THE GALAPAGOS AVIFAUNA

M. P. HARRIS

Edward Grey Institute of Field Oxford University and Research Station Gakpagos, Ecuador

The last review of the of the Galhpagos Haematopus ostralegus ( H. palliatus) , Himan- archipelago was made by Swarth (1931) al- topus himantopus (H. mexicanus), Tyto alba though L&&que et al. (1966) summarized the (T. punctatissima), Asio flammeus (A. gala- migrants on and near the islands. Several pagoensis), Pyrocephalus rubinus ( P. nanus authors (Brosset 1963; Curio and Kramer 1964; and P. dubius). The nomenclature and se- Harris 1968, 1969a,b,c,d, 1970a,b; L&$que quence of the migrants follows that used by 1963,1964; Nelson 1968; Snow and Snow 1969) LBv&que et al. (1966). Overall, the of have published distribution maps and/or is very close to that of the 1957 edition counts of several species of resident birds but of the A.O.U. Checklist of North American the status of the birds needs to be looked at as Birds. a whole to determine recent changes, due The naming of the islands is as much a either to man or other factors or to gaps in our problem as naming the birds (see Bowman knowledge. This paper attempts to document 1960). Most islands have several names, En- the past and present status of all the birds glish and Spanish. I hope that this summary known to have occurred within the islands or will be useful to field workers, therefore I have in the seas immediately surrounding them, and used the names now in common usage, but to suggest some possible reasons for the dis- give below a list of alternatives. The official tributions of the land birds, Only a handful Spanish name is italicized. The relative loca- of wader species, Heteroscelus incanus, Nu- tion of each island is shown in figure 1. menius phaeopus, Arenaria interpres, Charad- rius semipalmatus, and Lobipes lobatus can be Main Islands considered to use the GaKpagos as regular Name in common use Alternatives wintering grounds. Migrants are not discussed Isahela Albemarle in detail (though all records known to me by Fernanclina Narborough James Santiago, Salvador 1 January 1973 are given) because of the lervis Rdhida review cited above. For general information Duncan Pinzon’ on the islands and the flora, the reader is Santa Crux Indefatigable referred to the accounts of Thornton (1971) San Cristdhal Chathan; Barrington Santa Fe ’ and Wiggins and Porter ( 1971) , respectively. Hood Espafiokz Floreana Santa Maria, Charles NOMENCLATURE Tower Genovesa Marchena Bindloe For the names of the resident species, I have Pinta Abingdon used Swarth (1931) as a basis but there have Culpepper Darwin been many generic changes since this time. Wenman Wolf I have incorporated the modifications sug- Minor Islands (where different) gested by Lack (1945, 1947, 1969a) for the Seymour North Seymour Geospizinae (Darwins’ ) as well as a Baltra South Seymour few other published opinions. Tortuga Brattle Inconsistencies are inevitable but no attempt Crossman Islands Los Hermanos is made to discuss the taxonomic status of SPECIES LIST species. is used here solely as a convenient “pigeon-holeing” system for nam- A general statement is made of the status of ing birds. Changes involving the specific each species, followed, where appropriate, by name have been made in the following species remarks on its status on each of the main (with Swarths’ nomenclature in parentheses ) : islands. Except where explicitly mentioned, Puffinus lherminieri (P. obscurus), Sula sula Seymour and Baltra are included with Santa ( S. piscator), Casmerodius albus ( C. egretta) , Cruz, Gardner-near-Hood with Hood, and the [2651 The Condor 75:265-278, 1973 266 M. P. HARRIS

0 CULPEPPER

0 WENMAN

I 50MILES I I

0 PINTA

OTOWER R°CA REDONDA 1 MARCHENA

------w-_-e______EQUATOR -----me_ ------____

CALETA PTA. MANGLE TORTLJGA

I- WRECK BAY VILLAMIL PT?_ I/ CHAMPION c-2./GARDNER GARDNER FLOREANA -PTA. CEVALLOS

/ HOOD f%. SUAREZ

FIGURE 1. Map of Galapagos Archipelago showing places mentioned in the text. various offshore islands of Floreana with Flo- names in the acknowledgments; my own reana, partly because we have relatively little records are designated (MH). For the sake information on the smaller islands and also of brevity, the sources of older records con- because it seems that many of the birds (ex- sidered by Swarth (1931) and Lkveque et al. cluding mockingbirds) from the small islands (1966) are omitted. Where recent specimens travel regularly to and fro from the main are available, I have designated their location islands. The other small islands are not men- as follows: Charles Darwin Research Station tioned in detail as observations are very sparse. ( CDRS ) , California Academy of Sciences Current ornithology in the islands is based (CAS ), and American Museum of Natural on the living birds, so that there are, in most History ( AMNH) . cases, no specimens to support recent records. Spheniscus men&c&s. Galapagos Penguin. En- For the most important sight records, the demic. Breeds in suitabIe areas around the coasts of initials of the observer(s) are given in paren- Fernandina and western Isabela (from just north of Cape Berkeley to southwestern tip ). Nonbreeding theses after the record. Identification of any birds, especially immatures, are sometimes seen else- set of initials can be made by comparison with where on Isabela, James, Santa Cruz, and Floreana. THE GALAPAGOS AVIFAUNA 267

Birds have been recorded displaying on Santa Cruz islands and cliffs of some larger islands throughout the and James but there is no evidence of breeding. Rare archipelago. Rarely recorded at sea. elsewhere but has been recorded on Tervis, Duncan, Oceanodroma tethus. Wedge-rumned Storm Petrel. and Hood. Total population probably in region‘ of Resident Nests on Tower lci 200,600 pairs (Harris 3-5000 birds (DB, MH). 1969a)l, Isla Pitt, and almost certainly on Rota Podilymbus podiceps. Pied-billed Grebe. Recorded Redonda. Common at sea. three times in mangrove lagoons at Academy Bay Oceanodroma markhami. Sootv Storm Petrel. Soec- ( Santa Cruz) : March lQ60;~ 15-19 ’ December- 1961; imen collected off Fernandina, I2 July 1925. _ 13 Tanuarv-6 Februarv 1966 I MC. MH ). Fregetta grallaria. White-bellied Storm Petrel. i)iomedea irrorata. ’ Waved Albatross. Endemic. Specimen collected SE of Isabela, 11 June 1906 and Breeds on the southern slopes and coasts and the a sight record. 17 Tulv 1948. N of Pinta. extreme eastern tip of Hood. In 1970-71, the popula- Pilecanus &c&n&is. Brown . Resident. tion was estimated at 12,000 pairs, far higher than Breeds commonly on all the central islands and has previous, but incomplete counts (Brosset 1963; L&v&- also bred on Hood and Marchena. Uncommon visitor que 19164). Common at sea throughout the archi- to Tower and Pinta; single record of two adults and pelago except January to March when birds migrate an immature on Wenman, 23 February 1962, ( RL ). to waters off Ecuador and Peru. An albatross seen off Unrecorded from Culpepper. Pinta, 17 July 1948, was probably D. exuluns or D. Sulk sula. Red-footed Booby. Resident. Breeds on epomophora. Tower [ca. 140,QOO pairs (Nelson 196S)J, Wenman, Daption capense. Cape Pigeon. Recorded 23 Octo- Culpepper, Gardner-near-Floreana, Punta and Isla ber 1961 and 7 October 19165 (MC) but for July- Pitt ( San Cristobal). Sometimes seen ashore in trees November lQ72 there is a total of 13 records, includ- where it does not breed, e.g., Seymour, Punta Cevallos ing corpses washed ashore on Santa Cruz and Isabela (Hood), and Rota Redonda. (DD, JG, TR, MH). Sea conditions during the latter Sula dactulatru. Masked Boobv. Resident. Com- part of 1972 were atypical (high sea water tempera- mon breeding species throughout the archipelago. tures) and many unusual records occurred at this Often roosts on islands and rocks where it does not time. breed. Puffinus pacificus. Wedge-tailed Shearwater. Sula nebouxii. Blue-footed Booby. Very common found eaten by Asio fkzmmeus on Plaza, 19 September resident. Breeds on most islands except those north 19866 (MH; specimen CDRS ). The species is not of the equator (but has bred on Tower). Often roosts uncommon between Galapagos and Panama. where it does not nest. Puffinus griseus. Sooty Shearwater. Dead bird Nannopterum harrisi. Flightless Cormorant. En- washed ashore at Punta Tortuga (Isabela), 16 August demic. Breeds on sheltered rocks and beaches on 1971 (MH; specimen CDRS) and a live bird ashore Fernandina and Isabela (except south coast from at Villamil 14 January 1972 (JG). Six sightings May- Punta Garcia to Punta Moreno). Total population June 1972 ( BJ, RT, DD ) . 1970-71 about 800 pairs (MH). Unrecorded outside Puffinus carneipes. Pink-footed Shearwater. A live the breeding range. bird ashore at Villamil (Isabela), 15 January 1972 Fregata mugnificens. Magnificent . Res- (JG) and sight records off Duncan (one bird) and ident. Main breeding colonies on Seymour, Gardner- off Punta Albemarle (three), 2& October and 22 near-Floreana, near Punta Moreno ( Isabela), Wreck November 1972, respectively (MH) . Bay, and Kicker Rock (San Cristobal). A few pairs Puffinus lherminieri. Audubons’ Shearwater. Resi- breed on Daphne Minor and in colonies of F. minor dent. Nests on cliffs throughout the archipelago. on Wenman, Tower, Tortuga Island, and Punta Pitt Procellaria parkinsoni. Parkinsons’ Petrel. Single ( San Cristobal). birds shot off San Cristbbal, 14 October 1905; off Fregata minor. . Resident. Main Floreana, 4 May 1906; and well south of Isabela 18 breeding colonies on Hood, Gardner-near-Floreana, June 1906, the only records for this area of the Pacific. Tortuga Island; also on Crossmans, Isla and Punta Pterodroma phaeopygia. Dark-rumped Petrel. Fairly Pitt (San Cristobal), Tower, Culpepper, and Wen- common resident. Breeds in the humid and thickly man. Both species of roost on rocks and vegetated highlands of Isabela (Santo Tomas, Alcedo), in trees at places where they do not nest. James, Santa Cruz, San Cristobal, and Floreana. Num- Ardea herodias. Great Blue . Resident. bers have been severely reduced by introduced preda- Breeds only in coastal areas of the main islands. tors and the clearing of the nesting areas for agricul- Rare on Hood and Tower; unrecorded on Culpepper, ture ( Harris 1970b). Wenman, Marchena, and Pinta. Pelagodroma marina. White-faced Storm Petrel. Butorides striatus. Striated Heron, and Butorides Three records within the islands: a bird aboard ship sundeualli. Heron. The situation regarding these off San Cristobal. 9 Tulv 1967 (Harris and de Vries two species is confused. B. sundewalli is endemic and 1968); and sight records off Santa Cruz, June 1968 and common in the coastal areas of all the islands, and off James, February 1969 (TR). In addition, speci- until recently there had been no suggestion that mens have been collected within 120 miles of the archi- another Butorides sp. ever occurred in Galapagos. pelago 18 June 1966 and September 1930 and a few However, birds resembling B. striatus have been other sight records were made further offshore. recorded on Isabela (nesting), Santa Cruz (nesting), Ocealodroma Zeucorhoa. Leachs’ Storm Petrel. Fernandina (see fig. ’ 2), Duncan, Santa Cruz, San Four records of birds aboard ships within the islands: Cristobal, and Pinta. Present knowledge is insufficient two off Tower, 11 February 1970 (MH); single off to ascertain whether there are two species of Butorides Hood, 14 April 1970 (MH); and single off Floreana, breeding in Galapagos which hybridize (as inter- 17 May 1971 (GM). A sight record was made near mediates occur) or whether there is one, very variable Wenman, 22 February 1971 ( MH ) . Several specimens species. If B. striatus does indeed breed, it is not a have been taken in this area of the Pacific, so may be recent colonizer since skins assignable to the species commoner than these records suggest. were collected by the Academy Expedition 1905-6 Oceanodroma Castro. Band-rumped Storm Petrel. (L. C. Binford, unpubl. data). Specimens of B. stria- Fairly common resident. Breeds on various small tus are at CDRS, AMNH, CAS. 268 XI. P. HARRIS

Occurs on all the main islands where there are salt lagoons or permanent or temporary freshwater. Largest numbers seen in the crater lake of Fernandina. Anus discors. Blue-winged Teal. Regular migrant in small numbers. Commonest October to March but has been recorded in most months. Normally fre- quents freshwater ponds. Buteo galapagoensis. . Endemic. Previously common on all the main islands except Tower. Culoeooer.. __ and Wenman. Now extinct on Floreana and San Cristobal, very reduced in numbers on Santa Cruz. Visits but does not breed on Jervis. Total population about 130 pairs (TV). Pun&on haliaetus. Osprey. Migrant in very small numbers. Occasional birds spend the northern summer in the islands. FaZco peregrinus. Peregrine Falcon. Migrant in very small numbers. A total of 16 dated records, most between 21 November and 14 March but also re- corded 16 May 1970 and 20 June 1967 (TV; MH). Laterallus spilonotus. Galapagos Rail. Endemic. Common in the highlands of James and Santa Cruz, formerly very common on Pinta but now less so. Also breeds on Floreana, probably on Isabela, possibly on Fernandina. Recorded normally from the wetter in- land areas but rarely in mangroves. Neocrex erythrops. Paint-billed Crake. Resident. FIGURE 2. Striated Heron (Butorides striatus) at Nests on Santa Cruz and Floreana (specimens CDRS). Punta Espinosa, Fernandina, July 1971. Recorded since 1953 (Bowman 1960) but possibly overlooked for sometime, could occur on San Cristhbal and Isabela. Dead bird on Tower, Dec. 1972 (Kumer- For many years it was not realized that there were loeve, pers. comm.). two species of Butorides, B. hiatus and B. virescens, Porphyrula martinica. Purple Gallinule. A recently on Cocos Island and it is still not certain which breeds dead adult found off Santa Cruz, 13 February 1964 or whether both are resident (Slud 1967). The tax- and a bird caught alive and photographed at Punta onomy of the group obviously needs some critical Suarez, February 1972 ( DD ) An immature aboard attention. ship between Galapagos and Ecuador, 20 February Bubulcus ihis. Cattle Egret. Regular migrant in 1968 (Castroand de Vries 1970) (specimenCDRS ) . small numbers. First identified in 1964 and now GaZZinuZu chloropus. Common Gallinule. Resident. recorded from most islands, not only in the highlands Confined to brackish lagoons and a few inland pools with cattle but in sea lion colonies (e.g., Plaza, Hood ) on San Cristobal, Santa Cruz, southern Isabela, and Extreme dates 2 August to 29 April. Floreana. Heard calling at the crater lake on Fernan- Casrnerodius ulbus. Common Egret. Resident. Oc- dina, July 1957 (RB). curs in small numbers in the coastal areas of the main Haematopus ostralegus. Oystercatcher. Resident. islands; rarely recorded inland. Rather few nesting Scattered pairs found on the coasts of main islands; records but there is no evidence that the species is total population very small, probably less than 100 a migrant to the area. An all-dark heron, slightly pairs. Most numerous on Hood where there are 10-12 smaller than this species, was seen at Turtle Bay breeding pairs. (north Santa Cruz) sometime in 1967 (CA). Gharadrius semivalmatus. Semiualmated Plover. Leuco$royx thuZa. Snowy Egret. Migrant in small Common migrant. Small numbers always present but numbers. First recorded 28 December 1965 (MC ), the species is most numerous August to April when and since this date one or two have been seen ir- flocks of over 50 birds have been seen. regularly in the southern part of Santa Cruz, Turtle Charudrius wilsonia. Wilsons’ Plover. Three birds Bay (Santa Cruz), southern Isabela, and at Punta seen at Punta Cormorant (Floreana), 108 May 1969 Cormorant (Floreana) (MH). At least one bird has (RTP). been seen in full breeding , so it is con- Charadrius vociferus. Killdeer. An adult photo- ceivable that the species might one day nest. Extreme graphed at Punta Tortuga ( Isabela), 8 February 1971 dates are July to 21 April. (HRH; MH). Nycticorar nycticorux. Black-crowned . Pluuialis dominica. Golden Plover. Two birds seen A subadult seen at Turtle Bay (Santa Cruz), 1 April at Tortuga Bay ( Santa Cruz ), 18 November 1962 1971 (HD). (Hatch and Hailman 1967 ). Nyctanassa uiolacea. Yellow-crowned Night Heron. SquataroZa squatarola. Black-bellied Plover. Reg- Resident. Breeds on all the main islands with the ular migrant in small numbers, recorded in all months possible exceptions of Culpepper and Wenman. but most common during northern winter. Phoenicopterus ruber. Greater Flamingo. Resi- Aphriza virgatu. Surfbird. A migrant in very small dent. Seen regularly on a few salt lagoons on Isabela, numbers. Recorded on James, 25-26 October 1961; James, Bainbridge Rocks, Jervis, Santa Cruz, and Santa Cruz, 6 November 1961; Plaza, 21 October 1966 Floreana. Breeds or has bred on all these islands. ( MH); and at Villamil ( Isabela), 20 March 1970 Total population over 501) but less than 1000 birds. ( MH), 7 September 1971, and April 1972 (JG). Dendrocygna autumnaZti. Black-bellied Tree . Arenaria interpres. Ruddy Turnstone. Very com- Single at Villamil (Isabela), 2 July 1960. mon migrant to rocky shores, rare inland. Always Anus bahamensis. White-cheeked Pintail. Resident. present but most common August to March. THE GALAPAGOS AVIFAUNA 269

Arena& melunocephalu. Black Turnstone. Single there is the possibility of confusion with the Long- bird seen with two A: interpres at a pond in the high- billed Dowitcher (L. scolopuceus) which has been lands of San Cristobal, 29 September 1968 (MC). recorded on the Ecuadorian mainland. Wumenius phueopus. Whimbrel. Migrant in large Ereunetes pus&s. Semipalmated Sandpiper. Three numbers. Although recorded all months, it is rather collected on Floreana, 24 November 1961 and two uncommon from end of April to mid-July. Up to 500 seen there, 14 March 1971 (MH); sight records at birds counted on a roost at Villamil ( Isabela) (TV). James Bay, 28 October 1972 and at Academy Bay, Act&s muculuriu. Spotted Sandpiper. Regular mi- 18 November 1972 (MH). Probably overlooked due grant in small numbers and not recorded every year. to difficulties in field identifications. Extreme dates 23 August and 2 May. Ereunetes muuri. Western Sandpiper. Few definite Tringu solituria. Solitary Sandpiper. Rare migrant. records: on James, 26 October 1961, 20 September Recorded mainly from the highlands of Santa Cruz, 1968 (EU), 2#8 October 1972 (MH); Floreana, 24 15 October 1961, 21 September 1964 to 7 April 1965 November 1961 and 2 December 1972 CMB): and (MC), September 1965 to 19 April 1966 ( MC), 30 Barrington, 30 October 1970 (MH). This or‘ E. October 1967’ (MC). Only other records are inland pusillus seen also on ca. 15 other dates. on San Cristobal. I2 October I897 and December Limosu fedou. Marbled Godwit. Several seen at 1938 to February 1939, and on the north coast of Academy Bay (Santa Cruz), 9 November 1957; a Santa Cruz, 2 October 1968’ (EU). single at Villamil (Isabela), ’ 6 January to end ’ of Heteroscelus &anus. Wandering Tattler. Common Xlarch (TV. MH). and start of August 1971 ( TG). migrant. By far the most common shore bird in Crocethia alha. ’ Sanderling.‘ Common migrant with‘ Galapagos, which must be one of the major wintering flocks of over 100 birds on sand beaches. Some birds grounds for the species. Recorded throughout the year oversummer. but only small numbers during the northern summer. Himuntopus himuntopus. Common Stilt. Resident. Cutoptrophorus semipulmutus. Willet. Regular mi- hlost common on salt lagoons but also occurs and grant in very small numbers, although up to 14 have nests on freshwater ponds and crater lakes. been seen together. Recorded in all months but rare Phuluropus f&cur&. Red Phalarope. A specimen May to July. Most frequently seen on sand beaches taken off Fernandina, 25 August 1929; a single bird at Villamil ( Isabela) and Tortuga Bay ( Santa Cruz ) . seen at Caleta Tortuga (San Cristobal), 16 December Totunus melunoleucus. Greater Yellowlegs. Rare 1965 (MC): five birds at Punta Cormorant (Flo- migrant recorded April 1923, 29 April 19832, February reana ), 11 December 1965 (TV) ; and many records to March 1939, 2 April and 31 July 1960, 20-21 Jan- in large flocks of L. Zobutns October-November 1972, uary 1962, 30 September to 7 October 1964 (MC ), a time when many other unusual sea birds were 16 February 1966 (MC), 4 January 1967 (TV), 6 to recorded (MH). Possibly overlooked in large flocks 11 April 1971 (JG), April 1972 (JG), November- of L. lob&us but in most years must be rare. December 1972 (JG, MH). Lohipes lobutus. Northern Phalarope. Common mi- Totanus f&pa. Lesser Yellowlegs. Migrant in grant. Numerous 1 August to 20 April, with flocks small numbers. Recorded in nine northern winters sometimes numbering several thousand birds during between 8 October and 9 May. Maximum numbers December and January. Normally on the open sea but recorded 193839, when up to I3 frequented the lake small numbers on salt lagoons. of El Junco (San Cristobal), 1965-66 (four records) Stegunopus tricolor. Wilsons’ Phalarope. Migrant. (MC, TV, MH), and end of 1972 (ca. 10 records) Small numbers occur between August and December (DD, MH). but odd birds may be seen at any time. Usually on C&&is cunutus. Knot. Single bird at Punta Cor- salt lagoons, rarely on tidal pools, and unrecorded on morant (Floreana), 10 May 1969 (RTP). the open sea. Maximum flock 20 birds. E~olia melunotos. Pectoral Sandpiper. Singles re- Stercorurius pomurinus. Pomarine Jaeger. An im- corded in Santa Cruz highlands, October 19864 and mature shot off Isabela. 15 December 1897 and three 1 October 1967 (MC). seen between Hood and Floreana, 8 December 1971 Eroliu b&&i. Bairds’ Sanduiuer. Birds collected (RF). Unidentified jaegers seen 6 July 1961 and 22 on Barrington, 6 October 1897* and on Hood, 23 November 1972 (MH). September 1957, and two found dead at James Bay, Cathuructu skuu. Skua. Single bird seen off Cape 28 October 1972 ( MH; determined R. C. Laybourne) Berkeley (Isabela), 12 March 1966 (MC) and a Eroliu fuscicollis. White-rumped Sandpiper. Single “probable” off Floreana, 17 January 1971 (CA). bird Cape Hammond (Fernandina), 1 November Larus utricillu. Laughing Gull. An adult collected 1972 and two photographed Punta Cormorant (Flo- off Santa Cruz, 19 February 1962; a bird seen off reana), 2 December 1972 (MH). Isabela, 12 March 1960; and several seen and photo- Eroliu minutillu. Least Sandpiper. Migrant in small graphed at several places in latter Dart of 1972, (DD. numbers in most years, the most common “peep.” KIti). “Probables-seen’‘ on several dates but specific Extreme dates 8 September and 9 May, and largest separations from L. pip&can are difficult as most birds flock 15 birds. are immature. Micropuluma himuntopus. Stilt Sandpiper. Three Lurus fuliginosus. Lava Gull. Endemic. Widely records: near Villamil (Isabela), 27-28 ’ April 1932, but sparsely distributed throughout the archipelago; and 22 February 1966 (MC, TV), and Santa Cruz, most numerous near human settlements. Total pop- 8 October 19162. ulation less than 400 pairs ( Snow and Snow 19698). Limnodromus griseus. Short-billed Dowitcher. Lurus pip&can. Franklins’ Gull. Migrant in very Dowitchers recorded near Villamil (Isabela), 14 March variable numbers, possibly increasing. Usually seen 19860, 4 January 1967 (TV), May 1971, January, from October to May; largest count 35 at Wreck Bay April, November and December 1972 ( JG); three (San Cristobal), 7 December 1967 (TV). Most birds birds (1 collected) on James, 26 October 1961 and are immature. two there, 21 February 1971 (MH); a single bird Creugrus fzcrcutus. Swallow-tailed Gull. Endemic collected on Floreana, 24 November 1961; and five [though a few pairs nest on Malpelo Island ( Bond and birds at Caleta Tortuga (San Cristobal), 20 December de Schauensee 1938) I. Nests commonly throughout 1965 (MC, MH) Apart from the collected specimens the archipelago except eastern and southern Fernan- 270 M. P. HARRIS dina and western Isabela. Total population probably Meguceryle ulcyon. Belted Kingfisher. Migrant in slightly higher than the 10,000 pairs estimated b) very small numbers. Recorded in seven northern L&&aue ( 1964) and Harris ( 1970a) due to the _ I winters between 1961 and 1971, extreme dates 20 discovery of several new colonies, e.g., western Fer- October to 20 March. Most records are from Academy nandina. Bay (Santa Cruz) and are probably due to concen- Sterna hirundo. Common Tern. Migrant in very tration of observers rather than of the birds. Also small numbers. An immature collected on Hood, 26 recorded on the north side of Santa Cruz, Daphne, November 1961 and two (out of a flock of 7) photo- Tower, Hood, Isabela, and Floreana. graphed at Villamil ( Isabela), March 1970 (TR ). Myiurchus mugnirostris. Large-billed Flycatcher. About 20 other records could refer to this or S. Endemic. Widesmead on the main islands exceDt pamdisaeu, especially common in latter part of 1972. Tower (recorded 14 September 1906), Wenman Extreme dates August and March. (three on 24 September 1906). A bird caught at Sterna fuscutu. Sooty Tern. Resident. Large num- sea near Culpepper 25 July 1897. bers breed on Culpepper but birds are rarely seen Pyrocephulus rubinus. Vermilion Flycatcher. Resi- elsewhere in the islands because the population feeds dent. Breeds on most main islands except Barrington in the warmer waters north of the archipelago. Rarely (five specimens, last seen 1929), Jervis (seven speci- heard at night calling at other sea-bird colonies. A mens, last 1906), Wenman (single 24 September bird killed by Asio flummeus on Tower in 1966 (MH). 1906)) and Hood [single sight record December 1970 Thulusseus maximus. Royal Tern. Migrant in very ( MH ) 1. Unrecorded on Culpepper and Tower. small numbers. Up to nine recorded at or near Villa- Rip&a riparia. Bank Swallow. Migrant. Recorded mil (Isabela) in 10 years between 1960 and 1972. on 12 dates spread throughout the year. Specimen Most records are between January and March but CDRS. noted in all months. Wirundo rustica. Barn Swallow. Migrant. Occurs Anous stolidus. Brown Noddy. Resident. Common most years in small numbers during the northern nester on cliffs throughout the archipelago. winter. Some hirtmdines either oversummer or, more Gygis &a. Fairy Tern. Single sighting off Tower, likely, survive that long before dying. 14 September 1906. Petrochelidon vvrrhonotu. Cliff Swallow. Single Zen&u gulupugoensis. Galapagos Dove. Endemic. bird seen at sea some‘ 60 miles north of Culpepper, Widespread on all the main islands. Most common 13 April 1932. on Fernandina, James, Barrington, Hood, and Tower Progne subis. Purple Martin. Migrant in small islands where there are few or no feral cats. numbers. First recorded October 19864 (MC) but Coccyzus erythopthalmus. Black-billed Cuckoo. .4n possibly overlooked before this date. Since 1964, immature found dead on Hood. 16 Mav 1970 (MH; females or immatures have been recorded on various determined L. C. Binford; spe&en CDkS). dates from Hood and Santa Cruz. Between June and Coccyzus melucoryphus. Dark-billed Cuckoo. Resi- December 1970 at least four were seen regularly at dent. Common on Isabela, Fernandina, Santa Cruz, Punta Suarez (Hood) (MC). Skin of an immature San Cristbbal, Floreana; uncommon on Duncan, rare at CDRS. The only record of males, two on Hood on James (apparently unrecorded before 1966 and on 17 October 1972 (MH ), may be due to confusion now only known from northwest coast and two areas with the similarly all-dark P. mode&z. inland), and a single record 10 July 1906 on Bar- Progne modestu. GalQpagos Martin. Endemic. rington. Found in small numbers throughout the archipelago CTotophugu sulcirostris. Groove-billed Ani. Single although unrecorded in the five most northern islands birds inland on Isabela, April 1962 (CA); Santa Cruz, and Jervis. Very rare on Hood and not resident there. 29 March 1966 (MC); and James, March 1967 (TR). Nesomimus trifusciutus. Charles Mockingbird. En- Because of its association with cattle and its local demic. Breeds on Champion and Gardner-near-Flo- reputation for removing ticks from cattle, it is possible reana with population of about I50 birds (Harris that these birds were introduced by farmers. The 1968). Now extinct on the main island of Floreana. seemingly weak flight might also support this view. Neiomimus mucdonuldi. Hood Mockingbird. En- Tqto ulbu. Barn Owl. Resident. Common on Fer- demic. Common resident on Hood and Gardner-near- nandina; uncommon on Isabela, Santa Cruz, James, Hood. and San Crist6bal: nrobablv extinct on Floreana. Nesomimus melanotis. Chatham Mockingbird. En- Asio flummeus. ’ Short-eared Owl. Resident. Re- demic. Common on San Crist6bal. corded from all the main islands except Wenman. Nesomimus par&us. Gal&pagos Mockingbird. En- Two records from Fernandina [specimen collected 9 demic. Common on all the main islands except those January 1929 is in the Chicago Field Museum (per with the three species above and Duncan, which lacks TV) and a sight record 3 July 1971 (DlB)], and a resident mockingbird. The two records of mocking- several recent sight records from Jervis. Commonly birds on Duncan, 14 August 1906 and 6 December seen on the smaller islands with nesting sea-birds but 1968 (JB), are probably referable to this species. not known to nest in such areas. Assumed to be Dendroicu vetechiu. Yellow Warbler. Resident. resident on Culoenner_ ^_ where seen twice, 29 Tulv 1897 Common on all the main and most of the smaller and January 1964 (RB), on two of the‘ -ve;y few islands. Very widespread but apparently prefers landings on the island. greener areas than Certhideu oliuaceu, where both Chordeiles minor. Common Nighthawk. Two birds occur on dry islands. seen at Punta Suarez (Hood), 22-24 November 1963 Dolichonyx oryzivorus. Bobolink. Migrant. Most and a single at Punta Cevallos (Hood), 12 December commonly seen between October and December but 1970 (MH). Although it is sometimes difficult to odd birds in other months. Numbers vary greatly separate this species from C. ucutipennis, the field from year to year. characters of these birds were exactly right for C. Pirungu rubru. Summer . A mummified minor. A bird heard calling at night inside the crater adult male found on Hood, 30 August 1963. One or of Alcedo (Isabela), 23 April 1971 (CM) and an possibly two female-plumaged, unidentified unidentified nighthawk seen at Academy Bay (Santa seen in the farm lands of Santa Cruz, October and Cruz), 22 October 1972 (MH). November 1964. THE GALAPAGOS AVIFAUNA 271

Geospiza magnirostr&. Large Ground . En- Swarth ( 1931) and Bowman ( 1961) would prefer demic. Status difficult to elucidate for field identifica- them placed in G. magnirostris. A party which spent tions on strange islands are thought to be unreliable 2 days on the top of the island in April 1968 failed because of the difficultv. if not imnossibilitv. of field to see any large-billed geospizids ( RB ) . Possibly the separation from the largest beakkd individuals of collected specimens may refer to stragglers or a very G. fortis. However, it has occurred on all the main short-lived population now extinct, but the possibility islands except San Cristobal, Culpepper, and Hood. still exists that the species still survives in vegetation Three specimens collected on Fernandina, January at the base of the cliffs, where the earlier specimens and April 1899, and two on Barrington, October 1897 were obtained. A single specimen from Pinta, 27 and October 1905, could refer either to stragglers or March 1963 (Curio and Kramer 1965). It is tempting to now extinct populations. The statements of Sam- to assume that the specimen labeled as having been malisto (1967) that 1% of the finches breeding on collected on Gardner-near-Floreana on 31 October Barrington can be assigned to G. magnirostris (sim- 1897 was an error of labeling, but Rothschild and ilarly the 3% referred to G. difficilis) are considered Hartert ( 1899. 1902) were at nains to ooint out that to be due to misidentifications. The status of G. this was not‘ so, so the species-must be- considered a magnirostris on Floreana is obscure. The unusually straggler to this island. large subspecies collected by Darwin with reasonable Pkztyspiza crassirostris. . En- certainty on Floreana was thought now to be extinct, demic. Breeds on all the main islands except Bar- but a bird collected 20 September 1957 may be rington, Culpepper, Wenman, Tower, and Hood, most assignable to this form (Bowman 1961). Another G. of which are too dry to support a purely vegetarian magnirostris [considered by Swarth ( 1931) to be an species. Four specimens collected on Duncan in 1897 abnormal G. fortis] collected there by the Academy and 18919 but not seen there since. Expedition is much smaller and probably refers to parut&. . En- a straggler from another island. The species should demic. Resident on all the main islands except Cul- probably be considered extinct, or nearly so, on pepper, Tower, Hood, Marchena, and Wenman (two Floreana. specimens collected 24 September 1906). fortis. . Endemic. Camarhynchus pauper. . En- Widespread on all the main islands except Culpepper, demic. Found only in the highlands of Floreana, Tower, and Wenman (single collected 24 September where common. 1906). In 1905-6. 15 nonbreedine birds were col- Camarhynchus psittacukz. . En- lected on Hood and they may be &her stragglers or demic. Has been recorded, and probably has bred, representatives of a now extinct population (see later). on all main islands except Hood, Tower, Wenman, An early record from Tower is now known to be due and Culpepper. However, status uncertain on San to a mislabeled specimen (Trimble 1943: Parkes, un- Cristobal (single collected 24 December 1900) and publ. data). - does not now breed on Barrington, though specimens Geospiza fuliginosa. . En- collected 20 October 19106 ( one), .6-7 October 1897 demic. Common on all main islands except Cul- (three). and sight records Mav 1966 CTV) and 16 pepper, Tower, and Wenman (though six individuals September 1968 ( EU). Extinct on Duncan, ’ last seen there 24 September 1906). 1906. Geospiza difficilis. Sharp-beaked Ground Finch. Camarhynchus pallidus. . En- Endemic. Occurs in the humid zones on James, Fer- demic. Breeds on Isabela, Fernandina (probably), nandina, Pinta and common on the arid islands of Santa Cruz, James, San Cristobal, and Duncan (rare). Tower, Wenman, and Culpepper. Previously in the Also recorded on jervis (specimens 1897 and 1905), humid zone on Santa Cruz but disappeared between Floreana (specimen11 October 1905’ ). and Barring- 1932 and the visit of Lack in 1939 (Lack lQ45). All ton [two sight records including one with cactus spine, subsequent records of this species on Santa Cruz 15 Sentember 19868 ’ (EU) and a sight record on Pinta. almost certainly refer to misidentified G. fuliginosa. 24 August 1968 (RP)]. ’ - Darwins’ specimens, which came from either Floreana Camarhynchus heliobates. . En- or San Cristobal (the former is considered more demic. Restricted to the dense mangrove areas on likely), belong to a distinct subspecies which is now eastern Femandina [seen 1971 at Punta Mangle (PK) extinct. Rothschild and Hartert ( 1899) reject a speci- but not for many years at Punta Espinosa (MN, TV, men collected bv Markham on Floreana in Februarv PK)] and the west coast of Isabela between just north 1880 on the grounds that the species had never been of Punta Tortuga to just east of Punta Moreno. Speci- recorded there, but since Markham visited no other mens were collected early this century from southeast island in the archipelago, there can be no chance Isabela and a single in August 1971 at Cartago Bay of any mislabeling. Other old specimens from San Isabela (RB) but the status there needs confirmation. Cristobal ( 19 October 1897) and Isabela (4 January The total population must be very small due to the 1901) may also refer to extinct populations. limited . Geospiza scandens. Cactus Finch. Endemic. The Certhidea olioacea. Warbler Finch. Endemic. drier areas of the main islands are the preferred Breeds on all the main islands. habitat but the species is missing from Fernandina and islands with G. conirostris. Now extinct on Dun- GEOSPlZA SPP. ON CROSSMAN can. AND DAPHNE ISLANDS Geospiza conirostris. Large Cactus Finch. En- Ground finches collected on Daphne and Crossman demic. Commonly nests on Hood and Tower; also islands are intermediate between G. fortis and G. bred in 1966 on Wenman [a male seen feeding a fuliginosa. Forty-two specimens collected on Daphne fledged young ( MH ) ] but otherwise only one other have significantly larger (P < 0.05) than 16 record of a specimen collected 8 January 19638 (Curio from the Crossmans (measurements in Lack 1945). and Kramer 1965). The situation of the large-billed Originally, Lack ( 1945) considered these populations Geospiza sp. on Culpepper is confused. Rothschild and to be the results of hybridization between G. fortis Hartert (19082) and Lack (1945, 1969a) refer speci- and G. fuliginosa but later ( 1947, 1969a) decided that mens collected 1897-1906 to this species, whereas they were more likely to be specialized derivatives of 272 M. P. HARRIS

G. fortis (Daphne) and G. fuliginosa (Crossmans). In 1932 Geospiza difficilis was quite com- There have been no identifications of finches on the mon in the highlands of Santa Cruz but it had Crossmans since 1906. These extremely interesting populations need further investigation, if indeed they apparently disappeared by 1939 (Lack 1945). are still extant, as the many individuals I have seen On the other islands where this species also on Daphne were mostly assignable to G. fortis and coexists with the rather similar G. fuliginosa G. fuliginosa.The latter has been collected on Daphne i.e., James, Fernandina, Pinta), G. difficilis (Lack 1945). ( is restricted to the highlands. Some of the CHANGES IN THE AVIFAUNA highlands of Santa Cruz were cleared for agri- It is difficult even today, after more than 10 culture about this time, but the amount of years ’ intensive field work by various ornithol- the habitat suitable for G. difficilis which was ogists working at the Charles Darwin Research affected was small and its removal cannot be Station, to be certain of the distribution of considered responsible for the loss of this bird. breeding land birds. Large fluctuations in If man influenced it in some other way, there numbers of many species occur and the less is no evidence how he did so; but G. difficilis common species may be so reduced that they is now absent on San Crist6bal (where re- are overlooked, only apparent when they corded 1897), Floreana ( 1835, 1880)) and suddenly appear in numbers. Therefore, any Isabela (1901), all islands which would appear species considered to be extinct on any island suitable, and which were the first to be settled could possibly still exist, though future rec- by man. So it is quite possible that on those ords, if any, could in fact refer to individuals islands man was responsible, albeit indirectly, coming from other islands. for the loss of the species. If there are difficulties in present-day GEOSPIZA FORZIS’ ON HOOD records, what of the older ones? Virtually all the older records are substantiated by speci- In 1905-6, the Academy Expedition collected mens but there still remains the chance of 15 specimens of Geospiza fortis on Hood, but mislabeling. The oldest specimens are those despite intensive searches by several people collected by Darwin in 1835, but since he did the species has not been seen since. Lack not realize the significance of interisland dif- (1969a) has considered the Hood G. fortis as ferences until he had collected on San Cris- stragglers from another island, but on ecolog- t6bal and Floreana, there are doubts as to the ical grounds it seems just as likely that they localities of some of his most important spcci- were representatives of a population which mens. has become extinct due to a change of climate Although some species have disappeared and/or the effects of introduced goats. from some islands, as far as is known no Hood is an arid, low-lying island on the indigenous species has become extinct. Some outskirts of the archipelago. It is to be ex- island have been brought about pected that it would have few species of land directly by man, notably Buteo galapagoensis birds. However, Tower, smaller and even on Floreana and San Cristbbal, and probably lower and having far fewer plant species, has Tyto alba on Floreana. The disappearance of three species of seed-eating Geospiza whereas Nesomimus trifasciatus from the main island Hood now has only two. The three species on of Floreana is puzzling. Porter found it com- Tower are the large-billed G. magnirostris, the mon there in 1813 (Beebe 1924) but there arc small-billed subspecies of G. conirostris, and no subsequent positive records, for the speci- the small-billed G. difficilis. Hood has the mens collected by the Beagle expedition could very large-billed subspecies of G. conirostris possibly have come from Champion, which and G. fuliginosa. G. difficilis and G. fuligi- was visited by FitzRoy. It may be that cats nosa probably have closely similar ecological on Floreana were responsible, but this seems requirements, for only one of the two species unlikely because Nesomimus spp. on other is ever found on an arid island. Both can islands have adjusted to introduced . coexist on high islands; G. difficilis is found Similarly perplexing is the absence of any in humid areas and G. fuliginosa, in the arid Nesomimus on Duncan, only 6 miles from zone. However, the factors which influence Santa Cruz where N. parvulus is abundant. which species exists on any single arid island On Duncan, introduced black rats (Rattus are unknown. rattus) have prevented successful nesting of Measurements of Geospiza from Hood and tortoises by eating the newly hatched young Tower are given in table 1. Immediately but, at least to my mind, it seems unlikely that apparent is the great difference between the these animals would have exterminated a land two G. conirostris subspecies. Bowman (1961) bird. Still the problem of mockingbird dis- has shown that the beaks of these birds are tribution in the archipelago remains unsolved. adapted to different foods, and so the seed THE GALAPAGOS AVIFAUNA 273

TABLE 1. Bill lengths and depths (mm) of Geospiza fortis. G. conirostris can cope with the very spp. discussed in the text (from Lack 1945). hard seeds of trees such as Prosopis or Acacia,

Bill length Bill depth and also uses its bill to dig into gravelly

Island and species Sex No. Mean SD No. Mean SD ground for buried seeds. These might also be available to G. fuliginosa with its pointed bill, Tower but not to G. fortis with its intermediate . Geospiza magnirostris $ 29 16.5 0.60 24 21.2 1.13 RECORDS OF THE ACADEMY Q 17 16.3 0.59 11 20.9 0.61 EXPEDITION 1905-6 G. conirostris 8 43 14.3 0.76 40 13.0 0.96 0 25 14.2 0.38 19 12.6 1.20 There are several other land birds besides G. CEifficilis 8 52 9.4 0.38 48 7.9 0.33 those mentioned above which may have be- 0 28 9.2 0.44 25 7.7 0.29 come extinct, but in their cases it is even more Hood difficult to know whether the earlier records 6. conirostris $ 87 15.4 0.94 80 16.0 1.10 refer to breeding populations or to stragglers. 0 52 14.8 0.86 51 15.1 0.91 My own conjectures are given in table 2, based G. fortis $ 8 11.7 0.66 8 12.5 0.84 0 7 11.7 0.86 7 12.1 0.78 on the number of specimens collected in rela- G. fuliginosa $ 32 8.8 0.36 380 8.3 0.33 tion to the collecting effort on each island, on 0 15 8.5 0.43 14 7.8 0.34 the number of years during which specimens were collected, and on whether or not the islands appear to offer suitable habitat. supply from 40 species of plants on Tower is Most of the old records come from the year- being shared among three species of finches. long collecting trip of the California Academy It is tempting to assume that G. conirostris on of Sciences in 1905-6. When compared with Tower has evolved a smaher beak due to 4 years of recent field work ( 1965-72), I find with the large-billed G. magni- their records a great contrast to present-day rostris (Lack 1947). The same might apply to findings. For instance, they spent just a few G. difficilis on Tower, since it is smaller there hours on Wenman and collected in only a very than on any other island. It is, however, pos- small area at the base of the cliffs. Yet they sible that the bill of G. conirostris on Tower brought off the island a total of 13 individuals is the normaI size, and that on Hood the of five species (Myiarchus magnirostris, Pyro- absence of competition with G. mugnirostris cephalus rubinus, G. fortis, G. magnirostris, has allowed G. conirostris to evolve a more Camarhynchus parvulus) which have not been massive beak and to exploit a food supply seen there during six subsequent visits, includ- of larger, harder seeds without competition. ing a 6-week stay by Curio and Kramer. The Whether the observed differences are due to Academy group also took unique records on competition or lack of competition depends on Hood (G. fortis), Tower (M. magnirostris), what the original bill was like, a question and Barrington ( Coccyzus melacoryphus) . which cannot be resolved since there is no Other expeditions between 1835 (the BeagZe) island where the species exists alone. and 1929 (the Pinchot) have collected 10 The large gap between the mean beak sizes species on major islands where they have not of G. conirostris and G. fuliginosa (table 1) on been seen since. Thus, in the course of 90 Hood suggests that there could once have years, 14 species were recorded on islands been an intermediate species, which is sup- where they are not now found. Much more ported by the number of plant species on intensive field work in the last 40 years has Hood (96) compared with those on Tower added only five new island records: G. coni- (40). Considering also the 15 specimens of rostris 011 Wenman and Pinta; C. palliclus on G. fortis, the most likely explanation is that Pinta; P. rubinus on Hood; and C. melacorpy- G. fortis was once resident on Hood. Even if phus 011 James. Additionally, the earlier ex- this conclusion is wrong, there must still be peditions also collected various species on an explanation as to why so many G. fortis smaller islands where they are now lacking: (and others must surely have escaped collec- R4. magnirostris (breeding), Dendroica pe- tion) were present on Hood and why the techia, G. magnirostris and N. parvulus on colonization of the island did not result. Another unanswered question is why G. Daphne minor; C. parvulus on Cowley and fortis, a generalized feeder, and not the more Champion; C. psittaculu, C. parvulus, and C. specialized G. conirostris became extinct on or pallidus on Seymour. On Floreana, C. pauper could not colonize Hood. Possibly the was collected on the coast whereas now, at presence of goats and a change of climate least in most years, it is restricted to the high- would have affected the number of annual lands. It is difficult to escape the conclusion plants which supply most of the seeds for G. that something has changed. 274 hl. P. HARRIS

TABLE 2. Distributionof GalBpagosland-birds on the major islands.

Buteo galapagoensis B B B x B B B E B E B B Zenaida galapagoensis B B B B B B B B B B B B B B B Coccyzusmelacoryphus B b B B B x B Tyto alba B b B B B l& Asio flammeus B x B x B B B B B B B B B b Pyrocephalus rubinus B B B E B B E B x B B B Myiarchus magnirostris B B B B B B B B B B x B B x (i) Progne modesta B B B B B B B x B Nesomimus spp. B B B B x B B B B B B B B B B Dendroica petechia B B B B B B B B B B B B B B B Geospiza magnirostris B (E) B B B B (E) X B B B B G. fortis B B B B B B B B E B B B G. fuliginosa B B B B B B B B B B B ; G. difficilis (:, B B E (E) (E) B B B B G. scandens B B B E B B B B B B G. conirostris B x B X ?B B Platyspiza crassirostris B B B B x B B B B B Camarhynchus parvulus B B B B B B B B B B (E) C. pauper B C. psittacula B B B B E B (E) x B B B C. pallidus B b B x B B x B X X C. heliobates B B Certhidea olivacea B B B B B B B B B B B B B B B

B = almost certainly breeds now. b = probably breeds now. E = once resident, now extinct. (E) = probably once resident, now not present (see text). x = has been recorded, probably simg&zrs (see text).

Further evidence of major recent changes is There is little information on the weather supplied by the records of birds at sea. The earlier in this century, but reminiscences of Academy collected or saw four G. fortis, six the older settlers leave little doubt that the D. petechia, and single G. scandens and Progne islands were much wetter in the 1920s than at modesta quite a distance from land. In addi- present. Possibly the tops of Wenman and tion, a single M. magnirostris was taken in Jervis may have been as humid as the higher 1897 just north of Culpepper. The only land islands and could easily have supported P. bird which I saw far from land was a single rubinus (now missing from both), and M. Zenaida galapagoensis midway between Isa- magnirostris and C. parvulus (now missing bela and Wenman. However, I did see mi- from Wenman). Hood and Barrington would grants, including Arenaria interpres, Crocethia have much more greenery now if it were not alba, Numenius phaeopus, Charadrius semi- for the extreme overgrazing by goats. Fewer palmatus and Bubulcus ibis. Local people, goats and a little more rain would allow P. capable observers who make their living travel- rubinus, C. psittacula, and G. magnirostris to ing within the islands, and an ever-increasing breed on Barrington and G. fortis, on Hood. number of amateur yachtsmen report a similar It is probable that a slight change in climate lack of sightings in the open sea, but doubtless would explain most of the changes noted in some wandering still takes place. Several the avifauna within the last 70 years. Further times I have seen flocks of G. fuliginosa set evidence is needed before one can speculate out to sea from Hood in the direction of whether there has been an overall trend to Floreana, but there was generally evidence drier years or a reduction in the frequency of that they returned, although it is likely that wet years-the “El Nifio” phenomenon. the odd bird or flock might continue on to If there were many interisland wanderings Floreana which is clearly visible. possibly with the establishment of species for THE GALAPAGOS AVIFAUNA 275 perhaps relatively short times following a species although they collected intensively in series of wet seasons, it is difficult to see how where the species is now common. the finches managed to speciate so widely. It The introduction of relatively large-scale land is generally accepted that geographical isola- clearance has either allowed the species to tion is a prerequisite for , but there colonize the island or to become more numer- is little evidence for the time scale of any such ous and widespread. . Johnston and Selander (1964) have Egretta thula was first recorded in 1965. It demonstrated that evolution to what would is becoming a regular visitor and has been normally be considered the subspecies level seen in breeding plumage, so it may well has occurred within no more than 111 genera- breed in the future. tions after the introduction of the House Spar- row (Passer domesticus) into North America. FACTORS INFLUENCING THE NUMBER OF It could well be that the Geospizinae evolved LAND-BIRD SPECIES ON THE MAIN ISLANDS initially at least as fast and that once evolution Lack ( 1945, 1947), in his studies on variation had progressed sufficiently to allow two closely in Darwins’ Finches, concluded that the num- related forms to coexist, further evolution, ber of endemic subspecies of Geospizinae was perhaps by , would be inversely correlated with isolation, namely, the even more rapid. distance between the island and its nearest neighbor. Later, Bowman ( 1961) considered RARE AND ENDANGERED SPECIES that the number of plant species on islands The continuing increase in the human popula- was more important than isolation in influenc- tion both by the high birth rate and immigra- ing both the amount of on islands tion, the associated spread of agriculture, the and the numbers of finch species found on defoliation of the islands by goats, and the any one island. Then Hamilton and Rubinoff destruction of the native fauna by feral dogs, ( 1963-67) treated the problem mathemati- pigs, and cats create a depressing picture of cally. They considered the roles of areas, num- the future of the Galapagos. However, as yet ber of plant species, and isolation (measured there are only a few bird species which can both by interisland distances and distance from be considered endangered. the approximately central island of Santa Cruz) Although it is still relatively common, Ptero- and found ( 1964) that in a linear analysis both droma phaeopygia has declined markedly in isolation and the number of plant species were the last 3040 years since its nesting habitat the best predictors of species numbers on is- makes good agricultural land, the killing of lands-accounting for 48% and 31%, respec- adults by dogs and pigs, and a very low nest- tively, of the observed variation. But in a linear- ing success due to rats (Harris 1970b). This nonlinear analysis, the log of elevation and the species also nests in the Hawaiian archipelago log of distance from the center of the archi- but there too it is endangered. Unless land pelago were the most important factors (49%, clearance can be stopped and some of the 26% ) . Later ( 1967)) they noted that numbers predators removed, the future of this species of insular species and endemics are respec- will be insecure. tively predicted by average and nearest-neigh- The populations of several of the birds are bor isolation, and not by area.-“only the small and there are no reasons to suggest Darwin Finches demonstrate emphatically the that the rarest were ever more common, e.g., importance of isolation in regulating endem- Phoenicopterus ruher (500-1000 birds ), Nan- ism and species abundance in the adaptive nopterum harrisi ( 800 pairs ) , Spheniscus radiation of monophyletic bird groups within mendiculus (few thousand individuals), Laws archipelagos.” In other archipelagos, island fuliginosus (less than 400 pairs), and Hae- area was a better predictor of species abun- matopus ostralegus ( 100 pairs). Buteo galu- dance. Lacks’ (1969a) recent survey showed pagoensis, although reduced in number by that the range of each Geospizid, or rather the man, is quite secure on the uninhabited islands lack of each species on any particular island, and might be able to recolonize lost areas if could be explained either by the absence of not molested. suitable habitat or food or by the presence of a species with which it was likely to compete. ADDITIONS TO THE AVIFAUNA Lack suggested that the findings of Hamilton The only addition to the breeding birds in and Rubinoffs’ (1967) evaluation that isolation recent years is Neocrex erythrops, which was was of the greatest importance might be due discovered in 1953 (Bowman 1960)) and now to incomplete and misleading plant lists for breeds on both Santa Cruz and Floreana. The some of the less accessible islands. Further Academy collectors may have overlooked this (1969b), in a survey of several island groups, 276 M. P. HARRIS

TABLE 3. Data used for multivariate analyses of numbers of bird species on islands and physical character- istics of the islands.

Distance No. of land bird species No. of species Distance to next to next high Breeding Ever Fl;l”a:tFg Total Altitude AIM (miles ) island now recorded FelXls plants (ft) ( sq. miles ) island (miles)

Isabela 20 21 59 244 303 5500 2249 2 2 Fernandina 18 20 7 90 97 4900 245 2 2 James 20 20 40 185 225 2974 203 3 11 Jervis 12 16 1 68 69 1203 2 3 3 Duncan 14 18 :: 30689 377100 1502 7 6 6 Santa Cruz 19 20 2835 389 6 6 Barrington 12 17 2 60 62 850 7 9 9 San Crist6bal 16 19 31 221 252 2350 195 25 36 Hood 9 12 1 96 97 650 18 24 24 Floreana 16 22 23 227 250 2100’ 64 26 26 Tower 8 9 0 40 40 210 4 25 25 Marchena 14 14 4 45 49 1125 45 17 17 Pinta 16 18 16 85 101 2550 20 17 17 Culpepper 7 8 0 10 10” 550 1 19 96 Wenman 7 12 1 21 22 830 2 19 75

Altitudes of islands and interisland distances me taken from the U.S. Government Chart No. 1375 (1946 corrected 1953), island areas from Preston (1962), and plants from Wiggins and Porter ( 1971). = Probably far too low (R. I. Bowman, pew. comm.). he thought that ecological impoverishment of plants), square root of the area, and square islands, including the number of plant species, of the distances between islands were sub- was correlated with island area and isolation. stituted for the linear values. The independent The recent publication of an up-to-date variables accounted for 94.9% of the insular flora of Galhpagos (Wiggins and Porter 1971) variation in the number of land-bird species. and the present re-evaluation of the status of By a step down procedure, the number of the land birds (table 2) enable us to take a variables was reduced until only those having new look at the problem of species on various a significant influence were left. These were islands. All the land birds, except rails which altitude (P < 0.05), log total plants (P < need rather specialized conditions, but includ- O.OOl), and square of distance to the next ing the hawk, have been taken into account. island (P < 0.1, one-tailed test). These ac- It seems likely that all these species are inter- counted for 90.5% of the variation. Leaving related as far as competition for food is con- out the distance factor only lowered the varia- cerned. Because of the difficulties of inter- tion to 87.7%. Plants on their own accounted preting the older records when referring to for 72.9% and altitude alone, 71.3%; these two stragglers or extinct populations, I have used factors are highly correlated (7 = 0.61, P < only the number of species thought to breed 0.01). It seemed unlikely that altitude per SC on each island at present. The information would have much influence on the number of used in the analysis is given in table 3. The bird species over the small range of altitude numbers of plant species are taken from an considered here, except secondarily through up-dated copy of the Wiggins and Porter flora plants. Nonetheless, removing the altitude kept at the Charles Darwin Research Station. component reduces the variation accounted As far as possible, plants introduced by man for from 87.7% to 72.9% so that altitude does are excluded. have some influence apart from that via the A multivariate analysis with a library pro- plants. Perhaps rainfall influences the birds gram and the ICL 1906A computer of the through its effect on and soil arthro- Computer Service, University of Oxford, was pods. In this analysis, Nesomimus trifasciatus made using the numbers of breeding land was taken as breeding on Floreana even birds as the dependent variable and other though it is now restricted to Champion and factors (columns 3-9) as the independent Gardner. A separate analysis omitting this variables. In some cases the number of bird species indicated that one less species had species correlated better with some transfor- little effect and only lowered slightly (to 87% ) mation of an independent variable than with the combined influence of log plants and the linear measurement. Hence, the log of altitude. plants (ferns, flowering plants, and total Another analysis using all land-bird species THE GALAPAGOS AVIFAUNA 277 ever recorded on each island as the dependent It is with great pleasure that I acknowledge variable again showed altitude and log plants my debt to these people. Perhaps the largest to be the most important factors, accounting contributions came from M. Castro and Tj. de for 81.7% of the variation compared with Vries but all the following supplied important 90.8% when all the independents were in- records which are acknowledged in the text cluded. Log plants alone accounted for 75.2% by their initials: B. Jacobs ( BJ), C. Anger- and altitude alone, 56.7%. The square root of meyer (CA), C. MacFarland (CM ), D. Boer- island area was also a significant factor (I ’ < sma (DB), D. J. Day (DD), E. Curio (EC), 0.05), but only accounted for 6.1%. Edinburgh University Expedition ( EU ) , G. Obviously, the total number of plants is Merlin (GM), His Royal Highness the Duke exerting the main influence on the number of of Edinburgh (HRH), H. van Derwerff ( HD ), bird species, whereas area, altitude alone, and J. Black (JB), J. Gordillo (JG), M. Castro isolation are much less important. The cor- (MC), P. Kramer (PK), R. I. Bowman (RB), relation being with the log of plants rather R. Lev&que (RL), R. S. R. Fitter (RF), R. than simply plants may be due to the fact Perry (RP), R. Tindle (RT), R. T. Peterson that the introduction of a few species to an (RTP), Tj. de Vries (TV), and T. De Roy impoverished island would significantly (TR). L. C. Binford and K. C. Parkes kindly change the environment, whereas one or two checked specimens in their care, and D. Ama- introductions to an island with many plants don and D. W. Snow allowed me access to would not be noticed. Also, given a humid material in the American Museum of Natural zone, governed mainly by altitude but also by History and British Museum of Natural His- exposure to the prevailing winds, many more tory, respectively. plant species would have to be introduced D. Lack, J. F. Monk, and A. W. Diamond before an increase in land-bird species would kindly criticized the manuscript. My field be predicted. On Cerro Azul ( Isabela), where work while at the Charles Darwin Research the humid zone occurs at a much lower alti- Station was sponsored by a research grant tude than usual, a semiarid area, including from the British Scientific Research Council cacti, occurs on the top of the , suggest- (1965-67) and by a fellowship given to the ing that even if the islands were higher there Charles Darwin Foundation by Lindblad would be no more diversification of the flora. Travel, Inc. of New York. This is a contribution of the Charles Darwin SUMMARY Foundation for the Galapagos. Details are given of the past and present status of all birds recorded in the Galapagos Archi- LITERATURE CITED pelago. Fifty-seven species are known to have AMERICAN ORNITHOLOGISTS ’ UNION. 1957. Check- bred, of these 28 are considered endemic. For list of North American birds. Fifth ed. A.U.O., Baltimore. the first time it is realized that either &tori&s BEEBE, W. 1924. Galapagos: worlds’ end. Putnam, hiatus breeds in Galapagos and hybridizes New York. with the endemic B. sundevalli to give inter- BOSD, J., AND R. M. DE SCHAUENSEE. 1938. The mediates, or that there is a single, extremely birds of Malpelo Island, Colombia. Proc. Acad. variable species. Nat. Sci., Philadelphia 90:’ 155-157. It seems likely that conditions have changed BOWMAN, R. I. 1960. Report on a biological recon- naissance of the Galapagos Islands during 1957. during the last 70 years and several island UNESCO, Paris. populations of land birds have become extinct. BOW~IAN, R. I. 1961. Morphological differentiation Although several birds have very small pop- and adaptation in the Galapagos Finches. Univ. ulations, the only species whose status gives Calif. Publ. Zool. no. 58. rise for concern is Pterodroma phaeopygia BROSSET, A. 1963. La reproduction des oiseaux de which has suffered from land clearance and mer des Ples Galapagos en 1962. Alauda 31:81- introduced mammals. 109. A multivariate analysis shows that the num- CASTRO, M., AND T. DE VRIES. 1970. Purple Gal- linule (Porphyrula martin&z) in eastern Pacific ber of plant species is the main factor influenc- Ocean. Ardea 58: 262-263. ing the numbers of land-bird species breeding CURIO, E., AND P. KRAMER. 1964. Vom Mangove- or occurring on any island, Isolation has little finken ( Cactospiza heliobates Snodgrass und effect. Heller). Z. Tierpsychol. 21:223-234. CURIO, E., AND P. KRAMER. 1965. Geospiza coni- ACKNOWLEDGMENTS rostris auf Abingdon und Wenman entdeckt. J. Omithol. 106:355357. Any work of this type depends to a large HAXLTON, T. H., AND I. RUBINOFF. 1963. Isola- extent on the keenness of other observers, both tion, endemism, and multiplication of species in in the field and in documenting their records. the Darwins’ Finches. Evolution 17:388-403. 278 M. P. HARRIS

HAMILTON, T. H., AND I. RUBINOFF. 1964. On LACK, D. 1969b. The numbers of bird species on models predicting abundance of species and en- islands. Bird Study 16: 193-209. demics for the Darwins’ Finches in the Galapagos L%QUE, R. 1963. Le statut actuel des vertebres archipelago. Evolution 18:339-342. rares et menaces de larchinel’ des Galanaaos._ - HAMILTON, T. H., AND I. RUBINOFF. 1967. On Tiere et Vie 110:39~7430. _ predicting insular variation in endemism and L~VPQUE, R. 1964. Notes sur la reproduction des sympatry for the Darwin Finches in the Galapagos oiseaux aux iles Galapagos. Alauda 32:544, 81- archipelago. Amer. Nat. 101:161-172. 96. HARRIS, M. P.- 1968. -eating in Galapagos mock- LI?V~QUE, R., R. I. BOWMAN, AD S. I. BILLEB. 1966. ingbirds. Condor 70:388-403. Migrants in the Galapagos area. Condor 68:81- HARRIS, M. P. 1969a. The biology of the storm 101. petrels in the Galapagos islands. Proc. Calif. LOOMIS, L. H. 1918. A review of the albatrosses, Acad. Sci. 37395-166. petrels and diving petrels. Proc. Calif. Acad. Sci. HARRIS, M. P. 196913. Food as a factor controlling 2: 1-187. breeding of Puffinus Zherminieri. Ibis 111:139- NELSON, B. 1968. Galapagos, islands of birds. 156. Lonnmans. London. HARRIS, M. P. 1969c. Factors influencing the breed- - I ing cycle of the Red-billed Tropicbird in the PHESTON, F. W. 1962.’ The canonical distribution Galapagos Islands. Ardea 57: 1491157. of commonness and raritv. Ecoloav 43: 185-215. HARRIS. M. P. 1969ld. Breeding seasons of sea-birds ROTHSCHILD, W., AND E. HARTERT. 1899. Review in the‘ Galapagos Islands. 1. Zool., Lond. 159: of the ornithology of the Galapagos Islands. 145-165. Novit. Zool. 6:85-205. HARRIS, M. P. 1970a. Breeding ecology of the ROTHSCHILD, W., AND E. HARTERT. 1902. Further Swallow-tailed Gull, Creugrus furcatus. Auk 87: notes on the fauna of the Galapagos Islands. 215-243. Novit. Zool. 9:373418. HARRIS, M. P. 1970b. The biology of an endan- SAMhrALrsro, L. 1967. Censusing Darwins’ Finches. gered species, the Dark-rumped Petrel (Ptero- Noticias de Galanagos 7/S: 13-17. drama phaeopygia), in the Galapagos Islands. SLUD, P. 1967. The birds of Cocos Island (Costa Condor 72: 76-84. Rica). Bull. Amer. Mus. Nat. Hist. 134:263-295. HARRIS, M. P., AND T. DE VRIES. 1968. White- SNOW, B. K., AND D. W. Slow. 1969. Observations faced Storm Petrels (Pelugodroma marina) near on the Lava Gull Lurus fuliginosus. Ibis 111:30- the Galapagos islands. Ardea 56:193. 35. HATCH, J. J., AND J. P. HAILMAN. 1967. Golden SWARTH, H. S. 1931. The avifauna of the Galapagos Plovers in Galapagos. Condor 69:320.’ Islands. Oct. Paper Calif. Acad. Sci. 18:1-299. HOSKING, E. 1972. Brown Booby Sula leucogaster, THORNTON, I. 1971. Darwins’ Islands. Amer. Mus. Galanaeos. Brit. Birds 65: nlate 76. _ v & Nat. Hist.. New York. JOHNSTON, R. F., AND R. K. SELANDER. 1964. House TRLMBLE, R. ’ 1943. Birds collected during the two Sparrows : rapid evolution of races in North cruises of the “Vanabondia” to the west coast of America. Science 144:548-550. South America. Ann. Carnegie Mus. 29:419441. LACK, D. 1945. The Galapagos Finches, (Geospiz- WIGGINS, I., AND D. M. PORTER. 1971. Flora of the inae). Oct. Papers Calif. Acad. Sci. 21:1-158. Galapagos Islands. Stanford Univ. Press, Stan- LACK, D. 1947. Darwins’ Finches. Cambridge Univ. ford. Press, Cambridge. LACK, D. 1969a. Subspecies and sympatry in Dar- wins’ Finches. Evolution 23:252-263. Accepted for publication 1 February 1973.