sign in sign up
<<
Home , Avarua, Cook Islands, Emoia, Lipinia noctua, Mangaia, Mauke (electorate)

Pacific Science (1986), vol. 40, nos . 1-4 © 1987 by the University of Press. All rights reserved

The of and , Cook Group, Oceania1

RONALD 1. CROMBIE 2 AND DAVID W. STEADMAN3

ABSTRACT: Eight of lizards are reported from the of Raro­ and Mangaia with comments on their distribution, ecology, reproduction, and variation. Particular attention is given to systematic problems in the Cryptoblepharus and pattern polymorphism in . Emoia trossula, recently described from , is reported for the first time in the . Historic and zoogeographic evidence suggests that most species oflizards arrived on Rarotonga and Mangaia in Polynesian voyaging canoes within the past 1000 years, although Gehyra mutilata and Hemidactylus garnotii have arrived by incidental boat or air transport in the past several decades.

UNDERSTANDING the distribution and zoo­ faunas of Mangaia and Rarotonga. Here we geography oforganisms in the Pacific is summarize the collections (181 speci­ hampered by the virtual absence of data on mens), which contain all but two ofthe species the biotas of many islands. The few sum­ previously known from the Cooks and include maries ofthe herpetofauna ofOceania (Stern­ two additional species previously unreported. feld 1920; Loveridge 1945; Brown 1956) are The two taxa not found in this survey appear generalized and incomplete. Published herpe­ to be restricted to the northern Cooks. The tological records for the Cook Islands, for nocturnal forest gecko pelagicus (for­ example, are very few. This scattered group of merly Cyrtodactylus pelagicus; see Kluge 1983 15 small islands lies between the Society and and Zug 1985for usage ofthis combination) is Tubuai groups to the east and , Tonga, widespread in Samoa, Fiji, and elsewhere in and to the west (Figure 1). Boulenger , but it has only been reported from (1887), Sternfeld (1920), Mertens (1931), Burt Island in the Cooks (Burt and Burt and Burt (1932), and Mcf'ann (1974) have 1932). The Emo ia adspersa has been provided some distributional information, found on (Burt and Burt 1932), mostly for the northern islands in the group, Samoa, and other scattered islands to the west but little published information exists for the (Schwaner and Brown 1984). lizards of Rarotonga and Mangaia, the two largest of the Cook Islands. One ofus (DWS) recently spentS weeks (13 March-23 April 1984, 28 May -13 1985) MATERIALS AND METHODS surveying the living and fossil vertebrate Data on sex, size, and habitat were taken for all specimens. DWS recorded snout-vent 1 Manuscript accepted March 1986. Fieldwork was length (SVL) in millimeters with a ruler and funded by the Smithsonian Institution (Fluid Research weight in grams with Pesola scales on freshly Grant through S. Dillon Ripley and Storrs L. Olson; funds killed specimens in the field. Sex and repro­ donated by Mrs. Alexander Wetmore to the Division of Birds; and the Smithsonian television series, ductive condition ofspecimens field-prepared through David Clark , Daria Sommers, and Rolfe as skeletons also were noted. On fluid­ Tessem). preserved material, sex, size, and selected scale 2Department of Vertebrate Zoology (Amphibians counts were taken in the lab by RIC. All cata­ and ), National Museum of Natural History , Smithsonian Institution, Washington, D.C. 20560. log numbers refer to the Amphibian and Rep­ 3 Biological Survey, New York State Museum, State tile collections of the U.S. National Museum Education Department, Albany , NY 12230. of Natural History (USNM). 44 Lizards of Cook Island Group-CROMBIE AND STEADMAN 45

~;

Rakahanga ,

Pukapuka c .~

. Nassau

. '. ....", . ~" '(o:. ~ : '" 0. ..'':'" :",

Palmerston ~.

Altutakt .« T'akute a . • COOK ISLANDS Ati'u . Rarotonga.. Mangaia •

I I I

FIGURE 1. Outline map of the Cook Island s with their relative position in Oceania (inset). Redrawn from Survey Dept. Rarotonga (1983).

DESCRIPTION OF THE ISLANDS population in 1981of9477 persons, an annual Rarotonga precipitation of 2040 mm, and mean maxi­ mum/minimum temperatures of 27.0 and Rarotonga (Figure 2) is the largest , highest, 20.8°C (Survey Dept. Rarotonga 1983). and most populous of the Cook Islands. Cen­ Entirely volcanic in origin and on its own tered at 21°14' Sand 159°47' W, Rarotonga bank, Rarotonga is a lush island whose steep, has a maximum elevation of653 m, and much knife-edged mountains are contrasted by a of the interior exceeds 300in elevation. Raro­ relatively flat, narrow coastal plain and a 2 tonga has an ofapproximately 65 km , a fringing surrounding the island. Four 46 PACIFIC SCIENCE, Volume 40,1986

Tereora College

Te Kakl Motu

Turangi Stream

Avana Stream

o 2 3 I o 2 RAROTONGA miles

FIGURE 2. Map of Rarotonga. Redrawn from Survey Dept. Rarotonga (1983). small islets lie off the southeastern shore, in­ (249660); Lepidodactylus lugubris (249661­ side the reef. (See Stoddart 1972 for a detailed 62, 249739-41 , 252378-85); Cryptoblepharus discussion.) The river valleys have been cul­ cf. poecilopleurus (252386-90); Emoia cyanura tivated for centuries with food crops such as (249667-73,249748-56,252392-400); Emoia , bananas, and sweet potatoes. Intro­ trossula (249663-66, 249742-47, 252391); duced trees, shrubs, and herb aceous plants noctua (249674-75, 249757-58, occur in even the most remote parts of the 252401-03). island. Many large, scattered patches of near­ Motutapu Islet: Emoia cyanura (249676­ ly pure bracken fen1 (Pteridium sp.) occur in 81, 249762). areas damaged by periodic fires. Stoddart and Islet : Emoia cyanura (249682- 86, Gibbs (1975) provided summaries of the his­ 249759-61 ); (249687). tory, climate, and "marine environments of Taakoka Islet: Emoia cyanura (249688-92). Rarotonga. A total of 97 specimens (USNM numbers in Man gaia parentheses) of the following species oflizards were collected on Rarotong a and three of its Mangaia (Figure 3) is the second largest, surrounding islets. second highest, and the third most populous Rarotonga: Gehyra oceanica (249654- 59, island in the Cook Group. Cente red at 249732- 38, 252377); Hemidactylus garnotii 21°55' Sand 157°57' W, it is the southernmost Lizards of Cook Island Group-CROMBIEAND STEADMAN 47

...... ;; \ ...

<11,. x ,;A. '. ~ -,,

1lfI' ...... -. : .,. fIf : ,! Lake Tepeuru Ngau ./ f / f , j , , "- / f I ". ./ t " ...... i f ; ~ ~ ...... •...... onero a,' f I ( ..,. . Village ( ~ Keia ·Swamp l '" \\. ", ..... "- "'" ...... ~ ~ ~ ~ ~ ~~.\.. " .... \1 ~ ~ .../ ......

kilometers o 2 3 I o 1 2 miles MANGAJA

FIGURE 3. Map of Mangaia. Redrawn from Survey Dept. Rarotonga (1983). ofthe Cook Islands, 200 km east-southeast of known as the "." The makatea, often Rarotonga. Mangaia has an approximate delimited by steep cliffs on both the seaward area of 51 krn", a maximum elevation of and especially the landward sides, reaches an 169m, an average annual precipitation of elevation of approximately 60m and is up to 1992mm, and a human population of 1364 1.2 km wide. The narrow coastal plain of in 1981 (Survey Dept. Rarotonga 1983). Mangaia has thin pockets of soil interspersed Mangaia is on its own bank, encircled by a with limestone outcroppings. Much of the fringing reef. The highly weathered, volcanic volcanic soil ofthe central uplands has eroded uplands in the center of Mangaia are sur­ away because of centuries of . rounded by a raised coralline limestone reef Except for near the three villages, the coastal 48 PACIFIC SCIENCE, Volume 40, 1986 plain and makatea are mainly forested, also on Inocarpus, in or at the edge of forest. although all is second growth with many in­ Four others were found in or on houses, an troduced plants. The uplands are largely de­ unusual habitat for G. oceanica, which nor­ void of trees, except in stream valleys, and mally prefers large trees with abundant crev­ are covered mainly with bracken fern or ices for diurnal retreats. Both houses were pineapples. Most agriculture, especially wet close to forested areas and not in a strictly taro in low-lying areas, occurs in volcanic soils suburban habitat. just inside the inner cliff of the makatea. Although males from both islands had en­ A total of 84 specimens of the following larged testes, none of the females was gravid; lizards were collected on Mangaia: Gehyra the largest ovarian follicles were 2-3 mm in mutilata (249693-94); Gehyra oceanica diameter. The stomachs ofseveral individuals (249695-97, 249763-69, 252404); Lepido­ from both islands were packed with small dactylus lugubris (249698-702, 249770-75, seeds. This was the only species of gecko ac­ 252405); Cryptoblepharus cf. poecilopleurus tive during the day, although this diurnal ac­ (249703-04, 249776); Emoia cyanura tivity was restricted to individuals in deep fur­ (249705-30,249777-96,252406-09); Lipinia rows of the trunks of Inocarpus. noctua (249731,249797-99,252410-11).

Hemidactylus garnotii (Dumeril and Bibron, 1836) SPECIES ACCOUNTS One juvenile female (46 mm SVL, 2.2 g) was FAMILY collected on a house in Nikao, a suburban area of Rarotonga. This is the first record Gehyra mutilata (Wiegmann, 1834) from the Cooks. No others were seen on Rarotonga or Mangaia. Two juveniles (22 and 21 mm SVL, respec­ This parthenogenetic species (Kluge and tively; each 0.2 g) were taken on Mangaia ina Eckardt 1969; Eckardt and Whimster 1971), suburban area. Although none was found on native to Southeast and the Indoaus­ Rarotonga, the species does occur there (G . tralian , has a restricted distri­ McCormack, pers . comm.). Its rarity in the bution in Oceania, suggesting the recency of southern Cooks suggests a recent arrival. Ge­ its arrival in the area. Since only a single indi­ hyra mutilata is widespread in Oceania (Crom­ vidual theoretically is needed to establish a bie) and occasionally is abundant in edificarian population, additional dispersal may be ex­ habitats. In the Cooks it has been previously pected. This species is well established in reported only from Manihiki (McCann Hawaii and has been recorded in the Gambier, 1974), Pukapuka (Slevin 1934), and an un­ Marquesas, Samoan, Society, and Tubuai is­ specified locality ("Danger Group," Burt and lands (Crombie; Gibbons, in litt. 1985). Burt 1932).

Gehyra oceanica (Lesson, 1830) Lepidodactylus lugubris (Dumeril and Bibron, 1836) Fourteen specimens (six males, 68-85 mm ; eight females, 67-81 mm SVL) of this large This small , gregarious gecko is widely dis­ gecko were collected along Turangi and tributed in Oceania, the Indoaustralian Archi­ Avana streams on Rarotonga. All were found pelago, Southeast Asia, and on or near the trunks of large "i'i" or "chest­ islands (Wermuth 1965: 99; Crombie). The re­ nut" trees (Inocarpus edulis) in both open and markable dispersal abilities of L. lugubris are closed forest. enhanced by its seemingly all-female, par­ On Mangaia , 7 of the 11 specimens (three thenogenetic nature (Cuellar and Kluge 1972; males, 72-78 mm; seven adult females, 65-79 Cuellar 1984). Although occasional males mm; one subadult female, 55 mm SVL) were have been recorded in some populations, all Lizards of Cook Island Group-CROMBIE AND STEADMAN 49

25 of our specimens from Rarotonga and mon) and Lipinia noctua (uncommon) were Mangaia were females or unsexable juveniles. also found at this locality . Four ofthese Cryp­ Of 13 specimens of L. lugubris from Raro­ toblepharus were adult females (46-48 mm tonga, all except one were collected on build­ SYL) with distended oviducts but only small ings. Ten adults (32-42 mm SYL) had very ovarian follicles. The fifth specimen was a sub­ small follicles; three others were juveniles adult female (37 mm). (21-27 mm SYL). A 41-mm female with small On Mangaia, three specimens were collected oviducal follicles was found on a forested at about 10m elevation. Two juveniles of 23 slope above Avana Stream, 1 km from human and 21 mm SYL, respectively, were found on a dwellings. This specimen was taken beneath cement water cistern. One 39-mm male with the rotting wood of a standing dead tree, ac­ small testes was from the base of a companied by two infertile and five already tree in open forest ofHernandia , Barringtonia, hatched eggs. Cocos, and Pandanus. Like the other species The 12 specimens from Mangaia were all of reported here, Cryptoblepharus was adult or subadult females. All specimens be­ strictly diurnal and difficult to find before tween 31 and 35 mm SYL had very small 0800 hr or after 1700 hr, preferring sunny follicles. A 37-mm individual had distended, exposures. convoluted oviducts. The three largest speci­ Mertens (1931, 1933, 1934, 1964) inter­ mens (41-45 mm SYL) contained shelled and preted the 40 or more taxa in the genus unshelled ova from 5 to 9 mm in length. Cryptoblepharus (which he erroneously called One of the Mangaian L. lugubris was on a Ablepharus) as subspecies of C. boutonii. papaya tree near Oneroa Village and another Several areas of sympatry in Indoaustralia was on a coconut tree in disturbed coastal involve two or more of these "subspecies" forest. These two specimens were darker than (Storr 1976; Covacevich and Ingram 1978; the ten individuals taken from man-made Auffenberg 1980), suggesting that they may structures. On both islands , specimens from be full species. We see no value or justification buildings are remarkably uniform, being pale in associating Pacific taxa with C. boutonii, dorsally with rows of darker paravertebral originally described from in the spots, vague sacral markings, and dark eye Indian Ocean. The taxonomic status and no­ stripes. These differences may be due to in­ menclature of the widespread Oceanic popu­ dividual color change, but more data are lations of Cryptoblepharus are poorly under­ needed. stood as well. The name most commonly ap­ Elsewhere in the Cooks, L. lugubris has plied to them is C. poecilopleurus Wiegmann, been recorded from Manihiki Atoll and Puka­ the type locality of which has been misinter­ puka (McCann 1974; Burt and Burt 1932). preted (Crombie and Dixon; Crombie in prep.). Elsewhere in the Cooks, Cryptoblepharus poecilopleurus has been reported from Mani­ F AMlLY SCINCIDAE hiki, Mui (Manuae), Mittiero (Mitiaro), Nas­ Cryptoblepharus cf. poecilopleurus (Wieg­ , Pukapuka, and Suwarrow (Mertens mann, 1834) 1931: 131 ; Sternfeld 1920:420; Burt and Burt 1932:514; McCann 1974). Five specimens were collected, and at least fiveothers were seen, at low elevation (about 3 Emoia trossula (Brown & Gibbons, 1986) m) in an open, grassy woodland ofHernandia , Casuarina, Barringtonia, and Coccoloba, This large, arboreal skink of the samoensis 0.3-0.7 km northwest of Avatiu Harbor, group was recently described from Fiji and Rarotonga. Each lizard was on the trunk ofa . Eleven specimens (four males, 73­ tree, usually less than 1 m from the ground. 84 mm; six females, 62-82 mm; one juvenile, Barringtonia was used less commonly than the 33 mm SYL) were collected along Turangi and other three trees. Both Emoia cyanura (com- Avana streams on Rarotonga; none was seen 50 PACIFIC SCIENCE, Volume 40,1986 on Mangaia. Except for a single individual these and other diurnal lizards. Acridotheres seen in the fronds of a young coconut tree occurs on Mangaia as well, although in lower (Cocos nucifera) near Avana Stream on 29 numbers and for fewer years. If predation by May 1985, all adults seen or collected were on Acridotheres has a significant impact on lizard Inocarpus or Hibiscus trees along forest edges populations, then we can expect to witness a near the streams. The single juvenile was decline in lizards on Mangaia over the next active among rocks along the banks ofTurangi decade or two as the number of Acridotheres Stream. These lizards often are found up to increases. 2.5 m above the ground and are very wary, Although the 14 adult male E. cyanura ascending to 5 m or more into the trees when (43-55 mm SVL) from Rarotonga all had disturbed. The males had enlarged testes; the moderate to very enlarged testes, none of the females contained only small follicles. six females (43-52 mm SVL) had mature fol­ No member of the E. samoensis group (as liclesor shelled ova . One female (52 mm SVL), defined by Brown and Gibbons 1986)has been collected on 31 May 1985, had enlarged, con­ reported previously from the Cooks, and the voluted oviducts and probably had laid her presence of E. trossula as as Rarotonga eggs recently . Four males and one female is surprising (see Zoogeographic Comments). (35-41 mm SVL) with small gonads were re­ garded as immature. The Rarotongan populations are more vari­ Emoia cyanura (Lesson, 1830) able in color and pattern than those on This small skink is widely distributed in Mangaia. The "typical" E. cyanura is boldly Oceania and the Papuan region .The consider­ striped with a bright blue tail, although many able inter- and intrapopulational variation in populations contain unicolored individuals morphology, color, and pattern in lizards of called "melanistic" by Oliver and Shaw the cyanura group has led to much systema­ (1953: 90) and the " bronze" morph herein. tic confusion. We conservatively refer our Although the bronze individuals are often dis­ material to E. cyanura, providing data on tinctive and have been recognized taxonomi­ color and pattern to aid future investigators. cally (Lygosoma cyanurum var. schauinslandi As elsewhere in Oceania, E. cyanura was the Werner 1901); all intermediates between the most common and ubiquitous lizard on striped and bronze morphs may occur in a Mangaia, Rarotonga, and even the smaller single population. Certain specimens with no islets. Most of the 25 specimens from Raro­ body striping have variable remnants of head tonga were found among rocks and leaf litter stripes . The tail on most striped, intermediate, (particularly coconut palm fronds) along and even full bronze lizards was blue, but streams in forest. Others were found on a large some bronze and striped individuals had root on a vertical road cut , on a horizontal bronze tails. All eight specimens from near limb ofa Hibiscus tree, in coastal strand vege­ Black Rock were striped with bronze tails. tation (mostly Convulvulus) with cobbles The bronze pattern does not seem to be corre­ near Black Rock, and in a grassy, rocky expo­ lated with sex. We do not know ifit is seasonal sure at 200 m, near Te Kaki Motu, the highest or permanent. Four males and two females elevation recorded for the species on Raro­ from Rarotonga were completely bronze; one tonga. Like the other lizards on Rarotonga, E. male and one female had intermediate pat­ cyanura seems to prefer lower elevations and terns. Tai l color apparently is not correlated may be absent from much of the montane with sex or maturity and may be constant forest. Nevertheless, E. cyanura usually is within a small, localized population. rarer in the coastallowlands of Rarotonga Of seven specimens collected in forest of than in similar areas on Mangaia. This distri­ Cocos, Casuarina, and Hibiscus on Motutapu bution may be related to the extreme abun­ Islet, five were juveniles or subadults (28­ dance on Rarotonga of the introduced Asian 36mm SVL) and two were adult females (45 mynah (Acridotheres tristis), an aggressive, in­ and 47mm SVL). One of the females had quisitive, omnivorous bird that may prey on enlarged, convoluted, and flaccid oviducts, in- Lizards of Cook Island Group-CROMBIE AND STEADMAN 51 dicating recent egg deposition, and another leaf litter beneath breadfruit trees (Artocar­ had only small follicles. Six of the seven had pus) in suburban areas; brush piles at edge of strongly striped dorsal patterns with blue field (Lake Tepeuru Ngau); tails; one juvenile was striped with a bronze coastal forest (mostly Pandanus); disturbed tail. coastal forest (Cocos, Barringtonia, and Her­ The eight specimens collected on Oneroa nandia); and Hibiscus, Cocos, and Manglifera Islet were from habitat similar to that on litter (Keia Swamp). Fallen fronds ofcoconut Motutapu. Six juveniles or subadults (26­ palms are frequently used for basking as well 41 mm SVL) were striped with blue tails. as shelter. Although E. cyanura is primarily One gravid female was striped with a gray tail terrestrial, several specimens confirm the (50 mm SVL) and another (50 mm SVL) was arboreality observed on Rarotonga. One juve­ bronze without a tail. Both contained two nile male (38 mm SVL) and three females large (13 x 6 mm) shelled eggs ready for de­ (42-46mm SVL) were collected up to 1m position, confirming Greer's (1968) suggested high on Inocarpus trunks in forest bordering a clutch size for the species. taro patch. The sample from Taakoka Islet contained The taxonomic confusion in the E. cyanura four juveniles or subadults (27-36mm SVL) complex precludes a complete summary ofits and one adult male (46 mm SVL) with slightly distribution. In the Cooks, E. cyanura has enlarged testes. Three of the smaller speci­ been recorded from Manuae, Mauke, Mitiaro, mens were striped; their tails were either blue, Nassau, Pukapuka, and Suwarrow (Crombie) . green, or missing. A subadult male was re­ Boulenger (1887 :291) previously reported E. corded as bronze with a blue tail in the field, cyanura from Rarotonga. The record of E. but in preservative the pattern is more inter­ cyanogaster from Rarotonga (Burt and Burt mediate and faded striping is evident on the 1932: 320) was based on misidentified E. head and body. The adult was greenish bronze cyanura (Brown 1956). with a blue tail. The sample from Mangaia (N = 50) was Lipinia noctua (Lesson, 1830) much more uniform than that from Raro­ tonga. All specimens were striped and, with On Rarotonga, one specimen of L. noctua the exception of a small series from a slightly was collected in a suburban area near Tereora higher elevation, all had blue tails. The four College, another in coastal strand vegetation specimens from Lake Tepeuru Ngau (30 m) mixed with coral cobbles near Black Rock , had bronze or greenish bronze tails. Repro­ and a third in open, grassy woodland of Her­ ductive activity occurred on Mangaia in April nandia, Casuarina , Barringtonia, and Cocco­ 1984, since all the adult males (40-54mm loba, 0.5-0.7 km northwest of Avatiu Harbor. SVL; N = 14) collected had enlarged testes Four females (31-46 mm SVL) were in open and most stages in the female reproductive forest near Turangi or Avana streams. Along cycle were represented. Follicles in adult fe­ Turangi Stream, one was active among rocks males (42-52mm; N = 19) ranged from tiny and leaf litter, whereas another was sunning (1-2 mm) to large and yolky; two specimens on a large steel pipe. Near Avana Stream, one (46 and 47 mm SVL) contained shelled eggs was under moist, rounded boulders and the ready for deposition and several others had other was on an Inocarpus tree. Only one the distended, convoluted oviducts offemales (43 mm SVL) of these, taken in March 1984, that had recently laid eggs. Thirteen juveniles was reproductively active, with a 6 x 5 mm and subadults ranged from 26 to 38 mm SVL. ovum in the left oviduct. In contrast, a female The reproductive status in May -June 1985 (41 mm SVL) collected in May 1985contained was unknown because the sample was small a fully formed, ready to be born fetus (about (two adult males of 46 and 44 mm SVL and 14mm SVL) in the left oviduct and several two juveniles of 28 and 31 mm SVL). small ova in the right oviduct. Another female Ecological associations on Mangaia from collected in May (46mm SVL) contained two which E. cyanura was taken include: rocky large (8-10 mm) yolky ova. 52 PACIFIC SCIENCE, Volume 40,1986

One specimen (40 mm male with enlarged Tubuai, Marquesas, Hawaii, and New Zea­ testes) was collected on Oneroa Island in land. Contact among the southern Cook forest ofCocos, Casuarina, and Hibiscus. Islands has occurred since their initial settle­ Of the six Mangaian specimens, one was in ment. rocky leaflitter beneath a breadfruit (Artocar­ The first recorded European landing on pus) tree in suburban Oneroa village, three Rarotonga was in 1814(Crocombe 1964). The others were in leaf litter of second-growth first significant European influence came with forest near the base of the inner cliff of the British and Tahitian in 1823 makatea, and two others were in disturbed (Carter 1984). Since that time, Europeans coastal forest (mostly Cocos and Barring­ have had a greater impact on Rarotongathan tonia) about 0.8 km southwest of the end of on any ofthe other Cook Islands, culminating the airstrip; The males (40-42 mm SVL; N = in the construction ofan international airport 3) from the March-April 1984 collection had in the 1970s. enlarged testes, and the single female (45mm Significant European influence on Man­ SVL) had one enlarged (6 x 5 mm) follicle. gaia began with the arrival of British and The adult females (46-47 mm SVL) collected Tahitian missionaries in 1823 (Gill 1894: in June 1985 were reproductively active, the 243-250, 323-325). Since that time, irregular smaller one with two large, yolky follicles and boat traffic has traveled to and from Mangaia, the other with one full-term fetus in the right an island with less European influence than oviduct. Rarotonga. A small airstrip was completed Sternfeld (1920: 402) reported L. noctua on Mangaia in the late 1970s, and there are from Rarotonga. It is also known from flights to Rarotonga and sometimes elsewhere Mauke, Mitiaro, Nassau, Pukapuka, and in the southern Cooks. Suwarrow in the Cooks (Crombie). Natural dispersal (rafting) of lizards on driftwood or palm trash is a likely means of colonization for islands near the Australasian landmass. With lowered levels during the ZOOGEOGRAPHIC COMMENTS Pleistocene, the area of many island groups in Possible explanations for the distribution of the western Pacific was much greater than lizards in Oceania include natural dispersal now (Gibbons 1985), thus facilitating over­ and human transport. If endemic genera or sea dispersal. Geckos in particular are well species oflizards were present in the Cooks, as adapted for oversea dispersal. Kluge (1969:46) in western Oceania, this would suggest a has listed 18 traits of gekkonids that render longer period ofisolation than would be pos­ them successful "rafters." To this list we may sible if human transport had been the disper­ add parthenogenesis (as in Hemidactylus gar­ sal agent. With ,the possible exception of notii and Lepidodactylus lugubris), whereby a Emoia trossula, we believe that most or all single individual can establish a population. species arrived on Rarotonga and Mangaia Despite this considerable potential for with people . Thus a briefreview of the human oversea dispersal, the immense distances be­ history of the islands is appropriate. tween islands in eastern and central Oceania The southern Cook Islands, including make the possibility ofa successful landfall by Mangaia and Rarotonga, were first colonized a rafting lizard very unlikely . Consequently, by , perhaps from the Society natural dispersal has been largely discounted Islands, about 1000 years ago (Bellwood to explain the presence oflizards on the more 1979: 348), although Kirch (1986: 33) has sug­ distant and isolated islands. The virtual ab­ gested that the initial colonization may have sence of endemic reptiles east of Samoa is been more than 1000 years ago and may have further, albeit indirect, evidence that colorii­ originated from anyone of several groups of zation of these islands has been relatively re­ Polynesian islands . The culture (including cent. The homogeneity ofthe lizard faunas on language) of the southern Cook Islands is islands east ofSamoa (usually fiveor six out of similar to that of the "eastern" Polynesian eight possible species on any island) suggests a island groups of the Society, Tuamotu, common and multidirectional mode of trans- Lizards of Cook Island Group-CROMBIE AND STEADMAN 53

port, such as the wide-ranging Polynesian edificarian habitats, suggesting a recent ar­ voyaging canoes. These vessels would have rival in the southern Cooks. Gehyra mutilata, offered numerous hiding places for eggs or collected only on Mangaia but reported to adult lizards in the thatching, , and occur on Rarotonga, is widespread in wooden objects carried on board; fruit , dried Oceania, including the northern Cooks, so its fish, and associated insects would have pro­ presence in the southern islands is expected. vided adequate food. Zweifel (1979: 18-19) The case of Hemidactylus garnotii is more has suggested this method of dispersal for problematical. This lizard has a spotty distri­ both Lipinia noctua and Cryptoblepharus. He bution in , so its discovery on Raro­ provides convincing evidence that the relative tonga was surprising. Details on the dispersal morphological homogeneity of L. noctua in ofthis species in Oceania are unknown, but we Oceania, compared to the presumed source assume that it was a recent event. Since H . area (), argues for a single intro­ garnotii is parthenogenetic and ecologically duction and subsequent wide dispersal within versatile, its potential for establishing new Oceania. Although comparable morpholog­ colonies is theoretically far greater than ical and genetic data are lacking for most bisexual species, and we would expect a wider widespread Oceanic species, Cuellar (1984) distribution in Oceania if it had been present has demonstrated high histocompatibility for a long time. Lepidodactylus lugubris , for within and even between distant Oceanic example , is also parthenogenetic and has been populations (Hawaii and the ) recorded from 22 island groups in Oceania, of Lepidodactylus lugubris . Indeed, he has compared to 7 for H. garnotii. The range ofH. gone so far as to suggest that "the entire garnotii may still be expanding, for although species, now distributed throughout the In­ the species was described in 1836 from dian and Pacific Oceans, originated from a and was recorded from Hawaii by the mid­ single ancestral female" (Cuellar 1984: 183). 1800s, documentation ofits presence on other Incidental transport of lizards between island groups has been much more recent. For islands undoubtedly continued (and may have example, earlier reports (Burt and Burt 1932; increased) with the wooden sailing ships of Higgins 1943) on the herpetofauna ofSavai'i, early European explorers and the resulting western Samoa, did not include H. garnotii. expansion of boat traffic between islands . John Gibbons (pers. comm. April 1985) in­ Although modern, steel-hulled vessels offer formed us that the species occurs on the island fewer refuges for stowed-away lizards , suffi­ today. Since H . garnotii is frequently found in cient shipments of wood , fruit , and thatch edificarian habitats, where collectors might pass between islands to ensure that this mode easily find it, the species probably arrived on oftransport continues today. Savai'i within the past 40 years or, as Gibbons Six of the eight species of lizards collected believes, within the past decade. Biochemical on Rarotonga and Mangaia are common and or histocompatibility studies on Oceanic H. widespread on distant islands in Oceania. Five garnotii populations may be informative, par­ of these (G. oceanica, L. lugubris, C. poecilo­ ticularly since Darevsky et al. (1984) have pleurus , E. cyanura, L. noctua) were found on demonstrated that some Southeast Asian both Rarotonga and Mangaia and probably populations previously referred to garnotii arrived with the Polynesians. The offshore are a distinct species. We assume that both G. islets of Rarotonga, inside the reef, are in­ mutilata and H . garnotii arrived in the south­ habited by a small subset of the Rarotongan ern Cooks in post-Polynesian times, perhaps fauna , with only E. cyanura on two islets and within the past few decades. E. cyanura and L. noctua on one. Additional The final species, Emoia trossula from species (particularly geckos) may be present Rarotonga, is a large, conspicuous, forest­ on the islets since no nocturnal collecting was dwelling skink that does not fit the pattern of done and comparatively little time was spent most Polynesian-transported species. Its pre­ surveying the fauna . sence on Rarotonga (and not Mangaia), in the The other two geckos (G. mutilata and H. absence of Fijian /Tongan species more likely garnotii) were found in low densities in to have been transported (such as Cryptoble- 54 PACIFIC SCIENCE, Volume 40,1986 pharus eximius and ), is dif­ garnotii) while reducing space for others ficult to explain and may indicate natural dis­ (Emoia trossula, G. oceanica). Modern Raro­ persal. This unexpected faunal relationship tongans and Mangaians are rather oblivious between the southern Cooks and Fiji/Tonga to lizards, except to recognize that geckos eat suggests that the Tonga Trench does not rep­ insects around their houses. There is only one resent a completely effective barrier to the word ("moko") on Rarotonga in general use dispersal of non volant terrestrial organisms. for all lizards, and the Rarotongans inter­ Niue, nearly midway between Rarotonga and viewed by DWS recognize only two or three of Tonga, would be a logical place to find a the seven species. The Mangaians recognize species common to Tonga and the southern only two types of lizards : the skinks ("motu­ Cooks, yet no Emoia of the samoensis group kutuku") and geckos ("moko karara"). Liz­ has been reported from Niue. The fauna ofthe ards do not seem to be hunted or otherwise island consists offour species of skinks, three intentionally molested on either island. Both geckos, and a sea snake (Gunther 1874; the interisland distribution and the intraisland Crombie). One skink (Emoia lawesii) has abundance of Cook Island lizards seem to be Samoan affinities, but the other lizard species unintentional results of human activities. (c. poecilopleurus, E. cyanura, G. oceanica, L. lugubris, L. noctua, N. pelagicus) are all wide­ spread in Oceania. Gunther's record of the SUMMARY Papuan Mabuya (= Emoia) baudinii was based on misidentified Emoia cyanura (BMNH Eight species of lizards (181 specimens) 71.4.16.43). Many other characteristic Fijian/ from the islands of Rarotonga and Mangaia Tongan herpetofaunal elements (ranid frogs, provide the most complete herpetological sur­ iguanid lizards, boid and elapid snakes) have vey of this poorly known area. Seven species also not crossed the Tonga Trench to Niue or (three geckos, four skinks) were found on the southern Cooks. We cannot rule out the Rarotonga and six species (three geckos, three possibility that E. trossula recently colonized skinks) on Mangaia. Five species (Gehyra Rarotonga with human assistance, but more oceanica, Lepidodactylus lugubris, Crytoble­ detailed comparative studies should clarify pharus cf. poecilopleurus, Emoia cyanura, this problem. Lipinia noctua) were present on both islands; Although most or all of the lizards discussed Gehyra mutilata may also be on Rarotonga, in this paper owe their modern Oceanic distri­ although we have specimens only from Man­ bution in large part to human activities, the gaia. Hemidactylus garnotii and Emoia trossula dates of their arrival in the southern Cooks were found only on Rarotonga. Three of the are speculative. The only lizard whose bones reef islands off Rarotonga were briefly sur­ have been found in the caves of Mangaia is veyed, and one or two species (E. cyanura and Gehyra oceanica, although these limited re­ L. noctua) were found on each island . mains were not from a clear chronological All the skinks on Rarotonga and Mangaia context (Steadman 1985). Further paleonto­ are diurnal and moderately to exceedingly logical and zooarchaeological work in Oceania abundant. Although E. cyanura may be the would provide information on the chronology most common and widespread lizard on and mechanisms of dispersal for lizards and Rarotonga, it appears to be absent from much other terrestrial organisms. ofthe montane forest. Its reduced abundance The human influence on other vertebrates in the disturbed coastal lowlands of Raro­ (birds) on Rarotonga and Mangaia has been tonga may be due to predation by the intro­ one of predation and extinction rather than duced mynah bird (Acridotheres tristis). dispersal (Steadman 1985). The human im­ Lipinia noctua is ecologically similar to E. pact on intraisland distribution and abun­ cyanura, and the two species are often found dance ofPacific lizards may be considerable as sympatrically. Cryptoblepharus cf. poecilo­ well. Clearing of forest would increase avail­ pleurus prefers open, disturbed woodlands very able habitat for some species (L. lugubris, H. near the coast on both islands, where both E. Lizards of Cook Island Group-CROMBIE AND STEADMAN 55

cyanura and L. noctua are found as well. may apply to a distinct South American Emoia trossula, the most ecologically re­ species, although the latter possibility seems stricted skink, is found primarily on trees unlikely. In light of morphological differ­ (Cocos, Hibiscus, Inocarpus) along streams in ences, we prefer to recognize the Oceanic lowland forest on Rarotonga only. Of the Cryptoblepharus as specifically distinct from geckos, Hemidactylus garnotii and Gehyra C. boutonii. The analysis of variation within mutilata were found only on buildings. Le­ Oceania is hampered by limited samples from pidodactylus lugubris was most abundant (or critical areas. at least most obvious) on man-made structures The systematics of the "azure-tailed" and rarely was found on trees. In contrast, skinks of the Emoia cyanura complex is com­ G. oceanica was most common on large trees plicated by pattern polymorphism and varia­ (Inocarpus edulis) in lowland forest, although tion in tail color. Boldly striped, partially a few specimens were collected on or in build­ striped, and unicoi or ("bronze" or "melan­ ings on Mangaia. Several individuals of G. istic") individuals can be found within a oceanica from both islands had consumed single, localized population. On Rarotonga small, unidentified seeds, as well as insects. and Mangaia, this variation is not correlated Most species were reproductively active clearly with sex or maturity. during the survey periods (March-April 1984 The presence of lizards on Rarotonga and and May-June 1985) . Both specimens of G. Mangaia is attributed primarily to inad­ mutilata were recently hatched juveniles, and vertent human transport. Five species (G. the single H. garnotii was an immature female. oceanica, L. lugubris, C. cf. poecilopleurus, E. No gravid females of G. oceanica were found, cyanura, L. noctua) probably arrived with the although the males all had enlarged testes. All Polynesians during the past 1000 years. Two 25 specimens of L. lugubris were females or others (G. mutilata and H. garnotii) may have unsexable juveniles, indicating an all-female, arrived more recently on post-European boat parthenogenetic population. No gravid fe­ or air traffic. The presence on Rarotonga ofE. males of Lrlugubris were found on Rarotonga, trossula, a Fijian /Tongan element, may be due although 'several recently hatched juveniles to natural dispersal. The lizard faunas of were collected and one adult was associated Rarotonga and Mangaia are very similar with a communal nest of seven infertile or except that the presence of two additional hatched eggs. On Mangaia, certain females species and the greater intraisland variation were either gravid , had enlarged follicles, or (especially in E. cyanura) on Rarotonga sug­ had distended oviducts. In the skinks, males gest multiple invasions oflizard stocks on that of E. trossula had enlarged testes whereas the island. This idea is supported by the human females showed no signs of reproductive ac­ history of the islands , which indicates that tivity. The one juvenile was recently hatched. Mangaia has been more isolated (especially in Juveniles of C. cf. poecilopleurus were present post-European times) than Rarotonga. on Mangaia, and females from Rarotonga Much remains unknown about patterns of had distended oviducts with small follicles. distribution and dispersal of Oceanic lizards. Certain females of the viviparous L. noctua We suggest that surveys of extant and extinct had enlarged follicles, while others contained faunas would contribute greatly to our under­ a single full-time fetus, and the males were standing of whence, when, and how lizards producing sperm. Gravid females and/or came to occupy these isolated islands. sperm-producing males of E. cyanura were found on both islands. Clutch size was invari­ ably two. ACKNOWLEDGMENTS Our application of the name C. cf. poecilo­ pleurus to widespread Oceanic populations is Research permits and many other cour­ problematical because of the confused no­ tesies were obtained through the kind efforts menclature of this species. Its type locality (in of Stuart Kingan (Prime Minister's Depart­ Andean Peru) may be in error, or the name ment) and Tony Utanga (Department of In- 56 PACIFIC SCIENCE, Volume 40,1986

ternal Affairs) . On Rarotonga, Sharon Fal­ Species of the group of cone , Douglas Henderson, Teariki , lizards (Scincidae) in the Fiji Islands, with and Andrew Toputopu assisted in various descriptions oftwo new species. Proc. Calif. ways. Discussions with Gerald McCormack Acad. Sci. 44(4):41 -53. were very enlightening. On Mangaia, assis­ BURT, C. E., and M. D . BURT. 1932. Herpeto­ tance and friendship were provided by many logical results of the Whitney South Sea persons, but especially Tua John (chief ad­ Expedition. VI. Pacific Island amphibians ministrative officer), Tua Uria (head of the and reptiles in the collection of the Amer­ Island Council), Atingakau Tangatakino ican Museum ofNatural History. Bull. Am. (senior advisory officer, Agriculture Depart­ Mus . Nat. Hist. 63(5):461-597. ment), Tiria Ngatokorua, Peter Ngatokorua, CARTER , J., ed. 1984. Pacific Islands yearbook. Tiriamate Ngatokorua, George Tuara, 15th ed. Pacific Publications, Sydney. Ngametua Kareroa, Paiti Kareroa, Maata COVACEVICH, J., and G. J. INGRAM. 1978. An Ora, Ngara Ora, Metu Ruatoe, Ngatamaine undescribed species of rock dwelling Cryp­ Ruatoe, and Poko Atariki. toblepharus. Mem. Queensland Mus. For helpful criticisms ofthe manuscript and 18(2): 151-154. access to unpublished information we thank CROCOMBE, M. 1964. They came for sandal­ the following individuals: John Gibbons wood. Whitcombe & Tombs, . (University of the South Pacific, , Fiji), CROMBIE, R. I. In prep. The herpetofauna of George R. Zug and W. Ronald Heyer (USNM), Oceania: Island lists and species distribu­ Robert P. Reynolds and Roy W. McDiarmid tions . Manuscript. (U.S . Fish and Wildlife Service, USNM). Molly CROMBIE, R. I., and J. R. DIXON. In prep. The Dwyer Griffin drew the maps on very short status of Mabuya deserticola (Sauria: Scin­ notice. She and Linda K. Gordon, who typed cidae) with comments on the influence of the final draft, consistently weathered our Oceania on the neotropical herpetofauna. unreasonable demands with good humor. Manuscript. ' Examination of critical specimens in Eu­ CUELLAR, O. 1984. Histocompatibility in ropean museums was facilitated by a Smith­ Hawaiian and Polynesian populations of sonian Research Opportunities Fund award the parthenogenetic gecko Lepidodactylus (1233F624) to RIC. This paperis contribution lugubris. Evolution 38(1): 176-185. number 484 of the New York State Science CUELLAR, 0., and A. G. KLUGE. 1972. Service. Natural parthenogenesis in the gekkonid lizard Lepidodactylus lugubris. J. Genet. 61(1): 14-26. DAREVSKY, I. S., L. A. KUPRIYANOVA, and LITERATURE CITED V. V. ROSHCHIN. 1984. A new all-female AUFFENBERG, W. 1980. The herpetofauna of triploid species of gecko and karyological Komodo, with notes on adjacent areas. data on the bisexual Hemidactylusfrenatus Bull. Florida State Mus. 25(2):39-156. from Vietnam. J. Herp. 18(3):277-284. BELLWOOD, P. 1979. Man's conquest of the ECKARDT, M . J., and I. W. WHIMSTER. 1971. Pacific. Oxford University Press , New Skin homografts in the all-female gekkonid York. lizard Hemidactylus garnotii. Copeia (I): BOULENGER, G. A. 1887. Catalogue of the 152-154. I lizards in the British Museum (Natural GIBBONS, J.R.H. 1985. The biogeography and I ~-" History). 2nd ed. Vol. 3. Taylor & Francis, evolution ofPacific Island reptiles and am­ London. phibians. Pp. 125-142' in G. C. Grigg, BROWN, W. C. 1956. The distribution of ter­ R. Shine, and H . Ehmann, eds., Biology of restrial reptiles in the islands of the; Pacific Australasian frogs and reptiles. ' Royal Basin. Proc. 5th Pac . Sci. Conf. 3A::1479- Society of New South , Sydney. 1491. " GILL, W. W. 1894. From darkness to light in BROWN, W.

ligious Tract Society, London. (1984 reprint Taxonomic status and distribution of the by University of South Pacific, Institute of scincid lizards Steindach­ Pacific Studies, Suva, Fiji.) ner and E. lawesii Gunther on islands of GREER, A. E. 1968. Clutch size in the genus the southwestern Pacific. Herpetologica Emoia. Copeia (2) :417-418. 40(2): 158-164. GUNTHER, A. 1874. A contribution to the SLEVIN, J. R. 1934. The Templeton Crocker fauna of Savage Island. Proc. Zool. Soc. Expedition to western Polynesian and Me­ London (20):295-297. lanesian Islands, 1933. No. 15. Notes on the HIGGINS, H. 1943.A few reptiles from western reptiles and amphibians with the descrip­ Samoa. Copeia (1): 59. tion of a new species of sea snake. Proc. KIRCH, P. V. 1986. Rethinking East Polynesian Calif. Acad . Sci. 21(15): 183-188. prehistory. J. Polynesian Soc. 95(1): 9-40. STEADMAN, D. W. 1985. Fossil birds from KLUGE, A. G. 1969. The evolution and geo­ Mangaia, southern Cook Islands. Bull. Br. graphical origin of the Hemi­ Ornith. Club 105(2): 58-66. dactylus mabouia-brooki complex (Gekko­ STERNFELD, R. 1920. Zur Tiergeographie nidae, Sauria). Misc. Pub . Mus. Zool. , Papuasiens und der pacifischen Inselwelt. Univ . Michigan (138) : 1-78. Abh. Senck. Naturf. Ges. 36: 375-436. --- . 1983. Cladistic relationships among STODDART, D. R. 1972. Reef islands of Raro­ gekkonid lizards . Copeia (2): 465-475. tonga, with list of vascular flora by F. R. KLUGE, A. G., and M. J. ECKARDT. 1969. Fosberg. Atoll Res. Bull. 160: 1-14. Hemidactylus garnotii Dumeril and Bibron, STODDART, D. R., and P. E. GIBBS, eds. 1975. a triploid, all-female species of gekkonid Almost-atoll of : Reef studies in lizard. Copeia (4): 651-664. the Cook Islands, South Pacific. Atoll Res. LOVERIDGE, A. 1945. Reptiles of the Pacific Bull. 190: 1-158. world. Macmillan, New York. STORR, G. M. 1976. The genus Cryptoble­ MCCANN, C. 1974. Reptiles from Manihiki pharus (Lacertilia, Scincidae) in Western Atoll. In J. S. Bullivant and C. McCann, . Rec. West. Australian Mus. Contributions to the Natural History of 4(1): 53-63. Manihiki Atoll, Cook Islands. N.Z. Ocean. SURVEY DEPT. RAROTONGA. 1983. Maps ofthe Inst. Mem. (31): 61-63. Cook Islands. , Rarotonga. 19 pp. MERTENS, R. 1931. Ablepharus boutonii (Des­ WERMUTH, H. 1965. Liste der rezenten Am­ jardin) und seine geographische Variation. phibien und Reptilien. Gekkonidae, Pygo­ Zool. Jahrb. (Syst.) 61(1/2): 1-162. pod idae, Xantusiidae. Das Tierreich. Lief. - --. 1933. Weitere Mitteilungen iiber die 80. xxii + 245 pp. Rassen von Ablepharus boutonii (Desjar­ WERNER, F. 1901 . Ergebnisse einer Reise nach din), I. Zool. Anz. 105(3/4):92-96. dem Pacific (Schauinsland 1896-1897). - -- . 1934. Weitere Mitteilungen iiber die Reptilien. Zool. Jahrb. 14(5): 380-387. Rassen von Ablepharus boutonii (Desjar­ ZUG, G . R. 1985. Pacific Island lizards ­ din) , II. Zool. Anz. 108(1/2) :40-43. status of type specimens from the U.S. Ex­ - - - . 1964. Weitere Mitteilungen iiber die ploring Expedition 1838-1842. Copeia Rassen von Ablepharus boutonii (Desjar­ (1): 150-154. din) , III. Zool. Anz. 173(2): 99-110. ZWEIFEL, R. G. 1979. Variation in the scincid OLIVER, J. A., and C. E. SHAW. 1953. The lizard Lipinia noctua and notes on other amphibians and reptiles of the Hawaiian Lipinia from the New Guinea region. Am . Islands. Zoologica 38(2): 65'---95. Mus. Novit. 2676 : 1-21. SCHWANER, T. D., and W. C. BROWN. 1984.