sign in sign up
<<
Home , British Solomon Islands, Gallirallus, Guam Rail, Hypotaenidia, Rail (bird), Solomon Archipelago

THE AUK

A QUARTERLY JOURNAL OF

VOL. 108 JULY1991 No. 3

A NEW OF FROM THE AND CONVERGENT OF INSULAR FLIGHTLESSNESS

JAREDDIAMOND Departmentof Physiology,University of CaliforniaMedical School LosAngeles, California 90024-1751, USA

ABSTRACT.--GalIiralIusrovianae is an extant new speciesof flightlessor weak-flying rail from the Solomon islands in the southwest Pacific Ocean. Within the Solomon islands it is known from and reported from four neighboring islands,all joined at Pleistocene timesof low sealevel. The new speciesbelongs to the G. philippensisgroup (sensu Olson 1973a) and is mostsimilar to the widespreadG. philippensisand to G. owstoni() and next most similar to G. wakensis(Wake). It exemplifiesthe phenomenonof convergentevolution in two respects.First, a volant ancestorsimilar to G. philippensishas independentlygiven rise to flightlessor weak-flying derivatives on numerousoceanic islands, including G. rovianae,G. owstoni,G. wakensis,G. australis(),and others.Emphasizing the easewith which rails evolve flightlessnesson islands,I note 10 other groupsof rails in which insular flight- lesshesshas evolved repeatedly. Second,I suggestthat the ancestralspecies had boldly patternedplumage similar to that of G. philippensis,and that insular reductionof bold pat- terning has proceededindependently in G. rovianae,G. owstoni,G. wakensis,and severalother G. philippensisderivatives. Received 13 February1990, accepted 14 February1991.

THE SOLOMON islands, which constitute an weak-flyingground , named "Kitikete" in archipelago in the tropical southwest Pacific the widespread Roviana language of that dis- Ocean,harbor many localizedendemic subspe- trict. In 1977Alisasa Bisili, a retired government ciesand speciesof (Mayr 1942, 1969).Or- officer, succeededin collecting a specimen. It nithologicalexploration of the Solomonislands has proven to be a distinctive new speciesbe- beganin 1838and climaxedin 1927-1930,when longing to a widespreadgroup of Pacificspecies the Whitney South SeasExpedition made large that includes the volant Gallirallusphilippensis collectionson every ornithologically signifi- (Buff-bandedRail), plus the flightlessG. owstoni cant island and discovered dozens of new taxa (), G. australis(), G. wakensis (Mayr 1945). Between 1936 and 1980 nine ad- (Wake Rail), G. modestus(Chatham Island Rail), ditional endemic forms were described (two andG. sylvestris (Lord Howe Rail) (Olson1973a). speciesand seven subspecies).Most of them Some characters of the new rail, however, are were from the mountains of and shared with the two species of Nesoclopeus Bougainville, the two largest and highest So- (Woodford'sRail, N. woodfordi,and Barred-wing lomon islands (Danis 1938; Cain and Galbraith Rail,N. poeciloptera;both flightless), a genusthat 1955; Hadden 1981, 1983; Ripley and Hadden had not been considered close to until 1982). Olson (1973a) noted some similarities. The dis- Between 1972 and 1976 I surveyedbirds on coveryconnects Nesoclopeus to the G.philippensis most of the Solomon islands that are of orni- group and strengthensthe relationship sug- thological interest. While I was on New Georgia gestedby Olson. The new speciesand its rela- and four neighboring islands,local residents tives are of interest asan example of convergent repeatedlydescribed a chickenlike,flightless or evolution,whereby rails independentlyevolve 461 The Auk 108: 461-470. July 1991 FRONTISPIECE.Gallirallus philippensis (upper left), G. wakensis(upper right), G. owstoni(middle right), Neso- clopeuswoodfordi (lower right), and G. rovfanae(lower left). Paintingby JamesCoe, from specimensin the AmericanMuseum of NaturalHistory. Publication of the colorplate was supported by the DonaldL. Bleitz fund. 462 JAREDDIAMOND [Auk,Vol. 108

ainville SOLOMON ISLANDS

Choiseul Kulambangra Wono

Ysabel Vella

Ganonga

Giza' , • . o ' •a:-, ,•..- I•lalada KohinggoRendova ' o ø Tetipari Gatukai ',...'• _.. • •,5 • Guadalcanal

155øE 160øE CristobalSan • • I I Fig. 1. Map of the Solomonislands. Dashed lines show the outlinesof land during late-Pleistocenetimes of low sea level, as deducedfrom presenthydrographic depth contours.Islands from which Nesoclopeus woodfordi,Gallirallus philippensis, and G. rovianaehave been reportedare markedby vertical lines, horizontal lines,and black,respectively. flightlessnessand, in somecases, similar plum- the language spoken on New Georgia and age characterson widely scattered oceanic is- neighboring islands,for the people occupying lands. thoseislands, and for the lagoonthat the islands surround. Gallirallus raviahoe, sp. nov. Measurements.--See Table 1. Diagnosis.--A medium-sized rail closest in to Gallirallusphilippensis and G. owstoni, Holotype.--AMNH no. 826433,collected near and similar in body size and in bill size and Munda, New Georgia,the Solomonislands, in shape to adults of the former, but with a much June 1977 by Alisasa Bisili. Sex unknown. The longer tarsus(Table 1). Unique among Galliral- specimen was initially prepared as a mummy lus taxa in its unmarked, nearly uniform chest- and then made into a study skin in New York. nut-brown upperparts. Shares boldly barred Distribution.--Known only from the type underpartswith G. philippensisand G. owstoni, specimen.Reported by villagers in the western and shares ochraceous or rufous breast band Solomon islands to be widespread on New with thosetwo taxaplus G. australisand G. wak- Georgia and to occur on the neighboring is- ensis,but these characters are not shared with landsof Kulambangra,Kohinggo, Wana Wana, other G. philippensisrelatives. Gallirallus rovianae and Rendova. To be sought on the islands of is lessboldly patterned than G. philippensisand and Tetipari as well (Fig. 1). G. owstoniexcept in the barred underparts. Spe- Etymology.--From"Roviana," the name for cifically, there is greatly reducedwing barring, July1991] Gallirallusfrom the Solomons 463 less distinct facial pattern, reduced barring of locality:Olson 1986)and the unique type of the the undertail coverts, and (vs. G. philippensis) remaining race of N. woodfordi,N. w. woodfordi disruptedbreast band and unmarked back.Dif- (Guadalcanal island in the Solomon islands: fers further from G. philippensisand G. owstoni Ogilvie-Grant 1889). in that the ventral barring is not black but char- coal brown (Table 2). DISCUSSION Description.--Theupperparts are a nearly uni- form unmarked dark chestnut brown from the COMPARISONS WITH IMMATURES crown through the back to the upper tail. The color is slightly richer on the nape, continuing Becausethe age of the unique type is un- into a richer, slightly reddish chestnut mask known, it is necessaryto considerwhether its through the eye. Starting at the posterior mar- distinctive charactersmight be marks of im- gin of each eye is an inconspicuousgray stripe maturity rather than of speciesstatus. that continues posteriorly almost to the nape, Alisasa Bisili, the New Georgian who col- but there is no clear indication that this stripe lected the type, said that he was familiar with continues anteriorly towards the lores. The the differencesbetween young and adult G. ro- wings agree in color with the back except for vianaeand that the type was full grown. During eight small white spots in the primary coverts my 1974 and 1976 fieldwork on New Georgia, of each wing. The chin and throat are whitish, Bisili volunteered accurate information about becominggray on the cheeksthrough the breast. virtually all of the 64 other resident bird spe- A pinkish-tan wash on the tips of the neck and cies. Hence if the type is not fully adult, it is breastfeathers forms a breastband. The upper neverthelessunlikely to be much smaller than abdomen, flanks, and sides of the neck and breast adults or much different from adults in plum- are barred charcoal brown and white, with the age. tips broadly washed with chestnut. The center The primaries of the G. rovianaetype are all of the lower abdomen is unmarked light buff; in sheath. This would be compatible with its the shanks, unmarked brown. The undertail being fully adult if the adult molt of G. rovianae coverts are broadly barred pale buffy (washed were synchronous,and would suggestan im- with pinkish-tan or brick towards the tips) and mature with primaries not fully grown if the black. Most of the wing is still in sheath, but adult molt were asynchronous.However, it is the visible feathering of the underwing is un- not possibleto guessthe adult molt pattern of markeddark brown exceptfor two obscurewhite G. rovianae, because Stresemann and Strese- spots. mann (1966) examineda single molting adult Softparts.--The collector did not record col- specimeneach of G. philippensisand N. woodfordi ors in life. In the skin the legs and bill are dull and noted that the former may molt its pri- in color. maries synchronously, the latter asynchro- Material exarnined.--Adultspecimens of the nously. following taxa in the American Museum of Nat- The tarsus/bill ratio of 1.65 is much higher ural History: 23 of the 27 currently recognized than the ratio in adultsof all likely relatedspe- races of Gallirallusphilippensis, plus G. owstoni cies (1.20-1.33; Table 1). But the young of ni- (Guam), G. wakensis(Wake), G. australis (New difugousbirds quickly acquireadult-sized legs Zealand), G. modestus(Chatham), G. sylvestris for running, and the bills do not reach adult (Lord Howe), G. torquatus(Philippines and Ce- size until later. This is true in particular of Gal- lebes), G. insignis(New Britain), G. striatus liraflusphilippensis and Nesoclopeuswoodfordi. Six (southeastAsia and Indonesia), Nesoclopeuspoe- immature G. philippensishad a tarsuslength (40- ciloptera(Fiji), N. woodfordiimrnaculatus (Ysabel 45 mm) comparableto the adult value (37-47.5 island in the Solomon islands), and N. w. tertius mm), but a bill considerablyshorter (25-29 mm ( in the Solomon islands). vs. 27-38 mm in adults). Similarly, the single Immature specimensof nine racesof G. philip- available immature N. woodfordihas an adult- pensis,and of N. w. tertius.I also compared G. sized(56 mm) tarsus(comparable to 53-65.5 mm rovianaewith published descriptionsof G. oki- for adults) but a short (36 mm) bill (41.5-44.5 nawae (Okinawa: Yamashina and Mano 1981), mm for adults). As a result, immatures of both the unique type of G. sharpei(unknown type specieshave high tarsus/bill ratios(1.56 for both 464 JAREDDIAMOND [Auk, Vol. 108

TABLE1. Measurementsa (mm) of Gallirallusrovianae and its likely Gallirallusand Nesoclopeusrelatives.

No. of specimens Exposedbill length Tarsuslength Species • • • • • G. wakensis 8 27.4 (25.0-30.0) 33.4 (32.5-35.0) G. philippensis 59 55 34.0 (29.5-38.0) 31.0 (27.0-36.0) 43.0 (37.0-47.5) 40.9 (37.0-46.0) G. rovianae 1 34.5 57.0 G. owstoni 2 2 42.0 (41.5, 42.5) 37.8 (37.5, 38.0) 50.3 (49.5, 51.0) 45.3 (45.0, 45.5) N. woodfordi 6 2 43.4 (41.5-44.5) 42.0 56.9 (53.0-65.5) 58.7 (54.0, 63.5) N. poeciloptera 4 47.5 (46.0-49.5) 61.9 (58.0-64.5) aMeasurements are of the type of G. rovianaeand of adult specimensof other taxa. Measurementswere made to the nearest0.5 mm and are givenas the mean,followed in parenthesesby the rangeor by individualvalues. Available specimens of G. wakensis,G. rovianae,and N. poeciloptera were not sexed. species),similar to that of the G. rovianaetype the belly. It differs additionally from the N. (1.65) but considerablyhigher than that of con- woodfordiimmature in its gray eye stripe con- specificadults (1.29 for G. philippensis,1.33 for trasting with the chestnutmask, the much more N. woodfordi).Both the tarsusand bill lengths markedventral barring, the gray breast,the dis- of the G. rovianaetype are similar to those of rupted tan breast band, and the palenessof the the N. woodfordiimmature. anterior belly comparedwith the posteriorbel- Iramaturesof the nine racesof G. philippensis ly and breast. examined differ from adults in the following In short, the relatively long tarsus,short bill, plumagecharacters: rufous of the crown, nape, and wings in sheath of the G. rovianaetype are and maskreplaced by dull darkbrown; eye stripe equivocal indicators of its age. Its plumage, more obscure, more washed with brown, less however, is quite different from that of irama- clear pale gray; upperpartsnot black with dis- tures as well as adults of G. philippensisand N. tinct white bars, but dark brown with more ob- woodfordi.Hence its plumage distinctnessin- scure, paler brown edges, bars, or spots;pos- dicates taxonomic distinctness. terior throat and anterior breast mottled or barred and with some ochraceous wash, rather ASSESSMENT OF RELATIONSHIPS than clear gray; black-and-white barring of the flanksand belly more obscureand washedwith Olson (1973a)concluded that Gallirallusphilip- brown; more extensive unbarred whitish area pensis,G. owstoni,and G. wakensisare closelyre- in the middle of the belly. The one available lated and that G. australis, G. modestus, and G. immatureN. woodfordidiffers in plumagefrom sylvestrisare more distant membersof the same adults only in the more extensivewhite of the speciesgroup. Ripley's (1977) interpretation dif- chin, and in the more brown, less gray under- fers. As already explained (see diagnosisand parts with light edgeson the breast. Table 2), G. rovianaeclearly is closestto the for- One of the two plumage charactersby which mer three species,with which it sharesthe breast the immatureN. woodfordidiffers from adults band and boldly barred underparts. However, (the pale chin) is shared with the G. rovianae G. rovianaeis much less boldly patterned than type, as are three of the charactersby which any race of G. philippensis,especially in the ab- immatureG. philippensis differ from adults(more sence of dorsal markings, the greatly reduced obscureeye stripe, reduced ventral barring,and wing bars and facial pattern, the reduced bar- somebrown or chestnutedges ventrally). How- ring of the undertail coverts,and the disrupted ever, the G. rovianaetype differsfrom iramatures breast band. Gallirallus owstoni, G. wakensis,and of both G. philippensisand N. woodfordiin its G. rovianae,all differ from G. philippensisin their unmarked, nearly uniform chestnut-brown unmarked backs and reduced breast band, but crown and upperparts. It differs additionally G. rovianaedeviates further from G. philippensis from G. philippensisiramatures in the greatly re- in its much more obscurefacial pattern, and G. duced barring and spotting of the wings, the rovianae and G. wakensis deviate further in their unmottled gray breast, the unspotted tail, and greatly reduced wing markings from G. philip- the lack of a large white area in the middle of pensisthan does G. owstoni.Gallirallus rovianae is July1991] Gallirallusfrom the Solomons 465

TABLE 1. Extended. clopeushas traditionally not been associated with the G. philippensisgroup. Olson (1973a), how- ever, noted three similarities: a tenuous nasal Middletoe and claw length Average tarsus / bar, variably barred underwing, and obscure • • bill ratio facial pattern.I noted severalother philippensis- 41.7 (40.0-44.0) 1.20 like featuresof someNesoclopeus taxa. These in- 46.2 (41.0-54.5) 43.3 (39.0-48.0) 1.29 clude the few white spotsin the upperwing and 53.0 1.65 the obscureventral barring of N. w. woodfordi 50.5 (50.0, 51.0) 47.0 (47.0, 47.0) 1.30 59.9 (57.0-63.5) 60.0 (56.0, 64.0) 1.33 and N. w. tertius,the slightly barred undertail 54.0 1.26 covertsof N. w. tertius,the paler throat of all Nesoclopeustaxa except N. w. woodfordi,and the slightly paler belly of N. w. tertiusand N. poe- ciloptera.Any one of these charactersalone unique in its uniformly dark chestnut-brown would carry little weight, but collectively they dorsalcoloration. Thus, reduction of patterning suggestthat Nesoclopeusmay be derived from has proceededthe leastin G. owstoni,further in the G. philippensisgroup. G. wakensis, still further in G. rovianae, and fur- Gallirallusrovianae resembles Nesoclopeus (es- thest in the fairly monotonouslyplumaged G. peciallyN. woodfordi)and differsfrom G. philip- modestusand G. sylvestris. pensisin the long tarsus,unmarked back, lack Of these seven species,G. philippensishas by of contrastbetween back and crown, unspotted far the widestgeographic distribution, from the upper tail, obscurefacial pattern, few obscure Philippines and Indonesia east to . The white spotsin the upperwing, reducedor ab- other speciesare eachconfined to singleremote sent barring of the underwing, and reduced Pacific islands or island groups and are vari- barringof the undertailcoverts. These similar- ously flightlessor weak of flight. One conceiv- itiesbetween G. rovianaeand N. woodfordicould able interpretationof the plumage differences be convergent,as I interpret the similarity of is that the ancestralcondition was a relatively G. rovianae to G. wakensis and G. owstoni to be. unpatternedplumage, that mosttaxa other than Alternatively, becauseG. rovianaeand N. wood- G. philippensisrepresent an earlier expansion fordi are both from the Solomonislands, their wave retaining the ancestralpattern, and that resemblancecould reflect relationship. the bold pattern of the now-expanding G. phil- ippensisis a newly evolved character.This in- REPORTS OF GALLIRALLUS ROVIANAE terpretation is unlikely, however, becausethe variousspecies of the G. philippensisgroup cover Most information about G. rovianae in life the whole spectrum from the most (G. philip- comesfrom reportsof Solomonislanders in 1974 pensis)to the least(G. modestus,G. sylvestris)pat- and 1976. On each island visited, I asked is- terned, with G. owstonisharing most of the bold landers (especiallythe older men) to describe featuresof G. philippensisand with G. wakensis all birds known to them. The questioningwas andG. rovianae sharing some (but fewer)of those donein sucha way asto avoid leadingquestions features.It is much more likely that the ancestor and to testreliability. I askedpeople to describe was boldly patterned. Presumablythe wide- birds they knew rather than to respondto par- spreadG. philippensis, which still shares much' ticular questions.When it was necessaryto raise of its range with numeroussympatric rail spe- the subjectof a particular species,I provided cies, retained the pattern becauseit servesthe just enoughdetails to identify the speciesand function of speciesrecognition. The other taxa then asked the informant to provide further are in the processof independently losing their details. I absolutely avoided yes/no questions patternsas a result of reachingislands with few (Diamond 1989).In mostcases the islandersgave or no other rail species.Under this interpreta- names(in their local language)and clearly rec- tion the resemblances between G. rovianae, G. ognizable detailed descriptions(including ac- wakensis,and G. owstoni--eachliving on an is- counts of breeding, diet, and seasonalmove- land 2,500-3,200 km from the other two--would ments) that encompassed almost all species be due to parallelism. recordedpreviously for that islandby European The two speciesof Nesoclopeusare alsoflight- collectors,plus somespecies not yet recorded. less endemics on remote Pacific islands. Neso- Islanders on New Georgia, Wana Wana, Ko- 466 JAREDDIAMOND [Auk, Vol. 108

'2,,5

•.•

8 .5•

ß •'• July1991] Gallirallusfrom the Solomons 467

hinggo, and Rendova islands independently describeda chickenlikeground bird, called"Ki- tikete" in the Roviana languageshared among these islands.A similar bird on Kulambangra islandwas called"Keremete," and bilingual is- landers said that it was the same as the bird called"Kitikete" in the Rovianalanguage. The Kitikete (alias Keremete) was said to differ from other chickenlike ground birds, which were readily identified as Amaurornisolivaceus (Pyu- Peyo),Porphyrio porphyrio (Balikuhu), and Mega- podiusfreycinet (Eo) (Roviana names in paren- theses).All accountsemphasized that the Ki- tikete ran very fast,zigzagged, could be caught only with dogs,and was flightlessor nearly so. One informant said that it could flutter but nev- er more than half-a-meterabove the ground. Its was described as forest and especially second growth, where young trees grew on abandonedgarden sites.The call was said to be a rapidly repeated high-pitched note, which gave rise to its Roviana name, Kitikete. Addi- tional details were provided by my most knowledgeableinformant, Teu Zingbite of Ku- lambangra.Zingbite describedthe diet as om- nivorous(e.g. worms, seeds, coconut shoots, po- tatoesand taro from gardens,and small crabs) and the nestas a depressionon the ground lined with debris,containing two or three ,and built in the dry season(June), when rainwater on the forest floor was not a problem for a ground bird. When informants on the other major islands near New Georgia (Gatukai, , Ganonga, , Gizo) volunteereddescriptions of birds known to them, they did not include the Kitikete or any bird suggestiveof it. I specifi- cally discussedthe Kitikete (as describedby people from New Georgia, etc.) with infor- mantson Vella Lavella and Gizo, and they de- nied that the Kitikete occurred on their own islands. No local information was obtained for Tetipari island, which lacks a village, or for Vangunuisland, where I did not visit a village. Thus,the reportedrange of the Kitikete consists of New Georgia,Wana Wana, Kohinggo,Ren- dova, and Kulambangra;and it is still to be sought on Tetipari and Vangunu. It was in re- sponseto my requestfor a specimenof a Kiti- kete that Bisili collectedthe type of G. rovianae and identified it as the Kitikete. There are three possiblereports of G. rovianae by western observers.First, when I was on Ko- hinggo island on 19 September1976, I heard a 468 JAREDDIAMOND [Auk, Vol. 108 soft,very rapidly repeatedcall like "kitiketeki- (Sibley 1951, Blaber 1990) and Kulambangra tiketekitikete..." etc.from densesecond growth (Finch1985). Bougainville, Choiseul, Ysabel, and 3-m tall at the edge of a garden.My Roviana- possiblyGuadalcanal were joined by Pleisto- speakingguide identified the caller as the Ki- ceneland bridges.Nesoclopeus woodfordi is still tikete. extant on Ysabel and Choiseul but apparently Second,Blaber (1990) observed on New Geor- becameextinct many decadesago on Bougain- gia a group of three rails that he identified as ville and Guadalcanal (Diamond 1987). G. philippensis,based on his experiencewith that Finally, in the easternSolomon islands the specieson Guadalcanaland . Fie noted volant G. philippensisoccurs on San Cristobal, them as similar in size to G. philippensisof Gua- Ugi, andGuadalcanal, which lacked recent land dalcanal and perhaps smaller than Australian connections to each other. The colonization of birds.Fie did not recordplumage details except Guadalcanalmay be recent,postdating the ex- for noting that the ochre breastband familiar tinction there of N. woodfordi. to him from Guadalcanal and Australian birds Thus,rails of the G.philippensis group or Neso- was narrow and "ragged" in the New Georgia clopeusprobably occur on many centralSolo- birds. Fiis descriptionmay correspondto what mon islands.The presenceof N. woodfordiand I describeas the disrupted pinkish-tan breast the discoveryof G. rovianaemay explain the band in the type of G. rovianae. otherwisepuzzling restriction of G. philippensis Finally, in a dictionary of the Roviana lan- in the Solomon islands to the eastern islands, guage (Waterhouse1949: 169) a list of Roviana despitethe superiorcolonizing ability that has names for plants and and of their En- permittedit to occupya rangeextending from glish identificationsincludes the entry "kitikete, Indonesia to Samoa (>8,000 kin). a dark, very nimble bird. [Fiypotaenidia sp.]." Gallirallusphilippensis was often placed in the REPEATED INDEPENDENT EVOLUTION OF Hypotaenidiawhen the dictionary was FLIGHTLESSNESS IN ISLAND RAILS written in 1928. It is unknown how Waterhouse was able to associatethe Kitikete correctlywith Gallirallusrovianae adds one more to the grow- G. philippensis,but he may have used Roviana ing list of flightlessor weak-flying derivatives informants familiar with the islands of Guadal- of G. philippensisstock endemic to oceanicis- canal and San Cristobal,where G. philippensisis lands.At leasteight are now known (Table 3), a common roadside bird. ranging from G. p. dieffenbachii(so similar to G. philippensisthat it is now usually considered DISTRIBUTIONS IN THE SOLOMON ISLANDS conspecific)through G. rovianaeand G. owstoni, whose specificdistinctness is clear but whose Distributions of rails of the G. philippensis relationshipto G. philippensisis equally clear, to groupand of Nesoclopeusin the Solomonislands strongly modified derivativessuch as G. aus- are summarized in Figure 1. tralis,G. sylvestris,and G. wakensis.Because none The five islandsof the New Georgia group of the islandsinvolved were connectedby land from which G. rovianaehas been reported were to each other, they must have been colonized joined at Pleistocenetimes of low sealevel until independently by a volant ancestor,and the approximately10,000 yr ago(Diamond and Mayr colonistsmust have evolved independentlyto- 1976).The nearby islandsof Vella Lavella and wardsflightlessness. Reduction in patterningof Gizo, whose inhabitants reported G. rovianaeas plumagehas also evolved independently in most not present, were not joined to New Georgia of these taxa. and its neighbors.Thus, the present distribu- At least 10 other groupsof rails show inde- tion of G. rovianaeis a legacyof Pleistoceneland pendent evolution of multiple flightlessor bridges. weak-flyingderivatives on oceanicislands (Ta- Nesoclopeuswoodfordi has been collectedon ble 3). Evidently,as discussed by Olson (1973b), Bougainville, Ysabel, and Guadalcanal (Mayr insular rails not only evolve flightlessnesseas- 1949) and observed on Choiseul (Fl. Fiamlin in ily, but they are also under strong pressureto the unpubl. journal of the Whitney South Seas do so.Olson noted that evolutionof flightless- Expedition; my unpubl. obs.). Sight observa- nessmay involve only few geneticchanges, for tions of large dark rails suggestiveof N. wood- instancein genescontrolling relative growth fordi have been obtained from New Georgia ratesof differentbody parts. The selective force July 1991] Gallirallusfrom the Solomons 469

TABLE3. Convergentevolution of flightlessnessin island rails.a Asterisksdenote extinct taxa.

Volant relativeb Flightlessor weak-flying taxa(range) c Fulica atra *F. c. chathamensis(Chatham), *F. c. prisca(New Zealand) Gallinulachloropus G. comeri(Gough), *G. nesiotis(Tristan de Cunha) Gallinula ventralis *G. hodgeni(New Zealand),G. mortieri(Tasmania) Gallirallusphilippensis *G. p. dieffenbachii(Chatham), G. australis(New Zealand),*G. modestus(Chatham), G. owstoni(Guam), *G. pacificus(Societies), G. rovianae(Solomons), G. sylvestris (Lord Howe), *G. wakensis(Wake) Gallirallustorquatus G. insignis(New Britain),G. okinawae(Okinawa) Porphyrioporphyrio *P. albus(Lord Howe), *P. kukwiedei(New Caledonia), P. mantelli(New Zealand), *P. paepae(Marquesas) Porzanapusilla *P. astrictocarpus(St. Helena), *P. palmeri(Laysan) tabuensis P. atra (Henderson), *P. monasa(Kusaie) Dryolimnaspectoralis? *Atlantisiaelpenor (Ascension), *A. podarces(St. Helena), A. rogersi(Inaccessible) Gallirallus *Nesoclopeuspoeciloptera (Fiji), N. woodfordi(Solomons) ? *Pareudiastespacificus (Samoa), P. silvestris(Solomons) a From Olson (1973a, b, 1975), Steadman(1986, 1988, 1989), Balouet and Olson (1989), and referencescited therein. bThe mostclosely related volant species. • Railsthat are endemicto differentoceanic islands and that appearto have evolvedto or towardsflightlessness independently. Not listedare other Porzanataxa from the Cooks, Societies, and Hawaii for which it is uncertain whether P. pusiliaor P. tabuensisis the closest volant relative, andother Gallirallus taxa from the Cooks,Marquesas, and Societies for whichit is uncertainwhether G. philippensis, G. torquatus, oranother congener is the closest volant relative.

is surelythe energetic(and weight) burdenof of the New Georgiagroup, for informationabout G. flight muscle.Muscle is doubly costlybecause rovianae;Hugh Paia,Permanent Secretary of the Min- of high initial investmentof biosyntheticen- istry of Educationand Cultural Affairsof the Solomon ergy, plus the continuing maintenanceexpense islands,for permissionto collectand export G. rovi- due to its high metabolicrate. On extensiveland anae;Mary LeCroy, for specimen measurementsand valuable discussion;David Schwendeman,for pre- masseswith predators,these costs are balanced paring the type as a study skin; JamesCoe, for paint- by the benefitsof dispersaland of escapefrom ing the plate; Stephen Blaber,Brian Finch, StorrsOl- predators.But on oceanicislands free of mam- son, and H. Price Webb, for information on rail malian predators, reduction of flight muscle specimensand sightings;and the National Geograph- bringsgreat energy savings with little penalty. ic Society,for supportof fieldwork. Convergentevolution of flightlessinsular rails is probablymuch more frequent than indicated. LITERATURE CITED Of the flightlessor weak-flying taxa listed (Ta- ble 3), half of those that were extant at the time BALOUET,J. C., & S. L. OLSON. 1989. birds from of Europeandiscovery have subsequentlybe- Late Quaternary deposits in New Caledonia. come extinct, and at least half of the survivors Smithsonian Contrib. Zool. 469. are now endangered--victimsespecially of in- BLABER,S. J.M. 1990. A checklist and notes on the troducedmammalian predators. Even more such currentstatus of the birdsof New Georgia,West- taxa must have gone extinct before European ern Province, Solomon Islands. Emu 90: 205-214. CAIN, A. J., & I. C. J. GALBRAITH. 1955. Five new discoveryand soonafter the first human colo- subspeciesfrom the mountains of Guadalcanal nization of remote Pacific islands by Polyne- (). Bull. Br. Ornithol. Club siansand Melanesians,who brought , , 75: 90-93. and dogs.Subfossil remains of extinctflightless DANIS,V. 1938. •tude d'une nouvelle collection or weak-flyingrails have now been found on d'oiseauxde l'ile Bougainville.Bull. Mus. Hist. mostpaleontologically explored Pacific islands Nat., Paris (2) 10: 43-47. (Steadman1989, Olson 1989).Many more surely DIAMOND,J. M. 1987. Extant unless proven extinct? await discoveryas subfossils.The example of Or, extinct unlessproven extant?Conserv. Biol. Gallirallusrovianae shows that some may also 1: 77-79. await discoveryas living birds. 1989. The ethnobiologist'sdilemma. Nat. Hist. 98(6): 26-30. ACKNOWLEDGMENTS --, & E. MAYR. 1976. Species-arearelation for birds of the .Proc. Natl. I am indebted to Alisasa Bisili, who collected the Acad. Sci. USA 73: 262-266. type of G. rovianae;Mr. Bisiliand manyother residents FINCH,B. W. 1985. Noteworthy observationsin Pa- 470 JAREDDIAMOND [Auk, Vol. 108

pua New Guinea and Solomons. Papua New trophe.Pp. 50-53in Conservationfor the twenty- Guinea Bird Soc. Newsletter No. 215: 6-12. first century(D. Westernand M. Pearl,Eds.). New HADDEN,D. 1981. Birds of the North Solomons. Wau, York, Oxford Univ. Press. Wau Ecol. Inst. RIPLEY, $. D. 1977. Rails of the world. Boston, Go- ß 1983. A new speciesof thicket warbler Cich- dine. lornis(Sylviinae) from BougainvilleIsland, North --, & D. HADDEN. 1982. A new subspeciesof SolomonsProvince, . Bull. Br. Zootheta(Aves: Muscicapidae: Turdinae) from the Ornithol. Club 103: 22-25. Northern Solomon Islands. J. Yamashina Inst. MAYR,E. 1942. Systematicsand the origin of species. OrnithoL 14: 103-107. New York, Columbia Univ. Press. SIBLEY,C.G. 1951. Notes on the birds of New Geor- 1945. Birds of the Southwest Pacific. New gia, Central SolomonIslands. Condor 53: 81-92. York, Macmillan. STEADMAN,D. W. 1986. Two new speciesof rails 1949. Notes on the birds of Northern Mel- (Aves: Rallidae) from Mangala, southern Cook anesia, 2. Am. Mus. Novit. No. 1417. Islands. Pacific Sci. 40: 27-43. 1969. Bird speciationin the tropics.Biol. J. 1988. A new speciesof Porphyrio(Aves: Ral- Linn. Soc. 1: 1-17. lidae) from archaeologicalsites in the Marquesas OLSON, S. L. 1973a. A classification of the Rallidae. Islands. Proc. Biol. Soc. Wash. 101: 162-170. Wilson Bull. 85: 381-416. 1989. of birds in Eastern Poly- 1973b. Evolution of the rails of the South nesia: a review of the record, and comparisons Atlantic islands (Aves: Rallidae). Smithsonian with other Pacificisland groups. J. Archaeol.Sci. Contrib. Zool. 152. 16: 177-205. 1975. A review of the extinct rails of the $TRESEMANN,E., •a:V. $TRESEMANN.1966. Die Mfiuser New Zealand region (Aves: Rallidae). Nat. Mus. der V6gel. J. Ornithol. 107 (Sonderheft):1-445. New Zealand Rec. 1: 63-79. WATERHOUSE,J. H. L. 1949. A Roviana and English 1986. Gallirallussharpei (Bi•ttikofer), nov. dictionary, rev. ed. Sydney, Epworth. comb. A valid species of rail (Rallidae) of un- YAMASHINA,Y., & T. MANO. 1981. A new speciesof known origin. Gerfaut76: 263-269. rail from Okinawa Island. J. Yamashina Inst. Or- 1989. Extinction on islands: man as a catas- nithol. 13: 147-152.