Ecology and Evolution of the Sympatric Mascarene White-Eyes, Zosterops Borbonica and Zosterops Olivacea

Home , Borbo, Mascarene Islands, Piton des Neiges

ECOLOGY AND EVOLUTION OF THE SYMPATRIC MASCARENE WHITE-EYES, ZOSTEROPS BORBONICA AND ZOSTEROPS OLIVACEA

FRANK B. GILL

THE Zosteropidae,or white-eyes, are a small paleotropical group of generalizedpasserines that have colonizedmore oceanicislands than any other passerinefamily (Moreau, 1964). While. two speciesof Zosterops rarely coexistin mainlandavifaunas, multiple invasionsof small islands haveresulted in severalcases of bonafide sympatry.One such case involves the endemicMascarene white-eyes, Z. borbonicaand Z. olivacea,which are sympatricon Mauritius and R•union Islands in the westernIndian Ocean. Whereasniche expansion in a reducedcompetitive environment seems to be frequentin islandcolonists (Mac Arthur and Wilson,1967; Keast, 1968), studiesof the Mascarenewhite-eyes, here reported,suggest that the first white-eyeto colonizethe Mascareneislands became more specialized, rather than moregeneralized, in the absenceof competitors.A secondsuccessful colonizationby a generalizedwhite-eye and subsequentsustained coexistenceof the two specieswas apparently facilitated by the specializationof the first arrival.

THE MASCARENE ISLANDS The two major Mascareneislands, Mauritius and R•union, lie approxi- mately 400 miles east of Madagascar and are separated from each other by only 90 miles (Figure 1). Mauritius includes720 squaremiles compared to R•union's 1,00.0square miles, and reachesa maximumelevation of only 3,041 feet while R•union rises to 10,069 feet, the highest point in the Indian Ocean. R•union,the morerugged of the two islands,is composedof two deeply eroded overlappingshield volcanos(Upton and Wadsworth, 1966), one of which (Piton de la Fournaise)is still active. In comparson, Mauritius is a low islandwith rolling,accessible hills. The locationof these islandsat the southernedge of the tropicsinsures a relativelystable climate, whichis dividedinto a warm rainy seasonlasting from Decemberthrough April and a coolerdry seasonfrom May throughOctober. Both Mauritius and R•union were originallycovered by evergreenforest composedof a variety of hardwoo,dspecies, but the presenceof many val- uable timber speciessoon promoted the wholesaledestruction of these forestsby the early colonists.The indigenousforest on Mauritius is now restrictedprimarily to less than 7,0,00acres in the so,uthwesternplateau region (Vaughan and Wiehe, 1937), while considerably more native forest has persistedintact in the ruggedinterior of R•union. One is never

35 The Auk, 88: 35-60. January 1971 36 FRA•: B. GILL [Auk, Vol. 88

• I 45 E 50E 55 E 60

f• Comoro

Mauritius !os

R•union

-25 S

IND AN OCEAN

Figure 1. Location of Mauritius and R•union Islands in the western Indian Ocean. very far from conspicuoushuman disturbanceand the encroachmentof exoticplant specieson either island. Open tree savannasonce replaced the evergreenforest on the dry westerncoasts of both islands,but thesewere completelydestroyed by the early 17th century (Rivals, 1952). R•union's extensivehighlands above 1,800 m supporta relatively undisturbedheath ( Philippia) vegetation. The Mascareneislands are perhapsbest known to ornithologistsfor the wholesaleextinction of a bizarre avifauna (see Hachisuka, 1953; Green- way, 1958) that resultedwhen the islandswere colonized. Over 50,per cent of Mauritius' known original avifauna and 40 per cent of R•union's are now extinct. Recent discussionsof the remainingendemic landbirds are availablein Rountreeet al. (1952), Newton (1958) for Mauritius, Milon (1951) and Berlioz (1946) for R•union, and Watson et al. (1963) for both islands. Jan. 1971] MascareneWhite-eyes 37

FIELD WORK The field data and specimensforming the basisfor this report were obtainedduring expeditionsto the Mascareneislands in 1964 and 1967. In 1964 I was on Mauritius from August through Octoberand on R6union for one week in August and all of November. Robert W. Storer, my field companionin 1964, contributedto the infor- mationgathered at that time. In 1967I spent9 monthso.n R6union studying variation in Zosteropsborbonica, and in the courseof this work made frequent observationsof both R6unio.nwhite-eyes. Finally, someinformation has been obtained from captive Zosteropsborbonica and Zosteropsolivacea (olivacea) maintained on R6union and at The University of Michigan. Data and specimenswere collectedat 102 different localitieson Rdunion,but most of the information discussedin this paper was obtained at six principal study areas. As theseare mentionedrepeatedly below, each is describedbriefly: Beaufonds(Ste. Marie).--Altitude 200 m, annual rainfall 2,425 ram. Located on the island'snorth side on the sugar cane estate of Armand and GeorgesBarau, this study area consistsof a private gardenwith a variety of tropical fruits, flowers, and vegetables,and an adjacent ravine with a flowing stream and densevegetation, primary Eugeniajambos (Myrtaceae). Both white-eyeswere most plentiful on the north and east sides of the island in such ravines. Field work at Beaufonds consisted of several visits from August through November. Etang Sal• les Bains.---Altitude60 m, annual rainfall 713 mm. This lowland tract is on the island's hot, dry west side. Work was conductedin the extensive,dry open forestswhich are under the supervisionof the Service des Eaux et For6ts. In addition to Casuarinaand Pithecellobiumdulce (Leguminosae)important plants include Melia azedarach (Meliaceae), Lantana camara (Verbenaceae),and Schinusterebinthifolius (Anacardiaceae).Field work consistedof regularvisits from July to November,and the color-bandingand releaseof 41 borbonicain July for subsequentobservation. Z. borbonica was plentiful here, but no olivacea was ever seen. TdvelaveHeights.--Altitude 1,650m, annualrainfall 2,000mm. This highlandlocality is directly abovele T6velave (altitude 1,000m) at the Ligne Dominiale on the island's west sideand upland from Etang Sal• les Bains. The forestshere consistprimarily of young, regeneratedTamarin-des-hauts (Acacia heterophylla) with an understoryof heath (Philippia) and bracken fern (Pteridium aquilinum). ScatteredHypericum are also present. Observationsand collectionswere made at this locality in May and from July throughNovember. Both borbonicaand olivaceawere common. Cilaos,For•ts du Grand Matarum.--Altitude 1,400 m, annual rainfall 2,666 mm. This tract lies 1 km north of Cilaos at the base of Piton des Neiges. The forest is indigenous "bois de coleur,"categorized as "moderatelywet forest of middle altitudes"by Rivals (1952); it containsmany hardwoodspecies, including Dombeya spp., Nuxia verticillata (Loganiaceae),Eugenia cotinifolia (Myrtaceae), Hypericum spp. (Hypericaceae),and someintroduced Cryptomeria (Taxodiaceae). Epiphytes,mosses, orchids, and lichens cover the branches of the older trees. This was the most important study area for information on the feedingecology and relationshipsof the two white-eyes. A total of 53 borbonica and 15 olivacea were color-banded and released here. For•t du B•bour.--Altitude 1,400 m, annual rainfall 4,704 mm. This study plot is in one of the wettest forests on the island. The indigenousforest, "high altitude wet forest" of Rivals (1952), includesmost of the important speciesfound at Cilaos and many additional ones,the most conspicuousbeing numeroustree ferns (Cyathea). 38 Fe.A•x B. GILL [Auk, Vol. 88

Other important plants include Forgesia borbonlca (Saxifragaceae),Ardisia (• Icac- orea) spp. (Myrsinaceae) and introduced Fuchsia cocclnea. In April and May, 44 borbonlca and 9 olivacea were color-banded and watched regularly. The plot was then not revisited until November and December when additional studies were made for comparison.Both white-eyeswere plentiful here, ollvaceaoutnumbering borbonlca. Nez de BoeuJ (Plaine des CaJres).--Altitude 2,060 m, annual rainfall 2,000 mm. This high altitude study tract in the ericaceousheath zone overlooks the ravine of the Rivi•re des Remparts. Interspersedin the extensiveheath vegetation,in the hollows and other wind-protected sites, are dense clumps of the endemic Sophora nitida (Leguminosae) which, when flowering in August, are an important food source for white-eyes. Hypericum, which blooms shortly after the Sophora and throughout the summer, is also abundant in this region. White-eyes of both specieswere numerous. A total of 33 borbonica and 6 olivacea were trapped and banded here in August, and watched regularly for flower feeding habits and species interactions from August through November.

TI-IE MASCARENE WHITE-EYES Nontenclature.--Forthe purposeof this paper,which is a discussionof the ecologicalrelationships of the sympatricMascarene white-eye species, I use the nomenclatureof Moreau (1967) as follows: Zosteropsolivacea olivaceafor "oiseau(zo zo) vert" of R6union,Zosterops olivacea chloro- nothosfor the "oiseaupit-pit" of Mauritius,Zosterops borbonica borbonica for the "oiseaublanc" of R•union,and Zosteropsborbonica mauritiana for the "oiseaumanioc" of Mauritius. The nomenclatureof Mayr (1967) is followedfor "Indo-Australian"species. Distribution and behavior.--Unlike the other remaining Mascarene en- demics,which seem to dependon the remnantsof indigenousforest, Z. borbonicais an abundantand ubiquitousspecies on both Mauritius and R6- union. Althoughit occursin nearly all kinds of vegetation,it is primarily a speciesof openand disturbedhabitats rather than intact originalforest. It is a commonbird throughoutMauritius and the lowlandsof R•union wherehuman influence is inescapable,but it is scarcein the relativelyundis- turbed central forestsof R•union except around the edgesof clearings, streambeds,and roadsides. Like most white-eyes,borbonica is a conspicuouslysocial species.On both Mauritius and R•union it is usuallyencountered in partiesof 6 to 20 individualsthat move cohesivelyand noisily throughthe forest, feeding, bathing, and resting together. These flocks move in a follow-the-leader stylewith conspicuousdisplay of the white rump patchin flight. Suchflocks are formed throughout the year. In at least some white-eyes,e.g.Z. lateralis,flocking is a winter phenomenononly and is replacedby territorial defenseby matedpairs during the breedingseason (Fleming, 1943). The greenMascarene white-eye, Z. olivacea,is decidedlyless common on both islandsthan is Z. borbonica. Z. o. chloronothosis now quite rare. It Jan. 1971] Mascarene White-eyes 39 usedto occurregularly in the gardensof Vacoasand Curepipeon Mauritius but now doesso only infrequently(Jean Vins.on,pers. comm.) and is restricted primarily to the remote forestsand scrub of the southwestern plateau.Here it mayactually be morecommon than is generallyrealized, as it is easilyoverlooked unless one is familiar with its call note; with a little experienceone can usuallyfind severalZ. o. chloronothosin the courseof a day's hiking. In 1964 they were most commonnear Mr. Cocotteand PlaineChampaigne, where I foundup to 10 individualsin a singleday. Z. o. olivaceais a much more commonbird than is Z. o. chloronothos, but it is rarely as abundantas Z. b. borbonica.Its distributionfrom sea level to above 2,300 m altitude excludesonly the dry westerncoastal regions; neverthelessit does occur in the wetter forests above 900 m on this side of the island. Z. o. olivacea is also rather uncommon in the lowland forests in the south of R•union but becomes more numerous on the northern and easterncoasts where it lives primarily in woodedravines with flowering Eugeniajambos. It is mostplentiful and outnumbersZ. borbonicain the wet highlandforests of the interiorwhere introduced Fuchsia grows in profusion. Its distribution in the high altitude heath zone follows that of protectedpockets of Sophoraand Hypericum. Unlike borbonica,olivacea doesnot appearto be an edgespecies; it rarely wandersinto gardensand is oftenfound in heavyforest in the vicinityof certainfavored flowers. Z. o. olivaceais an aggressive,asocial white-eye that occursin isolated pairsthroughout the year rather than in small flocks. Rarely are several individualstogether without some interaction; in areaswith many flowers muchof theirtime is spentin aggressivechasing and chattering, for they are highlyterritorial throughout the yearin relationto flowers. Estimatesof total populationsize of thesespecies on Rdunion,based primarilyon roughestimates of populationdensity established by capture- recapturetechniques at Cilaosand Etang Sal• les Bains, are 556,000 for Z. b. borbonicaand 154,000 for Z. o. olivacea. The borbonicafigure is probablyan underestimateas minimaldensities. were usedpurposely in the calculations,whereas the olivaceaestimate may be somewhathigh. Z. b. borbonicaand Z. o. olivaceaboth breedduring the australspring in late Septemberthrough November and December.I foundone nest of olivaceaunder construction as earlyas 10 August,but whetherthis species normallystarts nesting earlier than borbonicais not dear. Z. b. borbonica is conspicuouslysocial throughoutthe breedingseason, adjacent pairs nestingas dose as 50 feet, sometimesin loose semicolonialgroups. "Helpers" sometimesfeed the young. There is no indicationof territoriality or nestarea defensein this species,which thus may resembleZ. 40 FRANK B. GILL [Auk, Vol. 88

Figure 2. Sytnpatric pairs of Mascarenewhite-eyes. A, R•union: upper, Z. o. olivacea; lower, Z. b. borbonica. B, Mauritius: upper, Z. o. chloronothos;lower, Z. b. mauritiana. virens of South Africa (Broekhuysenand Winterbottom, 1968). Z. o. olivacea,on the otherhand, breeds in isolatedpairs dose to productive floweringplants and defendsthe immediatearea against all otherolivacea. Z. borbonicaand Z. o. olivacea both huddle and engagein mutual preeningor "allopreening"as hasbeen noted in otherspecies of white-eyes (Harrison,1968). This is observedfar morefrequently in Z. borbonica than in Z. o. olivacea,in whichit is restrictedto occasionalsessions between (mated?)pairs. In my aviarieswhen both species were present, olivacea usuallyhuddled and preened with conspecificindividuals, but interspecific huddlingand preeningwere not uncommon. Plumagecoloration.--Z. borbonica differs markedly from typicalwhite- eyesin plumagecolor, for insteadof beingyellow-green it variesfrom dull gray to rich brown,and it lacksthe conspicuouswhite eye-ringof most Zosterops.Furthermore it has three groupsof white feathers,the upper tail coverts,the axillaryfeathers, and the crural feathers(ankles), which are often effectivelydisplayed in combinationagainst the rest of the plumage.The white rumppatch in particularis unusualin Zosteropidae. Color variation in Z. borbonicais pronouncedand involvesthe presenceor absenceof brownpigments (phaeomelanins) as well as variationin the intensitiesof both gray and brownpigmentation. A detailedanalysis will be publishedelsewhere; suffice it to sayhere that in colorZ. b. mauritiana, an all-grayform, is similarto oneof the colorvariants of Z. b. borbonica. Jan. 1971] Mascarene White-eyes 41

AlthoughZ. olivaceais grayer than any continentalAfrican speciesand lacksa yellowthroat, it conformsto the basicwhite-eye color pattern in being conspicuouslygreen above with yellow or yellow-greenupper and undertail covertsand in having a large, contrastingwhite eye-ring. In both subspeciesthe lower back, rump, and upper tail covertsare green. The napeand upperback of Z. o. olivaceaalso. contain considerable green and thus differ from the predominatelygray upperpartsof Z. o. chloronothos. The crownof olivaceais usuallyblackish and contrastswith the restof the upperparts,while in chloronothosit is gray and continuouswith the colorationof the restof the head,nape, and back. Only occasionally are there tracesof black or tingesof greenin this form. Both subspecies haveblackish lores. Belowthe two differ with respectto. 1) the intensity of gray on the belly, chloronothosbeing much lighter than olivacea;2) thecolor of the flanks,which have a strongbrownish wash in chlorono'thos; and 3) the presenceof greenfeathers along the midlineof the belly of most olivacea but of no chloronothos. Both yellowpigmentation and the white eye-ringappear to be rather flexibleevolutionary characteristics in the Zosteropidae, for theyhave been lost repeatedlyin insular forms (Moreau, 1957). The white rump of Zosteropsborbonica is probablya derivativeof the pale rump color of many Zosterops(Moreau, op. cit.) and the axillary tufts are only slightly more elaboratethan thosefound in someother species,for exampleZ. lateralis. Soft part coloration.--Infreshly taken specimens of bothZ. borbonica andZ. o. olivaceathe irides are brown, the feetdark gray, and the bill black with a gray baseto the lower mandible. Z. o. chloronothosdiffers from Z. o. olivaceain havinglight brownfeet and a light brownbase of the lower mandible. Size.•--Z.borbonica is a normalwhite-eye in sizeand proportions, though thebill tendsto beslightly weaker and more pointed than in typicalZoster- ops (Moreau, 1957). Wing lengthsaverage 55 mm in the lowlandsand increasewith altitudeon R•union2to an averageof about58 mm. The averagetail/wing ratio of thespecies is 0.74. Bill lengthsaverage 13.6 mm. There are no significantmensural differences between Z. b. borbonicaand Z. b. mauritiana. Z. o. olivaceaand Z. o. chloronothoshave unusuallylong bills, which average15.6 and 16.8 mm respectively.Their bill/wing ratios are higher than thoseof any otherAfrican white-eyes and are exceededby only a few specieson smallPacific islands (Moreau, 1957). Z. o. chloronothosis also unusuallyshort tailed (tail/wing ratio = 0.64), whereasZ. o. olivaceais

Summariesof measurementscan be found in Moteau (1957) and Storer and Gill (1966). Detailed analysis of mensuralvariation in Z. borbonicawill be published elsewhere. 42 Fv,As• B. G•x [Auk, Vol. 88

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 SECONDS Figure 3. Spectrograms(using wide bandpassfilter and normal pattern) of call notes of Zosteropsbo•bonica. Top row: two patterns of si,ngle-noted"eeee" from feedingZ. b. borbonlca.Second row: two patternsof double-noted"eeee-eeee" from feedingZ. b. bo•bonica.Third row: •'chipeeee" from agitatedZ. b. borbonica.Bot- tom row: three patternsof "p-treee"call from feedingZ. b. mau•itiana. normalin this respect(tail/wing ratio = 0.71). The two subspeciesof Zosteropsolivacea differ markedlyin their measurements,Z. o. chloronothosbeing shorter winged, proportionatelymuch shorter tailed, and slightlylonger-billed than Z. o. olivacea. Both racesof Z. olivaceaare, of course,longer-billed than the Z. borbonicawith whichthey are sympatric.The bill lengthratios of the two Jan. 1971] MascareneWhite-eyes 43

.2 .4 .6 -8 1.0 1.2 1.4 1.6 1.8 SECONDS

Figure 4. Spectrogram(as above) of portion of eveningsong of Z. b. borbonica.

pairs of sympatricspecies are 1.2 for Z. o. chloronothos/Z.b. mauritiana and 1.1 for Z. o..olivacea/Z. b. borbonica.Also chloronotho,is slightly smaller (wing length 3 mm shorter) than mauritiana. Z. o. olivaceais slightlylarger than most Z. b. borbonicabut hasa winglength comparable to largeborbonica of theR•union highlands. Vocalizations.--Thevocalizations of the Mascarenewhite-eyes differ betweenspecies and alsobetween conspecific populations on the different islands.The commonfeeding and flock call of Z. b. mauritianais a soft, inflected,and slightly warbled "p-tree" or "p-tee" whereasthe call Z. b. borbonicauses in similarsituations is a plaintive"eeee" or double"eeeee- eeee"(Figure 3). Whenagitated, borbonica gives emphatic chips followed by a piercing"eeeeeee," a call that resemblesthe "p-treee"of mauritiana (Figure 3). Chipsare occasionallygiven by mauritiana,and individuals of bothpopulations sometimes use rapid series of chipsin flight. In late eveningand againfor a few minutesjust after daybreakduring the breedingseason on R•unionfrom late Augustthrough December, male Z. b. borbonicarepeatedly deliver an elaborateloud warblingsong (Fig- ure 4) from a tall exposedperch, such as a Casuarinatree, heath bush,or eventelevision antenna. Occasionallyfull songis heardlater in the day, and foragingindividuals sometimes render short warbled phrases. So conspicuousand locally well-knownis the singingon R•union that it is surprisingnot to hear it on Mauritius. I neverheard this songfrom Z. b. mauritianadespite numerous visits to the forestin late afternoonin Sep- temberand October,and when asked about the possibilityof sucha vocalization, Jean-Michel Vinson, an experiencedlocal ornithologiston Mauritius, was also quite certain he had never heard it. Thus it would seemthat conspicuousdawn and eveningsinging is not characteristicof Z. b. mauritiana,but is of Z. b. borbonica. Z. b. borbonicaalso. has a differentmuted song, a catbird-like(Mimidae) assemblageof sweetwarbled phrases, chip notes,sputters, and squeaks. 44 FR.•x B. GXLL [Auk, Vol. 88

0 I i t i j [ i i I i i I I i [ t .1 .2 .3 .4 .5 .6 .7 .8

SECONDS Figure 5. Spectrograms(as above) of call notes of Zosterops olivacea. Top row: series of four emphatic "tchip" notes of Z. o. olivacea. Middle row: series of "tu" notes from active Z. o. olivacea. Bottom row: seriesof "pit" notes from Z. o. chloronothos.

This songis heard frequentlyat any time of day as the white-eyescome into breeding condition in August and September. Finally as they are preparingto roost,Z. b. borbonicagive a distinctiveextended soft twitter- ing. I could predictably and immediately elicit this vocalization from isolatedcaptive individualssimply by turning off the lights late at night. The vocalizations of Z. o. olivacea and Z. o. chloronothos are also dis- tinct, both from eachother and from thoseof the Z. borbonicawith which they are sympatric.The only call I ever heard from chloronothoswas an explosive "pit-pit-pit" (Figure 5), the basis of the local creole name Jan. 1971] Mascarene White-eyes 45

"oiseaupit-pit." Particularly in flight thesesame notes are often repeated in rapid successionto form a prolongedsputter. Both thesecalls are similar to the "pit" call note and flight sputter of Foudia rubra (Ploceinae), another Mauritian forest endemic. The call usually renderedby active Z. o. olivaceais a single"tsip" or a seriesof "tu-tu-tu" notes (Figure 5). When flying or excited,as well as in aggressivesituations, these white-eyes give an emphatic"tchip-tchip-tchip," the componentsof which are struc- turally unlike the "pit" notesof Z. o. chloronothos(Figure 5). When two or moreZ. o. olivaceaare together,they utter loud, excitedbursts of "tu" and "tchip" notes. Finally, Z. o. olivaceahas a loud, warbledsong similar to the muted songof Z. b. borbonica,but it is alwaysmuch louder and usually has "tu" and "tchip" notes interspersedamong the warbled phrases,one of whichis very similarto the principalphrase of borbonica's eveningwarbling. This songcan be heard throughoutthe day; there is no burst of singingin the late eveningor early morningas in borbonica. The commoncall note or contactnotes many white-eyesuse as they forageis a plaintivesparrow-like chirp or cheep,which often differsvery little betweenwidely separatedspecies. To my ear the "p-teee" of Z. b. mauritianais very similar to the contactcalls of Z. japonicain Hawaii and Z. lateralisin New Zealand,both of which have been describedas "creee" (Fleming, 1943; Peterson,1961). The contactnotes of Z. virensand Z. senegalensismust alsobe very similar in that they are closeto. Z. lateralis (Mees, 1969: 332). The only describedcall notesthat mightresemble those of Z. olivacea(either population)are the staccato"tsip-tsip-tsip" of Z. japonicasimplex (Mees, 1957: 129) and perhapsthe "splutter" of Z. virens (Skeadand Ranger,1958). Many white-eyespecies are known to have elaboratemelodius songs. The "warbling song" of Z. lateralis (Fleming, 1943) is similar to the muted breedingsong of Z. borbonicain context,timing in relation to the breedingseason, and composition.The typical daytime songsof Z. o. olivaceaand of Z. virens(Skead and Ranger,1958), with their rambling and mimickingqualities, also belong to this set of presumablyhomologous vocalizations.Restricted dawn and/or eveningsinging has been reported for relativelyfew species,e.g.Z. virens(Skead and Ranger,1958), Z. cey- lonensisand Z. japonica (Mees, 1957). The "challengesong" of Z. lateralis (Fleming, 1943) is probablyhomologous, but can be heard throughoutthe day, presumablyfunctioning in territoriality. Discussionof evolutionarydifferences in vocalizationsmust, of course, remain highly speculativewithout detaileddocumentation and comparison of many speciesrepertoires. However the vocalizationsof Z. borbonica, especiallyits call notes, are similar to those of virens and lateralis. The componentsof Z. b. borbonica'svocal repertoire seem to be the same as in 46 FR•sX B. Gtr• [Auk, Vol. 88 lateralis. The vocalizations of Z. b. borbonica and Z. b. mauritiana have divergedto the extent that their contactcalls are different, and mauritiana has lost the eveningwarbling (challengesong). More detailed study o.f mauritiana may reveal other differences.The call notesof Z. o. olivacea and Z. o. chlorono.thosare quite different from thoseof typical white-eyes, and Z. o. oliva,cealacks a specialdawn song (challengesong). More information on Z. o. chloron.othos' vocalizations is needed before detailed com- parisonwith Z. o'.olivacea will be meaningful. Feeding.--Z. borbonicatypically feeds in small flocksof 6 to 20 individualsthat, on Mauritiusin Augustand September,often include some of the other endemicforest speciessuch as Foudia rubra (Ploceidae),Co.quus (= Coracina) typicus (Campephagidae),and occasionallyZ. o. chlorono.thos.Z. borbonicais alwaysthe most numerous"nucleus" species, the othersappearing to join the flock but temporarily. On R•union interspecific mixingis lesscommon, Tchitrea borbon'niensis(Muscicapidae) being the only regularassociate of white-eyeflocks. Z. o.. olivace'ararely joins suchflocks and is normallyquite independentin its feedingbehavior. Z. borbonicausually forages like a warblerfor insects,which it captures as they flush, extractsfrom crevicesin the bark, leaf buds,and epiphytes, or gleans from the undersidesof leaves and branches. Titmouselike hangingupside down is common,and Z. b. borbonicaoften creepsalong tree trunks and branches,both horizontal and vertical, in a manner quite like that of the Black-and-whiteWarbler (Parulidae-Mnio.tilta varia). Only oncehave I seenborbonica fly-catching, one eveningin 1964 when severalZ. b. mauritiana flew repeatedlyfrom a sapling into. a cloud of small midges. The insectfood borbonica takes includes flies, beetles, moths, pupae, and eggs. Although these white-eyesare frequently seen taking caterpillars, few stomachsexamined contained them. Similarly spidersare taken regularly, but they do not form the bulk of borbo,nica'sarthropod diet. Z. b. mauritianaoften feedson Flatopsisnivea (Hemiptera,Flattidae), which sometimesoccur in large concentrations(J. Vinson, pers. con:an.). Z. b. borbonicafeeds heavily on white flies (Homoptera-Aleyrodidae)that are abundanton the undersidesof brackenfern (Pteridium aquilinum) fronds in the understoryof young tamarin (Acacia heterophylla) forests. Birds in captivity consumehouse flies and Drosophilain large quantities. Z. borbonicaalso eats a variety of fruits, particularly during the winter months when insects are relatively scarce, and is reported to damage fruit in private gardensand commercialorchards in the lowlands. Of the 357 Z. b. borbonicastomachs examined between May and August,44 per cent containedfruit, comparedto only 11 per cent of 333 stomachsexamined betweenSeptember and December,the normalbreeding season of the Jan. 1971] Mascarene White-eyes 47

TABLE 1 FLOWERS VISITED BY R•UNION WHITE-EYES

Plant species Z. b. borbonica Z. o. olivacea

Fuchsia coccinia (Onagraceae) x x Camellia sinensis (Theaceae) x Kalanchoe bryophyllurn (Crassulaceae) x Eugenia jambos (Myrtaceae) x x Melia azedarach (Meliaceae) x Lantana camara (Verbenaceae) x Pelargonium capitaturn (Geraniaceae) x Musa sp. (Musaceae) x Cuphea platycentra (Lythraceae) x Geniostomasp. (Loganiaceae)• x Aphloiatheae]ormis (Flacourtiaceae) • x Forgesiaborbonica (Saxifragaceae)• x Dombeyaspp. (Malvaceae)• x x Hypericumspp. (Hypericaceae)• x x Sophoranitida (Leguminosae)• x x

Native to R•union Island. species.Fruits that Z. b. borbonicaeats include the berriesof Lantana camara,Schinus terebinthifolius, Solanum sp., Cordylinesp. (Liliaceae), Eugeniacontini folia, A phloia theaejormis, papaya ( Carica) , guava( Psid- ium), and raspberries(Rubus). Suchunusual items as flowerstamens, a tiny snail,and heath (Philippia) buds were found once or twicein the Z. b. borbonica stomachs. Z. b. borbo.nicaregularly visits and extractsnectar from flowers,al- thoughthe amount of suchfeeding depends on the locality and time of year. Wherecertain flowers grow in profusion,as Kalanchoebryophyllum does at EtangSal• les Bainsin July, Eugeniajambos at Beaufond(Ste. Marie) from Septemberto December,Fuchsia coccinia in the For&tdu B6bour fromSeptember to April,and Sophora nitida at Nez desBoeuf (Plaine des Carres)in August,flower feeding is common.In the highlandsthe white- eyes'foreheads often turn yellowwith accumulatedpollen. Table 1 lists the flowerspecies at whichI saw borbonicafeeding. In the highlands Sophoraand Hypericum are most important. In the wet forestsDombeya treesare visited frequently and may be especiallyimportant in that, unlike mostof the nativeplants, some Dombeya is in flowerthroughout the year. An indicationof the importanceof nectarrelative to insectforaging and fruit eatingcan be obtainedfrom the countsI madeof feedingindividuals in the Cilaosstudy area from Septemberto December.These counts were obtainedby recordingwhat eachwhite-eye was doing when first noticed as I walkedslowly through the area. Flowervisits constituted 6.5 per cent (8 of a total of 122observations). Fruit feedingwas observed only 5 per centof the time (6 observations).Thus flowerfeeding is aboutas frequent 48 Fs^•x B. G•r [Auk, Vol. 88 as fruit feeding during the breeding season,perhaps more so, but it is much less frequent than insect foraging,which comprised80.5 per cent o.f the observationsrecorded in this period. To get at the nectar of Fuchsia flowers the white-eyespierce the base of the corollawith their bills. Usually they work on the flowerswhile danglingon a cluster of three or four blossoms,but they often reach out and pull the desiredbloom to a branch with the bill, pin it to the perch with one foot, and then punctureit. Alternatively they simply reach out from a perch with one foot and hold the flower in position. They use similar techniqueson the Kalancho'eflowers. The basic elementsof borbonica'sfeeding ecologyseem to be the same on Mauritius and R•union, but perhapsbecause of the scarcity of flowers in the Mauritian forest,I rarely saw Z. b. mauritiana feedingat flowers; once near the beach at Le Morne several visited the red flowers of intro- ducedflame trees (Erythrina) in the companyof SpottedWeavers (Plo- ceuscucullatus). Unlike Z. b. mauritiana, Z. o. chloronothosrarely foragesfor insects amongthe leavesand branchesof the forest vegetation,but rather restricts its feeding activities to flowers. Single individuals or pairs constantly moveabout and visit isolatedflowers, often coveringsubstantial distances between flowers. This behavior is partly responsiblefor the apparent scarcityof this species,for they appearsuddenly, feed briefly at the few flowerspresent, and then leave. The paucity of flowersin August and September,even in the vicinity of Mt. Cocotte and the Macabe Forest, the strongholdsof this species,may forcethem to visit isolatedproductive flowers. Two of the flowering speciesat which chloronothosfed were Bakerella hoyifolia and Tro.chetiauniflora. In addition to nectar these white-eyesdo take someinsects, which I foundin severalstomachs. Once I saw an individual insert its bill between the overlapping petals of a Trochetia flower and spreadthem apart slightly. The only chloronothos I saw feedingat a nonfloweringsite was a singleindividual that dangled briefly from a tree fern (Cyathea) frond and then flew to a secondtree fern where it probedat the baseof the crown. Z. o. olivaceaalso seemsclosely tied to and dependentupon available flowers. Rarely does one find this white-eye far from some conspicuous flower, and it is truly plentiful only near a profusionof such favorites as Hypericum, Soph.ora,and Fuchsia. In addition to the usual white-eye favorites, o.livctceavisits a wider variety of flowers than does borbonica (Table I). Nectar is almostcertainly the primary attraction, thoughsome insectsare also taken. The flower parts themselvesare sometimeseaten, as I found stamensin the stomachsof three specimenscollected in the Jan. 1971] Mascarene White-eyes 49

Cirque de Salazieand noticeableamounts of pollenin two other stomachs. Several times I saw o.livacea hover in front of a flower. Nearly all Z. o'.olivacea stomachs examined contained some insect mate- rial but I rarely noted this form gleaningthe foliagefor insects.In 9 monthson R•union I accumulatedonly 20 sightingsof insectforaging by olivacea, though such behavior was recordedwhenever noted. I never saw olivaceaforaging in the creeperlikefashion so often usedby borbonica. Z. o. olivace'ais an adeptflycatcher. In November,1964, Storer and I watched several individuals fly-catching from telephonewires along the road near St. Rose. This may be unusual,as I did not seeit in 1967, but I did frequentlysee them fly-catchingfrom tall pine trees. Probablythe mostfrequent mode of fly-catchingis in floweringbushes and sap trees (see below)where aerial insects are capturedas the white-eyeshop from one branchto another.Insects may be grabbedwhile the bird is in the air, or often just snappedout of the air without leavingthe perch. Bill clicksare audibleat closerange. Similarlyin my aviary olivaceacaught Drosophila by flying out from a perch,whereas borbonica usually caught those flies that had landed. Z. o. olivace•anever or only very rarely eats fruit. I neversaw it do so in the fieldor fo.undfruit in any stomachsexamined. In captivityolivacea ignoredvarious fruits that borbonicaate readily. An apparent offshoot of nectar feeding by the R•union white-eyes is their habit of drinkingsap. In early Septembernear CilaosI founda young Claoxylonglan'dulosum (Euphorbiaceae) that was being usedas a source of sap. At many places along the branches,particularly the younger, green ones,the bark had been strippedaway from the underlyingwood, and sap was oozing from the exposedbroken edges. Each stripped area was extensive,o.ften up to.6 incheslong, but the sap was flowingonly from a relatively small unhealedsectio.n. To me this sap tasted rather bitter, but the tree was tended almost constantlyby one banded olivacea,which moved fro.m one break to another in quick successionand drank the freshly accumulatedsap dropletsby running its slightly openedbill along the oozingedge and flicking its tongue in and out. This particular white- eye actively defendedthe tree againstall other white-eyesof both species that came near. That this was not just an unusualcase of sap feedingwas confirmedthe followingweek near Plaine des Chicors. There I found an old Claoxylon (appropriatelycalled "bois d'oiseaux" by the creoles),its branchescovered with gaping,bare scars,most of them old and dry, but an olivaceawas feedingat severalrunning fresh ones. A short time later a borbonicacame and fed at the same sites. A month later near La Possession I watched several borbonicaand some butterflies feeding at the exposededges of 50 FRANK B. GILL [Auk, Vol. 88

somebroken bark o.na large Acacia that had apparentlybeen damagedin a storm. The generalizedfeeding behavior o.f borbonicaresembles that of most other white-eyes,which take a variety of food items includinginsects, fruit, and nectar. White-eye insect-foragingis usually done in flocksand includeswarblerlike foliage gleaning,searching of terminal leaf branches, and in some speciestitmouselike hanging upside down (Z. virens, Z maderaspatana)or occasionalfly-catching (Z. pallida). The regulartrunk- creepingbehavior o.f boronicadoes not have commonparallel elsewherein the family, though both Z. chlorisinter•nedia and Z. lateralis have been noted clingingto the trunks of trees and movingupwards in small jumps (Mees, 1957: 41; 1969: 23). Most Zosteropsalso eat ripe fruits and berries. But nectar is apparentlyof minor importancein the diet of most white-eyes(Moreau et al., 1969: 42). Apparentlymost speciesof white- eyestake nectaronly occasionally,and thosefew that take nectarregularly alwaysdo so as a supplementto the fruit and insectsthat form the major portion of their diet. Thus in feedinghabits borbontcais similar to or perhapsslightly more generalizedthan other white-eyes,but olivaceaem- phasizesflower-feeding far more than typical white-eyesand also replaces the usualmodes of insectforaging by fly-catching.The aggressivebehavior and defenseof floweringbushes by olivaceais also unusualfor a white-eye as is the avoidance of fruit. Interspecific]eeding relationships.•Although the activitiesof borbonica and olivaceaare usuallyconducted separately, positive associations of the two speciesdo occur. OccasionallyZ. o. chloro.nothosis seenwith feeding flocksof Z. b. mauritiana (Niven, 1965; pers.o.bserv.), but my impression was that the associationwas alwaysa looseone that was interruptedby chloronothos'tendency to fly betweenwidely separated flowers rather than to foragein the continuousfashion of borbonica.Similar associations of the two speciesin the same feedingflock on R•union were extremelyrare, for only oncein 9 monthsdid I seeolivacea (three individuals)with a feeding flock of borbonica.Associations of o.livaceaand borbonica,however, are frequentin the vicinity of certain floweringplants, especiallyin grovesof floweringSophora, extensive groves of Fuchsia,and in ravineswith flowering Eugenia jambos. Elsewhere,in the absenceof such concentrationsof flowers,an associationbetween the two'species is likely to be brief, occur- ring only when passing borbonicapause to feed at relatively isolated Hypericum bushes. Interspecificaggression is commonbetween the two white-eyespecies in the vicinity of these flowers,and it is invariably one-sided: olivaceais aggressive,borbonica is submissive.The interspecificencounters are dis- tinctly lessintense than intraspecificones. Whereasdefense is not relaxed Jan. 1971] Mascarene White-eyes 51

TABLE 2 COIV[PARISONOF MASCARENIg WItlTIg-EYES TO TYPICAL •/VItlTE-IgYES

Typical white-eyes (virens, lateralis, Zosterops b. Zosterops o. Character patpebrosa,etc.) borbonica olivacea

Coloratio.n Carotenoids; eye-ring No carotenoids; Reduced carote- no eye-ring noids; eye-ring Proportions Normal Normal Long bill; short tail (Z. o. chloronothos) Vocalizations Seasonality Breeding season only Breeding season All year?? only Warbling song Present Present (muted) Present Challenge song Present: all day (lat- Dawn and evening Absent eralis); dawn and (absent in Z. b. evening ( virens) mauritiana) Flocking Winter o•nly (lateralis) All year Rarely all year in other ssp.? Territoriality Breeding-nest area None Flowering bushes (lateralis); absent or sporadic (virens) Food Insects, fruit, nectar Normal, though Flower-feeding; in- with increased sectsby fly-catch- creepingon tree ing trunks and branches until nonresidentoliva'cea are drivenaway, intrudingborbonica are chased briefly and then toleratedif they do not leave. Admittedlyborbonica often gainssome advantage by appearingin smallgroups, forcing the de- fendingolivacea to displaceseveral individuals in succession.Nevertheless persistenceby the chasedbird seemedto aggravatean intraspecificchal- lengebut resolvean interspecificone. In addition,wing quiveringby borbo.nicaseemed to easethe interspecificrelationship, while at other times the borbonicasimply left a floweringHypericum when an olivaceaappeared nearby,suggesting a well-defineddominance relationship. The net result was that the borbonicausually succeeded in feedingat someof the flowers defendedby an olivacea. Aggressionat floweringHypericum and Sophorawas especiallyconspic- uousand includedactual contactof fightingbirds at times. While resident olivaceadominated other white-eyes of both species,they wereoccasionally displacedfrom their ownbush by the introducedweaver, Foudia madagas- cariensis. DISCUSSION On both Mauritius and R•union islandslive two white-eyes,one of which is a typicallygeneralized, highly social species and the otheran unusually 52 FV,•K B. Gxz• [Auk, Vol. 88

TABLE 3 SLr•VI2VIAR¾OF MASCARENE WIZITE-EYE SPECIES CZIARACTERISTICS

Zosterops olivacea Zosterops borbonica

Character olivacea chloronothos borbonica mauritiana

Crown color Black Gray Gray/brown Gray Back color Green Gray Gray/brown Gray Belly color Dark gray Light gray; White to gray White with green no green and brown in center Eye-ring Present Present Absent Absent Foot color Dark gray Light brown Dark gray Dark gray Bill color Black wth Light brown Black with Black with gray base gray base gray base Tail/wing (mean) 0.71 0.64 0.74 0.74 Bill/wing (mean) 0.27 0.32 0.24 0.25 Feeding call "Tsip or tu tu" "Pit-pit" "Eeee-eeeee" "P-tree" Mimic warbling Loud ?? Muted ? ? Challenge song None ? ? Dawn and None evening Flower feeding Frequent; Frequent Occasional Occasional habitual Insects Fly-catching q- (method?) Warbler q- Warbler (q- creeping some fly- catching) Fruit Never ? ? Frequently ? ? in winter Flocking Very rarely ? ? Habitually Habitually aggressive,specialized flower-feeding species. The generalized species, borbonica,is closerto other white-eyesin virtually all aspectsof its biology than is the specializedspecies, olivacea (Table 2). Z. borbonica'sdepar- tures from the "norm" includerelatively simplechanges in plumagecolor, an increasedemphasis on trunk creepingas a modeof insectforaging, and somemodification of the vocal repertoire. It is primarily a speciesof edge and disturbedhabitats. Z. olivacea,on the otherhand, is primarily a species of the indigenousforests and differs from normalwhite-eyes in bill and tail proportions,call notes,absence of socialand flocking behavior,increased aggressivenessin relation to feedingsites, and a shift to flowersas its primary foodsource. The many differencesbetween the two insularpopulations of eachspecies are summarizedin Table 3; the differencesbetween Z. o. olivacea and Z. o. chloronothosare muchgreater than thosebetween Z. b. borbonicaand Z. b. mauritiana. The two racesof olivaceadiffer in vocalizations,plumage Jan. 1971] Mascarene White-eyes 53

coloration,overall size,and proportionsof the bill and tail, whereasthe two racesof borbonicadiffer primarily in their vocalizations.Thus, interinsular divergencehas been greater in olivaceathan in borbonica. Sequenceof colonizationand evolutionarychange.--In view of the uni- formity of continentalZosterops, one assumesthat the ancestor(s)of both borbonicaand olivaceawere typicalgeneralized white-eyes. Several hypo- thetical sequencesof colonizationof Mauritius and R•union by white-eyes and differentiationof the endemicforms exist. As interchangebetween the two islandshas been frequent (Moreau, 1957: 402), a white-eye that colonizedone of the islandsprobably reached the othera shorttime later, at least in relationto the timing of a secondsuccessful colonization of the islandsfrom the mainland.• Hence a single invasiono.f the Mascarene islands or separate invasionsof Mauritius and R•union followed by speciationand exchangeof speciesbetween islands seem unlikely hypothe- ses. ConsequentlyI favor a hypothesisof doubleinvasion of the islands, the first arrival giving rise to one of the modernspecies and the second arrival givingrise to the other. The question then follows whether the first successfulcolonist was the ancestorof the specializedolivacea or the generalizedborbonica. Z. olivaceahas certainlydiverged farther from a typical white-eyein habits and morphology,excluding plumage coloration, than has borbonicaand is a bird of the indigenousforest rather than disturbedhabitats. Its populations on Mauritius and R•union are also more different from each other than the comparablepair of borbonicapopulations, a fact suggesting, thoughnot necessarilyindicating, a longerseparation. These facts alone suggestthat o.livaceawas established on Mauritius and R•union before the arrival of borbonica's ancestor. Furthermore if borbonica colonized the Mascarenesfirst and retained its preferencefor edgehabitats and generalized feedinghabits, then the ancestorof olivacea,which was also a generalizedwhite-eye, would have had to shift its ecologicalpreferences, adapt to the indigenousforest, which was for somereason not inhabitedby borbonica,and perhapsdominate the establishedborbonica in order to colonizethe islandin the faceof congenericcompetition. Rather than occurringimmediately after and makingpossible the second colonization,it seemslikely that the major ecologicalshift gradually followed the first successfulcolonization, for as Lack (1969: 48) has noted, "any new arrival must competeimmediately with the establishedspecies and either its existingadaptations enable it to do so successfullystraight- away or it is eliminatedwithout time to evolve adjustments." I submit that the ancestorof olivacea increasedits flower-feedingfrequency and

3 For the purposes of these arguments the actual geographical origin of the colonists is not important. 54 FP•r•c B. GtLL [Auk, Vol. 88 dependencyas it adapted to the Mascareneforests, and that some time later the ancestorof borbanicawas able to colonizethese same islandsby virtue of beinga generalizedwhite-eye favoring edge and disturbedhabitats as was the ancestorof olivacea. Subsequentcompetitive interaction betweenthe two speciesmay then have causedolivacea to becomean even more specializedflower-feeder, to shift from normal modes of insect foraging to fly-catching,and to give up fruit-eating. The changefrom socialto aggressive,nonsocial behavior in olivaceapresumably accompanied its increaseddependence on flowers. With time alivacea'saggressive defensesof flowersmay haveaffected borbonica which, by losingthe white eye-ringand carotenoidpigmentation, became less subject to the interspecific aggressionevoked by thesetypical white-eyeplumage characteristics. I suggest,therefore, that the first white-eyeto arrive in the Mascarene islandsbecame more specialized, rather than moregeneralized, in a reduced competitiveenvironment. "Specialized" and "generalized"are admittedly only relative and somewhatambiguous terms. We do not know to what extent ancestralolivacea fed on flowersas opposedto fruit or insectsand to what extent it sacrificed fruit and insects for nectar before the arrival of borbonica. The initial increasein flower-feedingby t•livaceawas pre- sumablya broadeningof its feedingrepertoire and an increasein general- ization. The subsequentshift to flower-feeding to a point that was essentiallyirreversible once borbanica colonized the island, or even to the point that made possibleborbonica's successful arrival and ecologicalex- pansion,was certainly a lossof the opportunisticgeneralization that char- acterizescontinental Zosterops. Selectionoperates solely as a functionof overall reproductivesuccess, and the degreeto which a speciesspecializes is determinedby the location of peak reproductivesuccess within an array of potential modesof exist- ence, the sum of which correspondsto the species'"fundamental niche" (Hutchinson, 19'65: 32). Interspecific competitionwill determine an op- timal "realizedniche," which, in the presenceof specialists,may emphasize generalizedbehavior rather than specialized.In the absenceof interspecific competition,the directionof such species'evolution may shift towardsa more specialized "realized niche," one of which is specialized nectar feeding. Competition between flower-feedingbirds is often intense. Aggressive chasingby feedingsunbirds (Nectariniidae) and honey eaters (Meliphag- idae) is a familiar sight in Africa and Australia respectively,and interspecific aggressionis commonwhen severalspecies of hummingbirdsfeed in the sametree (Cody, 1968). Competitionand aggressionbetween similar- sizedsunbirds and honeyeaters has alsobeen described from the Moluccan islands(Rip.ley, 1959). In New Zealandaggressive Bellbirds (Anthornis, Jan. 1971] MascareneWhite-eyes 55

Meliphagidae),often preventsilver-eyes (Z. lateralis) from taking nectar at feedingstations and it seemsthat honeyeaters may also limit nectar feedingfrom flowersat many localities(Falla et al., 1966). Thus in the absenceo.f specializedflower feeders white-eyes may experiencea "competitive release"with respectto flower-feeding. The processof evolutionarychange hypothesized above for the Mas- carenewhite-eyes bears a strongresemblance to the "taxon cycles" described by Wilson(1961) andGreenslade (1968) for Melanesianant andSolomon Islandbird faunasrespectively. In this "cycle,"islands are colonizedby speciesthat occupymarginal, disturbed, and coastalhabitats in source areas. On the islandsthese species move into. the indigenousforest that lacksthe competitiveforest species of the mainland.Replacement by new colonistsoccurs in the lowlandsand with time rarity, extinctions,restric- tion to mountainareas of the interior,and relictdistributions result. Many white-eyesare foundprimarily in openand disturbed vegetation. Like certaincoerebids (Moynihan, 1969) they seemto. be "marginal" species,which are opportunisticgeneralists thriving in disturbedenviron- mentsand characteristicof islandsituations, both montaneand oceanic. The frequentecological separation of two or morewhite-eyes in the same geographicalregion (Table 4) suggeststhat this processof specialization and progressiveadaptation to. indigenous forest in the absenceof competi- torsmay be very importantin white-eyeevolution. Sympatryin white-eyes.--Thisstudy was initiated because of the un- usualoccurrence of sympatryin thegenus Zosterops despite its broadpaleo- tropicaldistribution. At first glancethere are, excludingthe Mascarenes, 22 possiblecases of sympatry,but only 7 of theseare bonafide, as alti- tudinal,geographical, or majorecological separations are involvedin the others(Table 4). Sixof the sevennon-Mascarene cases of actualsympatry are on smallislands, but unfortunatelyfew detailson the habitsof these white-eyesare available. On Ponapeand Palau in Micronesia,cinerea, the more abundantspecies is knownto dominateand chaseconspicilla'ta (Baker, 1951). A greaterincidence of congenericsympatry on islandsthan on continents, suchas that in white-eyes,is contraryto the trend found in other groups (Grant, 1968). It seemsthat white-eyespersist in continentalfaunas by virtue of beinggeneralized and marginalspecies. While the coexistenceof two suchgeneralists is not possible,specialization is alsoprevented by the presenceof alreadygood specialists in othergroups. But, by virtueof their socialbehavior, preferences for edgehabitats, and perhaps other attributes, white-eyeshave provedexceptionally good at colonizingislands, where increasedspecialization by the first arrival, which is madepossible by 56 FRANK B. GILL [Auk, Vol. 88

TABLE 4 SY2V•PATRY IN T:HE GENUS ZOSTEROPS

Locality Species Ecology Reference

AFRICA Zululand a) virens a) evergreenforest Moreau, 1957 b) senegalensis b) open vegetation Kenya and a) abyssinica a) dry lowlands Moreau, 1957 Somalia b) senegalensis b) mountains Grand Co- a) mouroniensis a) high alt. heath Benson, 1960 moro I. b) senegalensis b) lowland forest R•union x and a) borbonica a) insectsand fruit Present study Mauritius b) olivacea b) flowers

AUSTRALASIA Ceylon a) ceylonensis a) above 3,000 ft Henry, 1955 b) palpebrosa b) below 4,0.00ft Malaya and a) everetti a) middle and high Delacour, 1947 Thailand elev. b) palpebrosa b) coastal Luzon a) montana a) high (3,000- Delacour and 8,000 ft) Mayr, 1946 b) nigrorum b) low and mod- erate alt., SE Luzon c) japonica c) lowlands Mindanao a) montana a) over 6,000 ft Mees. 1957 b) everetti b) 0-5,000 ft Sumatra a) montana a) mountain tops Mees, 1957 b) atricapilla b) above 1,000 m c) palpebrosa c) 200-1,600 m Borneo a) atricapilla a) mountainsabove Smythies,1960 3,00,0ft b) eveterri b) N and NW Bor- neo-sub- montane c) palpebrosa c) lowlands (W Born.), mangroves and coast Java a) montana a) mountains Delacour, 1947; b) palpebrosa b) middle elevation Mees, 1969:314 c) ]lava c) coast Celebes a) chloris a) southwesternarm; Mees, 1961 open cultivated country, 0-1,200 m b) anomala b) southern arm; de- forested hills c) consobrinorum 3) southeasternarm d) atri]rons d) northern arm New Guinea x a) atri]rons a) 400-1,200 m, J. Diamond all New Guinea (in litt.) b) ]uscicapilla b) 1,200-2,200 m, west New Guinea c) novaeguineae c) 1,200-2,400 m, east New Guinea Australia a) lutea a) mangroves Serventy and b) lateralis b) scrub Whittell, 1951 Jan. 1971] Mascarene White-eyes 57

TABLE 4 (Co•rTI•rUEI))

Locality Species Ecology Reference

Yunnant a) palpebrosa ?? Mayr, 1967 siamensis b) japonica simplex

PACIFIC ISLANDS Fiji and Tonga• a) lateralis a) open country and J. Diamond light forest 0- (in litt.) 850 m b) explorator b) light forest and hill forest 200- 1,350 m New Caledonia a) xanthochroa a) forest Mees, 1969:138 b) lateralis b) bush and semi- open woodland Lifu • (Loy- a) minuta a) all habitats Mees, 1969: 141, alty Is.) b) inornata b) forest and scrub 123, 97 c) lateralis c) gardensand forest New Hebrides• a) f•vifrons a) forest, scrub Mees, 1969: 135, b) lateralis b) scrub and openings 98; J. Diamond (in litt.) Kulanbangra a) murphyi a) hills Mayr, 1945 b) rendovae b) lowla.nds Bougainville a) ugiensis a) mountains Mayr, 1945 b) metcalfii b) lowlands Ponape• and a) conspicillata seetext (p. 55) Baker, 1951 Palau (Caro- b) cinerea line Is.) Norfolk I. • a) tenuirostris a) forest and sec- Mees. 1969 ondary growth b) albogularis b) forest c) lateralis c) secondarygrowth

Actual sympatry involved. reducedcompetition, facilitates a secondsuccessflfi invasion and ultimately sustainedsympatric coexistence of two species.

ACKNOWLEDGMENTS

The 1964 field work was supported by the National Science Foundation as a part of the U.S. Program in Biology, International Indian Ocean Expedition, and was conducted under the auspicesof the Smithsonian Institution in Washington, D.C. The 1967 studieswere financed by a National ScienceFoundation Predoctoral Fellow- ship and a National Science Foundation Grant (GB 3366) to T. H. Hubbell at The University of Michigan for researchin Systematic and Evolutionary Biology. The Frank M. Chapman Memorial Fund of the American Museum of Natural History provided a tape recorder and accessoriesfor my use. In 1964 I was fortunate to have the generousassistance of the late Jean Vinson, Director of the Mauritius Institute Museum, who obtained the official permits needed to collect certain specimens,made available the facilities of his institution, and extended many courtesieso.n my behalf. In both 1964 and 1967 Armand Barau, Conseiler 58 FRANKB. Gn;L [Auk, Vol. 88

Generalde la R•union, Harry Gruchet, Conservateurdu Museum d'Histoire Naturelie de St. Denis (R•unio.n), and Christian Jouanin of the Museum National d'Histoire Naturelie in Paris rendered invaluable assistanceand encouragement. The frequent co- operation of the Service des Eaux et For•ts de la Rdunion and in particular Messrs. Moulin, Miguet, and Soroquereis also gratefully acknowledged.M. ThereseanCadet provided most of the plant identifications mentioned in this paper. George E. Watson directed portions of the 1964 field work and Robert W. Storer, in addition to contributing to the 1964 observations,provided regular counselduring this study. I am also indebted to Harrison B. Tordoff, James Bond, George Cox, and Jared Diamond, who criticizedpreliminary drafts of this paper.

SUMMARY

As one of the few casesof sympatryin the genusZosterops, the Masca- rene white-eyesinclude a typical generalizedspecies, borbonica, and a specializedflower-feeding species, olivecee. Z. borboniceis a speciesof edgeand disturbedhabitats and differs from typical white-eyesin plumage coloration,in frequenttrunk creepingas a modeof insectforaging, and in some modification of its vocal repertoire. Its two insular populations, Z. b. borboniceon R•union and Z. b. meuritiane on Mauritins, differ primarily in their vocalizations.Z. oliveceeis a speciesof the indigenous forestsand differs from normal white-eyesin bill and tail proportions, vocalizations,absence of socialand flockingbehavior, increased aggressivenessin relation to feedingsites, and a shift to flowersas its primary food source. Its two insularpopulations, Z. o. olivaceaon R•union and Z. o. chloronothoson Mauritius, differ in vocalizations,plumage coloration, overall size, and proportionsof the bill and tail. It is argued on the basisof comparisonspresented in this paper that the ancestor of oliveyea colonized the Mascarene islands before the ancestor of borbonica.In the absenceof flower-feedingcompetitors olivacea then adapted to the indigenousforest and increasedits dependenceon nectar feeding. This increasedspecialization enabled the ancestorof borboniceto colonizethe Mascarene islands sometimelater. The rarity of sympatric Zosteropson continentsis explainedin terms of competitionbetween two generalists. On islandswhere "competitiverelease" encourages specialization, doubleinvasions lead to sympatriccoexistence of white-eyes.

LITERATURE CITED

BAKER,R.H. 1951. The avifauna of Micronesia, its origin, evolution and distribution. Univ. Kansas Publ. Mus. Nat. Hist., 3: 1-$59. BENSON,C. V•r. 1960. The birds of the Comoro Islands: Results of the British Ornithologists'Union Centenary Expedition 1958. Ibis, 105b: 5-106. BERLXOZ,]. 1946. Faune de l'empire francais. IV. Oiseaux de la R•union. Paris, Librairie LaRose. BROEKltUYSEN,G. J., ANDJ. M. WINTERBOTTOM.1968. Breeding activity of the Cape Jan. 1971] Mascarene White-eyes 59

White-eye Zosterops virens capensisSunderall in the South-west Cape. Ostrich, 39: 163-176. Co•)Y,M.L. 1968. Interspecificterritoriality among hummingbirdspecies. Condor, 70: 270-271. DELACOXJR,J. 1947. Birds of Malaysia. New York, Macmillan Co. DELACOIJR,J., AND E. MAYR. 1946. Birds of the Philippines. New York, Mac- millan Co. FALZA,R. A., R. B. S•BSON,.aND E.G. TUmaOTT. 1966. A field guide to the birds of New Zealand. London, Collins. FLg•NG, C.A. 1943. Notes on the life history of the Silver-eye based on colour- banding. Emu, 42: 193-217. GP•NT, P. R. 1968. Bill size, body size, and the ecological adaptations of bird species to competitive situations on islands. Syst. Zool., 17: 319-333. GR•NSZADE,P. J. 3/1. 1968. Island patterns in the Solomon Islands bird fauna. Evolution, 22: 751-761. GInruNWAY,J.C. 1958. Extinct and vanishingbirds of the world. SpecialPublication No. 13. New York, Amer. Comm. Intern. Wild Life Protection. HAC•SXJKA,3/1. 1953. The Dodo and kindred birds. London,H. F. and G. Witherby Ltd. HA•a•SON,C. J. O. 1968. Allopreeningas agonisticbehavior. Behaviour, 24: 161- 209. H•;NaY, G.M. 1955. A guide to the birds of Ceylon. Londo•n,Oxford Univ. Press. HXJTC•NSON,G. E. 1965. The ecologicaltheater and the evolutionary play. New Haven, Connecticut,Yale Univ. Press. K•AST, A. 1968. Competitive interactions and the evolution of ecologicalniches as illustrated by the Australian honeyeatergenus Melithreptus(Meliphagidae). Evo- lution, 22: 762-784. LACK, D. 1969. Tit niches in two worlds; or homage to Evelyn Hutchinson. Amer. Naturalist, 103: 43-49. MAc AaT•Ua, R. H., ANDE. O. W•-SON. 1967. The theory of island biogeography. Princeton, Princeton Univ. Press. MA¾1•,E. 1945. Birds of the southwestPacific. New York, Macmillan Co. MAYa, E. 1967. Family Zosteropidae,I.ndo-Australian Taxa. Pp. 289-326 in Check- list of birds of the world, vol. 12 (R. A. Paynter, Jr., Ed.). Cambridge, Massa- chusetts,Mus. Comp. Zool. M•s, G.F. 1957. A systematicreview of the Indo-Australian Zosteropidae,part 1. Zool. Verhandelingen, 35: 1-204. M•s, G. F. 1961. A systematicreview of the Indo-Australian Zosteropidae,part 2. Zool. Verhandelingen,50: 1-168. M•s, G.F. 1969. A systematicreview of the Indo-Australian Zosteropidae,part 3. Zool. Verhandelingen,102: 1-390. M•LON, P. 1951. Notes sur l'avifaune actuellesde File de la Rdunio.n. 'rerre et la Vie, 98: 129-178. Moa•AXZ,R.E. 1957. Variation in the westernZosteropidae (Aves). Bull. Brit. Mus. (Nat. Hist.), Zool., 4: 312-433. MOa•A•J,R.E. 1964. White-eye. Pp. 885-886 in A new dictionary of birds (Sir A. L. Thomson, Ed.). New York, McGraw-Hill Book Co. Moa•AXZ,R. E. 1967. Family Zosteropidae,African and Indian Ocean taxa. Pp. 327-337 in Check-list of birds of the world, vol. 12 (R. A. Paynter, Jr., Ed.). Cambridge,Massachusetts, Mus. Comp. Zool. 60 FRANK B. GILL [Auk, Vol. 88

MOREAV,R. E., M. PERRINS,AND J. T. HuGH•s. 1969. Tongues of the Zosteropidae (White-eyes). Ardea, 57.' 29-47. MOYNInAN, M. 1968. The "Coerebini": a group of marginal areas, habitats and habits. Amer. NaturaList, 102: 573-581. NEWTO,r, R. 1958. Ornithological notes on Mauritius and the Cargados Carajos archipelago. Proc. Royal Arts and Sciencesof Mauritius, 2: 39-71. NivEN, C. 1965. Birds of Mauritius and R•union. Ostrich, 36: 84-87. PETERSON,R.T. 1961. A field guideto westernbirds. Boston,Houghton Mifflin Co. RIPL•¾, S. D. 1959. Competition between sunbird and honeyeater speciesin the Moluccan Islands. Amer. Naturalist, 93: 127-132. RIVALS,P. 1952. Etudes sur la v6g6tation naturelie de File de la R•union. Tray. Lab. Forest. Toulouse, Tome V, Section 3 •, vol. 1, Art. 2. ROUNDTREE,F. R. G., R. GUERL•,S. PELTE,AND J. Vr•soN. 1952. Catalogueof the birds of Mauritius. Mauritius I.nst. BulL, 3: 155-217. SERVENT¾,D. L., A•m H. M. WmTTELE. 1951. Birds of western Australia, secondEd. Perth, Paterson Brokensha Pty. Ltd. SKEAD,C. J., AND G. A. RANGER. 1958. A contribution to the biology of the Cape Province White-eyes (Zosterops). Ibis, 100: 319-333. S•¾TmES, B.E. 1960. The birds of Borneo. London, Oliver and Boyd. STORER,R. W., Am F. B. GIL•. 1966. A revisionof the MascareneWhite-eye Zoster- ops borbonica (Aves). Occ. Pap. Mus. Zool., Univ. Michigan, No. 648. UPTON,B. G. J., ANDW. J. WADS.WORTH. 1966. Thebasalts of R6unionIsland, Indian Ocean. Bull. Volcanologique,29: 7-24. VAUG•LaN,R. E., AND P. O. WtEttE. 1937. Studies on the vegetation of Mauritius. J. Ecol., 25: 289-343. WATSON,G. E., R. L. ZuSX,AND R. E. STORER. 1963. Preliminary field guide to the birds of the Indian Ocean. Washington, Smithsonian Inst. WIESON, E. O. 1961. The nature of the taxon cycle in the Melanesian ant fauna. Amer. NaturaList,95: 169-193.

Museum of Zoology,The Universityof Michigan, Ann Arbor, Michigan 48104. Presentaddress: The Academyof Natural Sciencesof Philadel- phia, Philadelphia,Pennsylvania 19103. Accepted5 November 1969.