interestof the student of geographicspeciation among ,insects, and othergroups of animals in the findingsof the zoogeographer. A studyof of the two regions. past and present distributionsyields much in- Australiaand Asia are connectedby a belt of formationon isolationof populationsand on the islands,the ,and the question dispersalof species. It-is in this connectionthat naturallycomes up as to where in this island I became interestedin zoogeography. regionthe borderlineis to be drawnbetween these The other reason is the introductionof new two fundamentallydifferent faunas. After re- methods. The intensiveexploration of all corners viewing the zoological evidence known to him, of the globe duringthe past fiftyyears has led Wallace (l.c.) comes to the followingconclusion: to an accumulationo; sufficientfaunistic data "We may considerit establishedthat the Strait to permitthe applicationof statisticalmethods. of [betweenBali and Lombok] (only 15 Furthermore,the science of ecology has reached mileswide) marksthe limit and abruptlyseparates a levelof maturityat whichit is beginningto affect two of the great zoologicalregions of the globe." profoundlyzoogeographic methods and principles. With thesewords he drewa zoogeographicbound- It seemedworth while to me to studythe contro- ary which was destinedto gain fame under the versialand stillwide open subjectof theborderline name of its author: "Wallace's Line," a termfirst between the Australian and Oriental Regions used by Huxley (1868) (Fig.' 1). It runsbetween withthe help of such modernmethods. and Lombok in the south, then through A. R. Wallace, who is generallyconsidered the between and Celebes, foremostrepresentative of classicalzoogeography, and finallyturns into the open Pacific between states in his famousessay On thezoological geog- Mindanao ()and the SanghirIslands. raphy of the Malay Archipelago(1860): "The This convenientborderline found quick acceptance westernand eastern islands of the archipelago in the zoological literature and was without belong to regions more distinctand contrasted hesitation adopted by nearly all 'the zoogeog- than any other of the great zoological divisions raphers publishing between 1860 and 1890. ofthe globe. SouthAmerica and Africa,separated Sarasin (1901) and Pelseneer (1904) should be 1

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions 2 THE QUARTERLY REVIEW OF BIOLOGY

consultedfor a historicalsurvey of the earlier facts became better known. Wallace himself literature.The echoin thepopular literature of was muchless positivein his laterwritings. Since this period was even more enthusiastic.A then many writershave insistedthat 'Wallace's mysteriousljne, only 15 mileswide, that separates Line was entirely imaginary (Weber (1902), marsupialsfrom , and honeyeaters and Pelseneer (1904), Mertens (1930), Brongersma cockatoesfrom barbets and trogons,could not (1936), and others). Van Kampen (1909), for failto appeal to theimagination of thelayman. example, asserted: "Such a sharp boundary as

-- Waltacestime ofi(uxle4)

V~ 1.I ~~~~~~~~~~~~~~~~~

1_7jj-- - Wallace'Ltne rinal) n F z Webeef t >m (Faunl Balanlc) ''-'-' tf | ~~/L fAtoPapuanb ainlandFauna

' - BDORNEO ______s -- - Ula_lacesItne -O- +i*a L

.0.0/ .. I

Ulbe' Ltne(Fana hBalance 9 / as

du I W~~~Uallace'slie (of ifuxleql) /_____ l--UIallaceslne (Original) /'.= Uleber~sI ne (Faunal B3alance) __ - - Ilieber'sJlJe (of Pelseneer) 1_ A +- +- LpitofA8ustralo-PapuanMainlandFauna .j_STRALI ZOO-mIme (Nd,e of Continental Thelves) .2 iUST 2 LIA* le'dta4Qy_deI. ||* J_t,

FIG 1. ZOOGEOGRAPHIc BORDERLINES IN THE MALAY ARCHIPELAGO The shaded areas are the continentalshelves.

E. Haeckel (1893) outdid all his contemporaries Wallace drewit does not exist. Not onlyis there by asserting:"Crossing the narrow but deep none where he drew it, but no such line exists we' go with a single step from anywhere in the archipelago." On the other the PresentEra to the Mesozoicum." hand,Wallace's Line has been vigorouslydefended Statementsof such exaggerationcall forrefuta- by such serious,authorsas Dickersonet al. (1928), tion and shortlyafter 1890 doubtswere expressed Raven (1935), and Rensch (1936). Curiously more and more frequentlyas to the validity of enoughmost of the writerson this subject seem Wallace's Line,particularly after the distributional to be definitelyin one or the othercamp, either

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions WALLACE'0S LINE 3

theyare forWallace's Line or theyare againstit, island belt,frofii,fhe east as a descendantof a and they tend to presenttheir data accordingly. group of species that was isolated in Australia Others treat one aspect only of this diversified at an early date. Such secondarilyeastern ele- problem. An impartialstudy of the situationis ments,as Merops ornatusamong the birds, are stilllacking at the presenttime. includedwith the easterngroup. The classifica- Actually, a whole complex of questions is tion of a few species will always remainopen to involved,of which the followingseem to be the doubt, but a differentdecision in these cases most importantones: would change the percentagesonly slightlyand (1) Is Wallace's Line the borderlinebetween would not basically affectthe followingfigures. the Orientaland the AustralianRegions, and if A specialistof a given group usually has no dif- not, whereis thisborderline? ficulties in deciding which species are Indo- (2) Does Wallace's Line representthe line of a Malayan and whichAustralian. major faunal break, and if this is true,how did Celebes. Weber (1902), the Sarasins (1901), such a break develop? de Beaufort(1926), Stresemann(1939) and other recentauthors agree that at least threefourths of IS WALLACE' S LINE THE BOUNDARY BETWEEN THE the Celebes animals are of westernorigin. Ac- ORIENTAL AND THE AUSTRALIAN REGIONS? cording to Rensch (1936: 252) the figuresare: The faunaof the Malay Archipelagowas rather Reptilesat least 88 per cent,Amphibia 80 per cent poorlyknown in Wallace's days. Wherehe knew TABLE 1 20 species of birds,we now know 120; wherehe Percentageof westernand eastern species on Lesser Sunda knew 5 species of reptiles,we know 40, and so Islands forth. This lack of informationcaused Wallace to singleout what he consideredtypical represen- REPTILES AND AM- BIRDS CHANGE tatives of the respectivefaunas, and to use the PHIBIANS OF PER- CENTAGE borderline of their ranges as zoogeographic IN BIRDS Western WesternEastern boundaries. The ,the squirrelsand other mammalsgo as far east as Bali, but are absent Per cent Per centPer cent fromLombok. Among birds the barbets (Capi- Bali...... 94 87.0 13.0 14.5 tonidae) and many other Oriental groups are Lombok...... 85 72.5 27.5 4.5 abruptlybrought to a halt by Lpmbok Strait. ...... 87 68.0 32.0 5.0 .78 37.0 The Australianhoneyeaters (genera Philemon and 63.0 5.5 AlorGroup . - 57.5 42.5 Meliphaga) and the cockatoe (Cacatua) reach Lombok, but not Bali.' The faunal difference on eitherside of Makassar Strait is even more and butterflies86 per cent. In birdsthe figureis striking:A rich Orientalfauna on Borneo and a slightlylower. Among 74 species of Passerine marsupia'l(Phalanger) on Celebes. It was on the birds 67.6 per cent are western. The percentage basis of such data that Wallace came to the con- for the old endemics (genera and good species) clusionthat Lombok and Makassar Straitsform and forthe morerecent immigrants is quite simi- the boundary between the Oriental and the lar. There is no doubt, Celebes must be in- AustralianRegions. cluded with the OrientalRegion. An analysisof the now available extensivefaunal LesserSunda Islands. Table 1 shows.the ratio lists does not bear out Wallace's conclusion. of the western and the eastern elements'on a Aftereliminating a few widespreadspecies, the numberof islands (the data ofreptiles and amphib- faunaof each of theislands of theMalay Archipel- ians are fromMertens, 1930; the data on birds ago can be dividedreadily into two groups: One are original). Rensch's (1936) careful analysis consistsof westernspecies, that is, species which shows that the Indo-Malayan element prevails are derived fromthe OrientalFauna, the other numericallyas fareast as the islandsof the Timor of easternspecies, that is, such whichare derived group. This is equally true for flyinganimals from the Australian Fauna. In a few species (birds and butterflies)and for flightlessgroups it is apparent that the genus or the familyto (mammals,land snails). whichthey belong was originallyof westernorigin, The figuresin Table 1 permitonly a singlecon- but that the particular species arrived in the clusion: Wallace's Line is not the borderline

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions 4 THE QUARTERLY REVIEW OF BIOLOGY betweel the Australianand the OrientalRegions. 120. It is even more true for freshwaterfish: The firstof the questions asked above is thus Borneohas 162 species'ofthe carp familyCyprini- answeredin favorof Wallace's opponents. dae, Celebes has none; has 55 species,Lom- bok has apparentlyonly a single one. Raven DOES WALLACE'S LINE INDICATE A MAJOR (1935) shows that the Mammalian fauna is FAUNAL BREAK? equally impoverished. The same is true for the The fact that Wallace's Line is not the border Philippines,their fauna is badly depleted, as between the Oriental and the Australo-Papuan comparedto thatof Borneo and Palawan (Dicker- Regionsis not thecomplete answer to ourproblem. son etal., 1928). A line which has been defendedso vigorously by so many zoogeographersmust have some THEE GEOLOGY OF THE MALAY ARCHIPELAGO significance.It is worthy of notice that its Why the islands , Java, Borneo, and staunchestdefenders were those naturalistswho Palawan should have a rich animal life,whereas actually studied and collectedthe animal lifeon the Philippines,Celebes, and the Lesser Sunda both sides of the line, like Dickerson and his Islands have a poor one, cannot be understood associates in the Philippines,like Raven who withouta studyof thegeological conditions. The repeatedlycrossed Makassar Straitin a sail boat British geologistEarle pointed out, as early as fromBorneo to Celebesand back,and likeWallace 1845, that geologicallythe Malay Archipelago and Rensch who crossedback and forthbetween consistsof threeparts, a westernone comprising Bali and Lombok. The actual impressionsof the greaterSunda Islands and the adjoiningparts theseworkers are vividlydepicted in a quotation ofAsia, whichwas verystable during the Tertiary, from\one of Rensch's books. Arrivingon Bali an eastern one consistingof and aftera prolongedexploration of Lombok, Sumbawa Australia,which was also stable,and an unstable and Flores,he asks himself: island belt in between. The unstablearea, com- "Whatabout the animal life? Is itreally as differentprising the Philippines,Celebes, the Moluccas, fromthat of Lombok, as has beenclaimed by so many and the , has a most compli- othertravellers? Is thesmall strait between the two cated geological structure. Deep sea basins, islandsactually a sharp faunaldivision? A strait, grabens,geosynclines and geanticlinesare scram- whicheven the smallest could cross without any bled togetherin a bewilderingmanner. Geologists difficulties?... Andthe difference is indeed quite extra- are still far from agreementin regard to the ordinary!Much more conspicuous than I wouldhave interpretationof these structures. So much, eyerimagined. As soonas I enteredthe woods on a however,is clear-that thisarea is highlyunstable sgall nativetrail a wholechorus of strange bird songs greetsme-in fact,among the real songstersthere is and that it has seen many and violent changes not a singleone withwhich I was familiar[from the in the recentpast. islandseast of Wallace's Line].... Onesurprise follows Originally,that is in late Mesozoic times, theother. The veryspecies that are most common on Celebes, the Moluccas, Misol, and westernNew Bali, are absenton theislands to theeast. The most Guinea, seem to have been situatedon the same characteristicbird of these woods is a greenbarbet . . . continentalshelf. The fossil marine faunas of it belongsto thefamily Capitonidae which is entirely the mentionedregions, as well as tectonicfeatures absentoh Lombok! The woodpeckersalso, which are prove this close relationship. In fact, most represente'don theislands farther east by a singlespe- geologists consider it as well established that cies are foundon Bali in fivedifferent only, species. Asia and Australia were in broad continental On theother hand I misseda wholenumber of species to the end of the Mesozoic. ofbirds which are characteristicfor the islands visited connectionup very previously..." (Rensch1930). The Tertiarywas a period of very active oro- genesis. Part of the Philippinesand of northern An unemotional statistical analysis of the Celebes seem to have been foldedup first. There faunal data tends to supportRensch's assertions. is some evidencefor the existenceof additional The moststriking feature of Wallace's Line is that islands during Eocene and Oligocene, as, for it separates a zone with a rich animal life from example, in the Timor region, but the exacd a badly impoverishedone. Borneo has about position,size and chronologyof such islands is 420 species of breedingbirds, Celebes only 220. unknown.In early Miocene,or accordingto other Java has about 340 breedingspecies, Lombok only authorsin very late Oligocene,the crustof the

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions WALLACE'S LINE 5 earth seems to have buckled down in a gigantic Tertiary-geologysupplies, however, no explana- manneralong a line, which is roughlyindicated tionfor a faunaldifference between Bali and Lom- by the west SumatranIslands, Timor,Kei, Seran, bok, a differencewhich seems to be due to events and-Halmahera. The verystrong negative anom- of a morerecent geological past. A considerable alies of the gravimetricmeasurements along this quantityof ocean water accumulatedin the polar lineare accordingto VeningMeinesz good evidence ice caps during the Pleistocene glaciations. It forthe occurrenceof such an event. The folding has been calculatedthat this resulted in a lowering was so violentthat it resultedin the widespread of the sea level of tropicalocean by at least 70 m., overthrustingof older strata over youngerones. but more probablyby 150-m. This caused the The so-calledouter Banda arc, consistingof the dryingup of all shallow seas and resultedin a islands Sumba, Timor, Babber, Timorlaut,Kei, considerableextension of land on Sunda and Sahul Seran,and Buru was formedalong part of this fold. shelves (see Fig. 1). Sumatra,Java, and Borneo- All of these islands are geologicallyvery similar. unitedwith the Malay Peninsulain the formation Slightlylater, but stillin theMiocene, a secondfold of "," an extensionof the Asiaticmain- was formedconsisting of parts of Sumatra and land and Bali became attached to this continent. Java, as well as of the so-calledinner Banda arc Lombok, however,which is separated fromBali (Bali, Lombok, Sumbawa, Flores, Alor, Wetar, by a straitof a depth of 312 m., remainedsepa- Dammer, and Banda). Most of this fold re- rated,even'though it was fusedtemporarily with mained, however,at firstsubmerged under the Sumbawa. ocean. In fact,some of the islandsmay not have The geological backgroundof Wallace's Line emergeduntil well in the Pleistocene. Later in is thus as follows:In its central part, between the Tertiary,particularly in the Pliocene and Borneoand Celebes,it followsthe edge of the con- Pleistocene,extensive fault lines developedwhich tinentalSunda shelf,in the south betweenBali led to the liftingof large blocks (marineterraces and Lombok (and the same is true in the north in Timor rose 1280 m.!) and the corresponding betweenBorneo-Palawan and the Philippines)it sinkingof otherblocks to formflat-bottomed deep indicates the eastern edge of the Pleistocene sea basins. The geological data indicate that Sundaland. The faunal break, which I have periodsof violent tectonic activity have alternated shownto existalong Wallace's Line, appears now in with periods of relative quietness,and the fre- a new light. It is due to the fact that the line quencyof earthquakes and thecontinuous volcanic separates,on the whole,a continentalfrom an in- activity in this region reveal that the orogenic sular fauna. This separation is clear cut in movementshave not yet completelydied down. Makassar Strait,but it is ratherobscured along The writingsof Molengraaf (1922), Umbgrove the Sunda arc, wherethe geanticlineof the inner (1932, 1934), and Kuenen (1935) should be con- Banda arc protrudesfrom Sundaland like a penin- sultedfor further details concerning the geologyof sula. Faunal breaks along this chain of islands thisregion. Three factsof zoogeographicsignifi- occur not only on Lombok Strait,but also on all cance seem to stand out among the geological the other inter-islandstraits. A number of data: (1) There is no evidence whatsoeverfor authors,among whom Mertens (1930) is foremost, any continentalconnection between Borneo and have contended,that some of the other straits, Celebes. In fact, the distance betweenthe two as thatbetween Java and Bali, or the one between islandswas, up to the Pleistocene,greater than it Sumbawa and Flores,are even moreefficient dis- is today; (2) Java, Bali, Lombok, and the other tribution barriers than Lombok Strait. This islandsof the innerBanda arc are situatedon the assertionis in conflictwith the above-given findings same geanticline;and (3) there is no geological of Pleistocenegeology and it becomes therefore evidencefor any cross connectionsbetween inner necessary to examine the relative efficiencyof and outer Banda arcs, except possibly between these water barriersin more detail. Sumba and Flores. The firstof these threeconclusions shows that THE EFFICIENCY OF THE WATER BARRIERS BETWEEN Makassar Strait is an ancient ocean barrierand THE LESSER SUNDA ISLANDS that at least thisparticular part of WTallace'sLine The faunalchange between Borneo and Celebes is geologicallywell-founded. Geologists and zoo- is abrupt,but it is much moregradual along the geographersare in full agreementon this point. west of Wallace's Line. The numberof species

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions 6 THE QUARTERLY REVIEW OF BIOLOGY of birdson thisisland chain is as follows:Sumatra icance ,ofBali Strait as a zoogeographicbarrier. about 440; Java, 340; Bali, 166; Lombok, 119; It would be entirelymisleading to subtractthe Sumbawa, 123; Flores, 143; and Timor, 137. In numberof Bali species fromthe total numberof the freshwaterfish family Cyprinidae,Sumatra Java species and state that the differenceconm- has 115 species, Java, 55; and Lombok only a prises the species that are unable to cross Bali single one. Of butterfliesSumatra has 334 Strait. This method was actually applied by species;Java, 270 species. Of reptilesSumatra has Mertens (1930) and Brongersma(1936), who ar- 193 species; Java, 136 species (Rensch 1936). It rived therebyat the erroneousconclusion that is obvious fromthese figuresthat the animal life was the most importantbarrier along of Java is considerablyimpoverished as compared the Sunda chain. with that of Sumatra (or Borneo). The reasons A faunal change betweenthe westernand the forthis are not entirelyclear, but twofactors seem easternend occursprobably not onlyon Java, but tQ be mostimportant. One is the heavy activity on all elongatedislands of the Sunda chainsuch as of the Javanesevolcanoes, particularly during the Sumbawa, Flores, and Timor. This fact invali- Pleistocene,which covered a goodpart of the island dates to some extentthe figureson the subsequent wvithlava and ashes and may have exterminated calculations,but it is fortunatelyof minorimpor- a numberof localized species. The second and tance in respectto the small and ratherround is- moreimportant reason is that Java is less humid lands of Bali and Lombok.

JAVA BALI LOMBOK SUMBAWA FLORES ALOR GROUP

+-3 W15 t12 28-4 I

I I~~~~~~~~~~~11

BALI LOMBOK ALAS SAPE FLbRES KAM1ING STR. STR. STR. STR. STR. STA. 87% 72.5% 68% 63% 57. 5%

FIG. 2. INTER-ISLAND STRAITS IN THE, LESSER SUNDA ISLANDS AND TIEIR EFFICENCY AS DISTRBUTIONAL BARRIERS FOR BUDS Explanation of figues in text. and poorerin habitats than Sumatra,also more This is by no means the only difficultythat is peripheraland thusless accessibleto colonistsfrom encounteredin the attempt to determinethe the Asiaticmainland. The climaticdeterioration, relativeefficiency of the various straits in theSunda which is already indicated in westernJava, ac- chain. It happens that there is a gradual but celeratesrapidly in the easternpart of the island steadychange of climate and plantcover from west where in the lowlands true tropical rain forest to east. Each more easterlyisland is somewhat seems to be largelyreplaced by monsoonforest. morearid thanits westernneighbor and one after The resultis that manyof the mostcharacteristic the otherof the humidity-lovingspecies drops out Java elements(including nearly all of the well- because the habitat becomes unsuitableand not knownendemic's) are restrictedto westernJava. necessarilybecause it can not crossthe waterbar- Of the340 speciesof Java birdsonly 245 are found rierseparating it fromthe next island. in theeastern half of theisland and it is reasonable The effectof six inter-islandstraits on the dis- to believe that some 70 of these species drop out tributionof birds is illustratedin Fig. 2. The beforethe easterntip of Java is reached,leaving top line of figuresrecords the numberof eastern only about 170 species for the easterntip of the species that findthe westernlimit of theirranges island. No naturalhistory survey has ever been on the inter-islandstraits. Lombok Straitshows made of thissection of Java. This is unfortunate thehighest figure with 15 species,but, on thewhole, because the fauna of a small area of easternmost the differencebetween the variousstraits is rather Java, equivalentin size to Bali, must be compared slight. This is not surprising,since all theeastern withthe Bali fauna,if one wantsto testthe signif- species have the ability to jump water barriers

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions WALLACE'S LINE 7 and it is probable that the ecologicalfactors on are, however,found on Wetar, Timor, or other the islandshave as much or moreto do with the more easterly islands, which proves that Flores limitsof the ranges than age or widthof the straits Strait is not the easternlimit of theirrange. betweenthem. The second row of figuresgives The data presentedin Fig. 2 and in Table 2 can the numberof breedingspecies knownfrom each be summarizedas follows:Each of the straitsin island (the secondfigure on Java givesthe number the Lesser Sunda Islands is a zoogeographicbar- of specieson the easternhalf of Java). The third rier. Lombok Strait, however,is more effective row of figures,and thisis the mostimportant one than any of the others. This is apparentlydue to ofall, givesthe numberof westernspecies that are the fact that this straitpersisted throughout the haltedby the variousstraits. The significanceof Pleistocene,whereas Bali Strait and Alas Strait Lombok Strait becomes at once apparent It driedup at theheight of the Pleistocene glaciation. preventsthe passage of 68 (41 per cent) of the 166 Rensch's data (Table 2) indicate that reptiles, Bali species. No other strait approximatesthis amphibia, butterflies,and land mollusks show figure. The last rowof figures gives the percentage similarconditions as birdsand it is possible that ofwestern species on each of theislands. a more thoroughexploration of Bali, Flores, and The relativeefficiency of Lombok, Alas, Sape, Alor would make the two sets of data even more and Flores Straitscan be expressedby calculating similar. Mertens(1930), forexample, who denres what percentageof the speciesoccurring on either that Lombok Strait is a more importantbarrier side are stopped by the straits. Lombok Strait, than Bali or Sape Straits,bases this claim on a forexample, is a barrierfor 83 ( = 68 + 15) species study of the very insufficientlyknown herpeto- TABLE 2 Relativeefficiency of straitsin LesserSunda Islands

AMPHIBIA BUTTERFLIES BIRDS (ORIG.) REPTILES, RANK AND LAND MOLLUSKS (RENSCE) (FOR BIRDS)

LombokStrait (Bali-Lombok) ...... 83 of285 = 29 percent' 84 of377 = 22 per cent 1 Alas Strait(Lombok-Sumbawa) ...... 22 of242 = 9 percent 34,of367 = 9 percent 3 (Sumbawa-Flores) ...... 15 of266 = 5.5 per cent 52 of364 = 15 per cent 4 FloresStrait (Flores-Alor) .28 of218 = 13 percent 2 1This figure indicates what percentage of the sum of the species of the two islands on eitherside of the strait have not crossedthe strait. ofa totalof 285 (166 + 119), thatis, 29.1 per cent. fauna of Bali. De Beaufort (1026) and other The percentagesfor the other straitsand for a studentsof freshwaterfish are also handicapped numberof other animal groups a're given in Table 2. in a discussionof this subject,since withfew ex- The figuresof Table 2 have, of course,only a ceptionsthere are no true freshwaterfish east of relativevalue sincethe species totals include many Lombok Strait. Even Bali has a badly depleted species twice, once east and once west of the fishfauna. That this island is so poor in fresh- straits. Still they are valid as indicatorsof the water fish is not really surprising,sitice it i?,a r,elativeefficiency of these straits and of their volcanicisland and has not a singlelarge and only rank. Many of the smaller islands (Penida, one medium-sizedstream. It wouldbe dangerous ,Sangeang, Rintja, etc.) are insufficientlyto base too manyconclusions on thedistributional exploredand have thereforebeen omittedfrom the data derivedfrom a singlegroup which is so exact- tabulation. Solor, , Pantar, and Alor ing in its ecologicalrequirements as are freshwater have beenunited as Alorgroup. In thetabulation fish. of the bordersof westernspecies only the eastern- Freshwaterfish are usefulas negativezoogeo- most occurrencehas been used. This explainsa graphicindicators. The fact that primaryfresh- seemingdiscrepancy of the figures. Flores Strait, waterfish (see Myers,1938, for a definitionof this for example,stops only 24 of the 143 species on term)are absentfrom Seran and Kei indicates,for Flores. One would expect the Alor group to example,that theseislands have had no continen- have 119 species (143 less 24), but it actuallyhas tal connectionwith New Guinea. The 1presence only 75. The "missing"44 (119 less 75) species of four species of freshwaterfish in the Lesser

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions 8 THE QUARTERLY REVIEW OF BIOLOGY

Sunda Islands-Rasbora elbertion Lombok and Salwati,and Misol) fromthe NorthernMoluccas. Sumbawa; Clarias batrachuson Bali, Lombok,and In thissection the line does not followentirely the Sumbawa; Aplocheilusjavanicus on Lombok; and 100 m. contour,which would exclude Koffiau, A. celebensison Timor-does not necessarilyprove Gebe, Batanta, and Waigeu. However,all these continentalconnections for these, but it casts islandsare so purelyPapuan and formsuch a well- doubt on the means of dispersalof these species. definedfaunistic unit that it seemsjustified to be The slight, or absent, differentiationof these slightlyinconsistent. It mightbe worthwhile to species demandsthat these islands had a recent emphasizethat the line, as just drawn,gives a continentalconnection. However, if such had betterdefined delimitation of the "Papuan main- existedone wouldexpect a muchricher fish fauna. land" and "Papuan island" fauna than does The transportof fishby waterspouts is well sub- Wallace's Line in the west forthe Indo-Malayan stantiated and it is also possible that Lombok fauna. Its validity is particularlyapparent for Strait had occasionalya surfacesheet of fresh- all groups with a limited dispersalfaculty, for waterwhile it was the outletfor the largestreams example freshwaterfish. De Beaufort's map of PleistoceneSundaland. It wouldbe dangerous (1926,p. 103) ofthe range of the subfamily Melano- to go too farin suchspeculations of possible chance taeniinae illustratesit quite graphically. This dispersalbut it is even more dangerousto base Australiangroup extendswestward as far as the sweepingzoogeographic conclusions on the pres- AruIslands and Waigeu,but is absentfrom the Kei ence of a fewspecies of so-calledfreshwater fish. Islands and from the Northernand Southern Moluccas. THE EASTERN COUNTERPART OF WALLACE' S LINE The significanceof this easternline has been emphasizedby a numberof authors. It indicates, It is obvious that theremust be a line at the like Wallace's Line, a major faunalbreak; it sep,- easternedge of the island belt whichcorresponds rates, like its westerncounterpart, a continental to Wallace's Line in the west. Such a line would froman islandzone, as wellas a zone witha moreor separatethe zone of a moreor less pure Australo- less undilutedPapuan faunafrom a mixedPapuan- Papuan mainlandfauna fromthe islands to the Orientalfauna, a contrastwhich is least apparent west wJithan impoverishedPapuan fauna and an in the north. It is forall these reasonsthat this Indo-Malayan admixture. This line has been line must be considereda major zoogeographic vaguelyreferred to by Lydekkerand othernine- boundary. teenthcen'tury writers, but I believe de Beaufort (1913) was the firstto point out its true signifi- SHOULD AN INTERMEDIATE ZOOGEOGRAPHIC REGION cance. It is not difficultto trace since it follows, BE RECOGNIZED? exceptfor a shortstretch in the north,the 100 m. depth line, that is, the edge of that part of the A gradual transitionbetween the Orientaland, Sahul shelfthat was dryland at the heightof the the Australianfaunas takes place in the island Pleistoceneglaciation (Fig. 1). It passes between belt between Wallace's Line and its eastern the Aru Islands,which have a purePapuan fauna, counterpart. This was realized quite clearlyby and the Kei Islands with an impoverishedfauna SalomonMuller (1846), the earliestzoogeographer with Oriental elements. Of birds, for example, of the Malay Archipelago. He lists correctly 166 species are knownfrom the Aru Islands, in- "Celebes, Flores, Timor, Gilolo and perhaps cludingbirds of paradise and many othertypical Mindanao" as islandson whicha mixtureof Indian Papuan types,while only 84 speciesare knownfrom and Australianelements is found. Wallace also, the Kei Islands, includingsome westernelements. in his laterpublications, admitted the intermediate The line thenpasses betweenthe mainland of New positionof thisregion and statedof Celebes that Guinea and Seran Island. There are 115 species it "hardlybelongs to either[Oriental or Australian] of birds(about 30 per cent western)known from region." Pelseneer (1904:1007) lists a whole Seran as against more than 300 species fromthe group of workerswho recognizedthe transitional Vogelkop,the neighboringpart of New Guinea. characterof this region. The linethat separates the Papuan mainlandfauna There are otherfactors, in additionto the lack fromthe island fauna swingsfrom Seran north of continentalconnection, which contributeto- and passes throughthe Gilolo passage separating ward the povertyof the fauna of this island belt. the westernPapuan Islands (Waigeu, Batanta, SalomonMuller (1846) veryably pointedout some

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions WALLACE'S LINE 9

of thereasons, such as thesmall size ofmost of the of Indo-Malayan and Australo-Papuanelements. islands, their low elevation and their aridity. As againstthese points which would favorthe rec- Thereis a moreor less arid corridorextending from ognitionof a transitionregion there are some the Philippinesand Celebes to Buru and to the very strong objections. Pelseneer (1904) has Sunda Islands fromeastern Java to Timorlaut. stated them clearly. He points out that it is This zone has acted as a barrierto manyhumidity- onlynatural that lovingforms and has preventedtheir passage from Sundaland to the Papuan Region or vice versa. a zoogeographicborder is nota linewithout width and Additionalreasons for the faunal povertyof this thatby necessity there is a mixtureof faunal elements zone are the younggeological age of manyof the alongthe border of two zoogeographic regions, caused a islands,which limits the number of chance coloniza- by reciprocalpenetration. But if one wouldadmit for this reason a special tions,and the heavyvolcanic activity over part of 'transitionregion' or a 'regionof intermediacy,'one, the region. There are threelines of volcanoesin wouldobviously double the difficulties ofdelimitation. this transitionzone, one extendingfrom Sumatra For now it wouldbe necessaryto traceboth of the throughJava to theinner Banda arc,a secondone borderlines which separate the transition region from followingthe westernedge of the northernMoluc- eitherof the two adjoining zoogeographic regions. cas, and a thirdone reachingfrom north Celebes throughthe SanghirIslands to the Philippines. These difficultiesof delimitationare fullycon- The volcanic activity is thus strictlylocalized, firmedby the two most recentlyproposed transi- but whereit occursit may be a veryserious factor tion regions. Dickersonet al. (1928:297) define indeed. As mentioned,it seems to be one of the theirsas follows:" is outlinedsharply by reasonswhy Java's animal lifeis so muchpoorer Wallace's Line (as modified)on the west and than that of Borneo or Sumatra (Rensch, 1936). Weber's Line upon the east." It thus includes There are not only 59 young volcanoes of more the Philippines,but it excludes the Moluccas, than 2000 m. altitudeon Java, but also manyex- Timorlaut,and Kei Islands. Rensch (1936:265), tinctlate Tertiaryones. This factoris even more however,includes in his Zwischengebiet"Celebes, evident on Lombok where heavy Pleistocene the Lesser Sunda Islands,Timorlaut (perhaps also eruptionsof Mount Rindjani seem to have de- Kei), and theMoluccas (at least the southernMo- stroyedmuch of the mountainfauna. The same luccas)." He definitelyexcludes the Philippines. is truefor the volcano on Ternate Island (Strese- Celebes and the Lesser Sunda Islands are, thus, mann,1939:381). the only two districtswhich the two transition All the mentionedfactors combine to give the regionshave in common. fauna of the transitionzone a peculiarcharacter. The "degree of intermediacy"of the various This has impressedsome of the authorsto such sectionsof the transition region is veryuneven. It an extentthat theyhave proposedto give formal seems,for example,that the percentageof Aus- recognitionto this fauna and elevate the island tralo-Papuanspecies in the Philippines(which are belt to the rankof a separatezoogeographic region includedin Wallacea by Dickersonand Merrill) or subregion. is smallerthan the percentageof Orientalspecies Dickerson et al. (1928), who coined the term in New Guinea or Australia. Still,nobody would Wallacea forthis region,and Rensch (1936), who want to suggestincluding Australo-Papua in the simplycalls it Zwischengebiet(region of intermedi- transitionzone. acy), are thetwo most recent champions of such an Stresemann(1939:403) adds another weighty arrangement.This region would include four objection. He pointsout that the transitionzone differentgroups of islands, (1) the Lesser Sunda comprisesfour separate districts which have much Islands fromLombok eastward;(2) the Moluccas less in commonwith one anotherthan eaph one and otheroutliers of the Papuan Region (Tenim- has with some outside region:The Moluccas are ber, Kei); (3) the Celebes group (with Sula and faunisticallyclosest to New Guinea, and Celebes Talaut); and (4) thePhilippines. Two reasonsare to the Philippines,but the Philippinesare closer usually quoted in favor of recognizingsuch a to Malaysia than to Celebes. The Lesser Sunda transitionregion. One is, that many endemic Islands, finally,have a close faunal relationship speciesand generaare confinedto it. The other with Java and Australo-Papua,but only a very reasonis, thatall of theislands, which are included slightand recentone withCelebes. To unitefour in thistransition zone, are populatedby a mixture such heterogeneousdistricts in a single "region"

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions 10 THE QUARTERLY REVIEW OF BIOLOGY violates all principlesof regionalzoogeography. Line." Pe'leneer drew the courseof this line owr Afterall, if a zoogeographicregion means any- the basis of non-zoologicaldata, that is, primarily thing,it means the homeof a moreor less homo- on the soundingsand otheroceanographic results, geneous characteristicfauna. "Wallacea", how- of the Siboga Expedition,many of whichare no ever, is the home of four differentfaunas. It longer valid today. However, Weber's Line is self-evidentthat the formal recognition of a zoo- actuallyseparates the islandswith a morethan 50 geographicregion of such heterogeneityis neither percent Indo-Malayan fauna from the islands with. practical nor scientificallydefensible. The term a-more than 50 percent Papuan fauna,as is evident transitionzone is justifiedonly if applied infor- fromRensch's (1936) carefuldata and fromall the mally as a descriptiveattribute. other zoogeographicwork of the region. With insignificantmodification the line suggestedby- WEBER 'S LINE Pelseneeris still acceptable as the best possible borderlinebetween the Orientaland the Australo- It is apparentfrom the preceding discussion that Papuan Regions. Line neitherWallace's nor the formalacceptance The courseof Weber's Line (Fig. 1) is as follows: a zone are of transition satisfactoryattempts of In thenorth it beginsbetween Talaut and Celebes delimitingthe Oriental against the Australian in the westand thenorthern Moluccas in the east. This to Region. leaves, my mind,only one other In this sectionthe line is extremelywell defined,, alternativesolution, namely, the recognitionof a since the fauna of the northernMoluccas consists, line east of Wallace's Line. Beforeattempting to of about 80-90 per cent and that of Celebes of draftthe best possiblecourse of such a line,a few about 20-40 per cent Papuan elements. The line wordsmust be said about the validityof any zoo- continuesfrom here betweenthe Sula Islands in geographicborderline. thewest and Obi in theeast and thenswings around A zoogeographicregion is usually definedas a Buru. The fauna of the Sula Islands is insuf- geographicsubdivision of the earth that is thehome ficientlyknown, but it is close to that of Celebes ofa peculiarfauna. Such a regionis characterized exceptmuch poorer and witha strongerMoluccan, the by presenceof manv endemicgenera and fam- element. Still the Papuan componentis probably iliesand by theabsence of the characteristic genera less than40 per cent,while it is about 63 per cent and familiesof otherzoogeographic regions. Its on Buru and even higheron Obi. It is difficultto bordershould be drawnalong the line wherethis trace Weber'sLine fromBuru on. Pelseneerat- specificfauna is replaced by a differentfauna. temptedto followthe contour of theocean bottom This procedureis logicaland presentsno difficulties and thiscaused himto runthe line between Banda in all the cases wherean efficientbarrier separates ("Indo-malavan") and Seran ("Papuan") and be- the two regions,such as is formedby the South tween Sermatta ("Indo-malayan") and Babber Atlantic between Africa and South America. ("Papuan"). The much more detailed informa- However,an interminglingof the twofaunas takes tion on the fauna of these islands,which is now in place a borderzone whenevertwo such regions available, indicatesthat a differentcourse might comeinto direct contact. This is exactlywhat has be preferable. The youngvolcanic Banda Islands happenedin theisland belt betweenAsia and Aus- have a fauna whichalmost completelylacks en- tralia. Both theIndo-Malayan and theAustralo- demicelements, and whichis veryclose to that of Papuan mainland faunas have spilled over into Ambon,Seran, Seranlaut, etc. There is no doubt the intermediateisland belt and it mightseem im- that the Banda Islands must be includedin the in possibleto delimit sutha mixedregion one fauna southernMoluccas. Babber, on the otherhand, fromthe other one. However,as Pelseneer(1904) has a faunawhich is closerto thatof Dammer and is a says correctly,"it evidentthat theremust be Sermatta,than to thatof Timorlaut. It is, there- line ... withinthe regionof mixture,on one side fore,preferable to place the line betweenBabber of whichthe faunal elementsof one regionprevail and Timorlaut. Rensch (1926:206) has already and on the otherside thoseof the second region. pointedout the impossibilityof separatingBabber This line can serveusefully to markthe borderline fromthe closelyrelated Sermattaand Dammer. betweenthe twobiogeographic regions." The fauna of Timorlautis about 62.5 per cent On the basis of these considerationsPelseneer Australo-Papuan. South West Islands, from establisheda borderlinebetween the Orientaland Roma and Kisar to Dammer and Babber are a the AustralianRegions, which he called "Weber's faunisticunit, but theprogressive decrease of Indo-

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions WALLACE'S LINE 11

Malayan elementswhich started on Java and Bali Other objections have been raised against continueson these-islands. It is possible that a Weber's Line. Some authors,for example,have futureanalysis may showthat the easternelement objected to Weber's Line because it separates on Babber and Dammer is already slightlymore islands which lie on the same submarineridges. than50 per centof the total'faunaof theseislands. Thus it cuts betweenBabber and Timorlaut,be- Even thenI wouldbe inclinedto retainthem in the tweenDammer and Banda, and betweenSula and OrientalRegion rather than to drawa linethrough Obi, each ofthese three pairs of islands lying on the the middleof the SouthWest Islands. same submarineridge. It seems to me that this One glanceat the map showsthat Weber's Line argumentis anotherinstance of confusingzoogeo- is situatedmuch closer to theAustralo-Papuan than graphicand geologicalinterpretations, exactly as to the Asiatic shelf. The reason for thisis two- in the case of continentalversus oceanic islands fold,faunal pressure and accessibility. The faunal (Mayr, 1941). The geology of an island, and pressureof the Indo-Malayan fauna is greaterthan particularlyof an oceanic island,is of no concern that of the Papuan faunabecause it is muchricher whatsoever,when we are attemptingto classify in speciesand families. The sphereof influence of its fauna. If the fauna of Seran and Kei is pre- thisrich fauna will,therefore, extend farther into vailinglyPapuan, I shallclassify these islands with the island belt than that of the poorerPapuan thePapuan region. The factthat Timorand Sumba fauna. The secondreason is that the chainof the witha prevailinglyIndo-Malayan fauna lie on the Lesser Sunda Islands,forming practically a penin- same tectonicarc has absolutelyno bearingon this sula of Sundaland, was infinitelymore easily decision. In fact thereis no evidencethat any of accessibleto colonistsfrom the westthan to those thesearcs wereever raised to the extentthat they fromthe east, whichhad to jump the wide gap wereexposed for their full length, and it is obvious eitherfrom Australia to Timoror fromNew Guinea that the underseageology can have no influence (and Aru) to the islandsof . The pre- over the distributionof formsthat are dispersed ponderanceof Orientalspecies in theLesser Sunda across thewater. Islands would be even more pronouncedif eco- De Beaufort (1926:184) also rejects 'Weber's logical factors(aridity) had not favoredcoloniza- Line fora purelygeological reason, because "the tion by Australianelements. These various fac- Moluccas are not the remainsof a formergreater torsexplain the present course of theline of faunal land mass." This argumentis entirelyirrelevant, balance, Weber'sLine. Wallace's argumentthat not only since Pelseneernowhere makes such a Celebes shouldbe includedin the AustralianRe- claim,but also because the formergeological his- gionbecause it had so fewOriental species as com- toryhas a bearingon zoogeographicclassification pared to Borneo,is beside the point. Every true onlyto theextent to whichit influencespresent day island has, of course,a muchimpoverished fauna, distribution. but its zoogeographicposition is deternminedby an I knowof only a singlevalid argumentagainst analysisof its existingfauna and not by the ele- the adoption of Weber's Line as the boundary ments it lacks. With an 80 per cent Oriental betweenthe Australianand the OrientalRegions. faunaCelebes can notbe includedin theAustralian It is the objectionagainst dividing arbitrarily any Region! continuousseries of values at the halfwaypoint Weber'sLine has foundcuriously few adherents between the extremes. In the case of Weber's among zoogeographers;Boden Kloss (1929) is Line the situationis aggravatedby the fact that one of the exceptions. There is nothing spec- the 50:50 balance betweenthe Indo-Malayanand tacular about this line and by crossingit one the Australianelements is not always the same in encountersa smallerfaunal change than is found the varioustaxonomic groups. The birdfauna of between Borneo and Celebes, or between New Wetar Island, for example,is more than 50 per Guineaand Seran,or in generalbetween the "main- cent Australo-Papuan,while in othergroups the land" and the "island" faunas (Fig. 1). The Oriental elementseems to prevail. On Celebes differencebetween the faunas of Sula and Buru about 67 per centof the birds are ofwestern origin, and of Babber and Timorlautis, indeed, rather whileamong mammals, butterflies, reptiles, amphib- small. Weber's Line is not acceptable to those ians, and land snails it is morethan 80 per cent. who look for a strikinglyconspicuous borderline On thewhole it seemsas ifamong reptiles and but- between the Oriental and Australian Regions terfliesthe western element pushes farther eastward (Rensch,1926:265). thanamong birds and snails. However,taking the

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions 12 THE QUARTERLY REVIEW OF BIOLOGY fauna as a whole,Weber's Line seemsto separate relationshipof the dominantspecies of each habi- rather neatly the islands with a prevailingly tat. It seems,for example, to judge by Rensch's Orientalfauna in the westfrom the islandswith a description(see above) thatthe differences between prevailinglyAustralo-Papuan fauna in the east. the dominantspecies of birdsof Bali and Lombok As stated above, the easternmostof the South is even morestriking than is apparentfrom a sta- West Islands (Dammerand Babber) possiblyhave tisticalanalysis of the total faunas. Such a com- slightlymore than 50 per cent Australo-Papuan parisonof thedominant types of local faunasmust elements,but it is inadvisableto separate them be based on accurate censusdata gatheredin the fromthe largergroup of islandsof whichthey are field and such data are not yet available. To an'integralpart. Lines of50:50 balanceface even gatherthem would be a worthwhile task offuture more difficultieson continentsthan in archipel- explorersof the Malay archipelago. agos. The line in North Americaon which the The combination of ecological and zoogeo- Palearctic and the N,earcticelements balance, graphicmethods promises to yield data of consider- would be entirelyunsuitable as a zoogeographic able interest. It seems,for example, that the faunal boundary. A 50:50 line is, thus, admittedlya compositionof each habitat is different.Of the moreor less arbitraryboundary and may have to eleven species of birds that are restrictedto the be modifiedin special cases. But it is no-more mountainforest of Timor (above 4000 feet) only arbitrarythan to accept March 21st as the first a single one is Papuan, the other ten are Indo- day of spring(regardless of the weather!),or the Malayan. The ratio is even, if not reversed, 21st birthdayas the day on whichan adolescent among the birds of the tree savanna of Timor. reaches-seniority. Such rigid divisions are of Lack of exact ecological data preventsa more practicalusefulness not onlyin humanaffairs, but accurateanalysis at thepresent time. Steenisand frequentlyalso in scientificmatters. Different other botanists have shown that a similar dif- faunal regionsare generallyindicated on zoogeo- ferenceof floristiccomposition exists between graphicmaps by differentcolors. It is obvious differentplant associations. Here is a practically that the 50:50 line is the most convenientplace untouchedfield for future investigators. where to replace one color by another. It is in The delimitationof biogeographicregions de- this sense that Weber's Line (as modifiedabove) pends to a considerableextent on the dispersal may be accepted as the boundarybetween the facultiesand on the natureof the speciationproc- regionwith a prevailinglyOriental and the region essesof theorganisms of whichthe distributionis witha prevailinglyAustralo-Papuan fauna. studied. It has become evidentin recentyears that thereis much differencebetween phytogeo- UNSOLVED PROBLEMS OF INDO-AUSTRALIAN graphic and zoogeographicclassifications. The ZOOGEOGRAPHY majorfloristic regions coincide fairly well withthe The conclusionsat whichI arrivedin thepresent major climaticregions. The majorzoogeographic analysis are not final. Many of the islands are regions,on the otherhand, indicateprimarily the insufficientlyexplored and it is certainthat future extent of formerly(or currently)isolated land explorationwill add a good deal to our knowledge. areas. The biogeographicclassification of New A furtherrefinement in thezoogeographic methods Guineais a goodillustration for this. New Guinea is also expectedto yield increasedresults. Salo- is, forthe phytogeographer,a part of the Malayan monMuller, P.L. Sclater,A. R. Wallace,and other region,but faunisticallyit is at least as close or earlyrepresentatives of the classicalschool of zoo- even closerto Australia. A comparisonof phyto- geography selected arbitrarily a number of geographicand zoogeographicmaps indicatesthat indicatorspecies and based theoutlines of thezoo- it is impracticalat the presenttime to construct geographicregions and subregionson the dis- biogeographicmaps, that is, maps that intendto tributionof these species. The preferredtech- illustratesimultaneously the distributionof plants nique ofthe present paper is to calculatein percent and of animals. the proportionof faunalelement in the totalnum- This is equally true,although to a lesserextent, ber of species of certain localities. All the foranimal groups with different dispersal faculties. percentagesin Fig. 2 and Table 2 are derivedby I have already mentionedabove the differences thismethod. betweenbirds and reptilesin regardto the faunal In the matterof faunal compositionan even compositionof some of the islands. Much more superiormethod might be to determinethe faunal accuratedata are needed. It is possiblethat some

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions WALLACE'S LINE 13 of the invertebratesshow a distributionalpattern Indo-Australianarchipelago (Scrivenor, et al., thatis muchmore similar to thatof plantsthan to 1943) have been publishedafter the completionof that of mammalsor birds. Progressin this field the presentwork. These papers containnothing dependslargely on a more thoroughfaunistic ex- that would require a major modificationof the plorationof theIndo-Australian Region. conclusionsat whichI have arrivedabove. How- amountof in- SUMMARY ever,they contribute a considerable terestingfactual data and raise a numberof ques- (1) Wallace's Line is not the boundarybetween tionswhich I have not treated. Corbet (op. cit.) the Indo-Malayan and the Australian Regions, shows that Weber's Line, at least in its northern but ratherit indicatesthe edge of the area (Sunda part betweenCelebes. and the northernMoluccas, shelf)that was dryat theheight of thePleistocene constitutesa morepronounced faunal divisionin glaciations. several familiesand genera of butterfliesthan (2) The equivalentline along the edge of the Wallace's Line in its most effectivesection (be- separates New Guinea and the Aru tweenBorneo and Celebes). Malcolm Smith(op. Islands fromthe Moluccas and Kei Islands. cit.) comes to the conclusion,on the basis of the (3) Weber's Line separates the islands in the distributionof vertebrates,that Weber's Line is west on which the Indo-Malayan elementis pre- preferableto Wallace's Line, if a singleborderline dominantfrom the islands in theeast on whichthe betweenthe Orientaland the AustralianRegions Australo-Papuan element has a numerical su- is to be found. The botanical contributorsem- periority. phasize the discrepancybetween the classification POSTSCRIPT of biogeographicregions of the zoologistsand of The results of an importantsymposium on the botanists. This disagreementis much less Wallace's Line and on the zoogeographyof the strikingin regardto the minordivisions.

LIST OF LITERATURE

BEAUFORT, L. F. DE. 1913. Fishes of the easternpart MAYR, E. 1941. The originand the historyof the of the Indo-AustralianArqhipelago, with remarks bird fauna of Polynesia. Proc. Sixth Pac. Sci. on its zoogeography. Bijdr. tot de Dierk., 19: Congr. (1939), 4: 197-216. 95-164. MERTENS, R. 1930. Die Amphibienund Reptilien . 1926. Zoogeographievan den IndischenArchi- der Inseln Bali, Lombok, Sumbawa and Flores. pel. VolksuniversiteitsBibliotheek. Vol. 35, 202 Abh.Senckenb. Naturf. Ges., 42: 115-344. pp. Haarlem. MOLENGRAAFF, G. A. ,F. 1922. Geologie. In: De BRONGERSMA, L. D. 1936. Some comments upon Zeen van NederlandschOost-Indie, pp. 272-357. H. C. Raven's paper: "Wallace's Line and the Leiden. distributionof Indo-Australian mammals." Arch. MULLER, SALOMON. 1846. tYberden Charakter der Nterl. zool., 2: 240-256. Thierweltauf den Inseln des indischenArchipels, DICKERSON, R. E., et al. 1928. Distributionof life etc. Arch. Naturg.,12: 109-128. in the Philippines. Bur. Sci. Monogr.,21, 322 pp. MYERS, G. S. 1938. Fresh-waterfishes and West Manila. Indian Zoogeography. Smithson.Rep. for 1937: HAECKEL, E. 1893. Zur Phylogenieder australischen 339-364. Fauna. Denkschr.Medic. Nat. Ges. Jena, 4: 5. PELSENEER,P. 1904. La "Ligne de Weber," limite HUXLEY,T. H. 1868. On the classificationand dis- zoologique de l'Asie et de l'Australie. Bull. Classe tribution of the Alectoromorphaeand Hetero- Sci. Acad. R. Belgique,1904: 1001-1022. morphae. Proc. Zool. Soc. London, pp. 294-319. RAVEN, H. C. 1935. Wallace's Line and the dis- tribution of Indo-Australian mammals. Bull. KAMPEN, P. N. VAN. 1909. De zoogeographievan Amer.Mus. Nat. Hist.,68: 177-293. den Indiscne Archipel. Bijbl. 3 en 4,-Natuurk. RENSCH,B. 1930. Eine biologischeReise nach den Ned. Indie. Tijdschr. Kleinen Sunda Inseln. P. 209. Berlin. The boundaries KLOSS, C. B. 1929. zoo-geographical 1936. Die Geschichtedes Sundabogens. Eine betweenAsia and Australiaand some orientalsub- tiergeographischeUntersuchung. Pp. 1-318 (with regions. Bull. Rag. Mus., 2: 1-10. an exhaustive bibliographyof 326 titles on the KUENEN, PH. H. 1935. Geological interpretationof zoogeographyof the Malay archipelago). Berlin. the bathymetricalresults. In: The Snellius Ex- SARASIN, P., AND F. 1901. Uber die geologischeGe- pedition in the eastern part of the Netherlands schichte der Insel Celebes auf Grund der Tier- East-Indies 1929-1930. Vol. 5, pt. 1, p. 1-124. verbreitung. 169 pp. Wiesbaden.

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions 14 THE QUARTERLY REVIEW OF BIOLOGY

SCRIVENOR,J. B., T. H. BURKILL, MALCOLm A. SNaTH, IndischenArchipel. Tijdsch.Kon.,Nederi. Aard- A. ST. CORBET, H. K. AIRYSHAW, P. W. RICHARDS, rijksk.Genootsch., (2) 49: 769-833. and F. E. ZEUNER. 1943. A discussionof the -, et al. 1934. Gravity,geology and morphology biogeographicdivision of the Indd-Australian of the .In: Gravity Archipelago,with criticismof the Wallace and Expeditionsat Sea, 1923-1932. Vol. 2. Pubi. Weberlines and of any otherdividing lines and Netheri.Geod. Commission. withan attemptto obtainunivormity in the names WALLACE, A. R. 1860. On thezoological geography used forthe divisions. Proc. Linn. Soc. London, of the Malay Archipelago.Journ. Linn. Soc. 154thSess., pp. 120-165. London,4: 172-184. STRESEMANN,E. 1939. Die V6gelvon Celebes. Zoo- geographie.J. Ornith.,87: 312-425. WEBER, M. 1902. Der Indo-australischeArchipel UMBGROVE, J. H. F. 1932. Het Neogen in den unddie Geschichte seiner Tierwelt. 46 pp. Jena.

This content downloaded from 134.173.140.68 on Tue, 3 Feb 2015 17:45:35 PM All use subject to JSTOR Terms and Conditions