Rats were collected with a blowgun or snap traps baited with 

 coconut, the former proving far more effective because most 

 traps were sprung by hermit crabs. We preserved all specimens 

 in formalin and sent them to the US National Museum. 



Seabird Species Accounts 



Eleven species of seabirds occur at Caroline, most of 

 which breed in large numbers. They include one tropicbird, 

 three boobies, two frigatebirds. and five terns. 



Red-tailed Tropicbird (Phaethon rubricauda) (Figs. 2.3; PI. 1 ) 



Red-tailed tropicbirds breed at widely scattered locations 

 throughout the tropical Pacific and Indian Oceans. In the Line 

 Group, they nest from Palmyra south to Starbuck ( Perry, 1980), 

 with a large population (8,500 birds) on Christmas Island 

 (Clapp, 1967). Prior to our expedition it was unknown from 

 Caroline, Vostok, or Flint. 



Distribution and Habitat Preference : Our first indication 

 that red-tailed tropicbirds nested on Caroline was the discovery 

 of the skull, tail feather, and broken egg ( Subchapter 1 . 1 , PI. 46) 

 under a small Tournefortia bush on a previously unnamed islet 

 between Pig and North Brothers Islets. We named this sparsely 

 vegetated collection of rubble "Skull Islet" (Subchapter 1.1. 

 pis. 46,49). We later found 47 nests on another islet, naming 

 it Bo' sun Bird (Fig. 1 ) after the species' common name. 



All nests were located under relatively open Tournefortia 

 scrub less than 3 m tall in open, windy locations, with the 

 majority (91%) on small islets (0.24-0.86 ha). All nests were 

 under shrubs with few stems within a 0.5 m 2 nest space, and 

 most had peripheral cover on the sides of the shrubs, both 

 important factors in nest-site selection (Clark etal., 1983). All 

 nests were in areas relatively free of Polynesian rats (Rattus 

 exulans) and coconut crabs (Birgus Ultra): 5 nests on Long 

 were within 50 m of the island's south point. 



There are large populations of Polynesian rats and coconut 

 crabs on Caroline's bigger, more wooded islets. This rat, 

 though basically vegetarian, is an effective seabird predator 

 (Kepler, 1967; Norman, 1975) that in some years has taken 

 65% of the red-tailed tropicbird eggs and 1 00% of the chicks on 

 Kure Atoll (Fleet, 1972). Coconut crabs are also known bird 

 predators (Clapp & Sibley, 1971a; Helfman, 1979; Reese. 

 1987). It may be no accident that tropicbirds on Caroline 

 are restricted to small, relatively open islets that harbor few, if 

 any. rats and crabs or occur only on the tip of Long Island, 

 where predator densities are low. We saw no rats on Bo'sun 

 Bird Islet. Although rats could swim the 1 65 m to the islet, the 

 nearly continuous presence of black-tipped reef sharks 

 (Subchapter 1.1. PI. 10) in the channels surrounding the islet 

 provides protection to its nesting tropicbirds. 



Numbers : In September 1988. we found a total of 

 56 active nests on 5 islets (Fig. 2, Table 1) and estimated a 

 minimum population of 60 pairs. The May 1990 expedition 

 found 130 nests on Bo'sun Bird: our revised estimate for 

 Caroline is approximately 300 birds. Bo'sun Bird Islet was 

 surveyed by POBSP in June 1965, and no tropicbirds were 

 located on the ground or in the air (F. Sibley, personal 



communication ). It is unlikely that red-tailed tropicbirds were 

 present but overlooked at that time, suggesting that they have 

 colonized the atoll only recently. The Caroline population is 

 now the second largest colony known from the Line Group, and 

 Caroline is only one of five islands in the archipelago where 

 red-tailed tropicbirds are known to breed. 



Phenology : Of the 56 nests found in 1988, 54 contained 

 eggs or chicks (Table 2). The 33 chicks were divided into 4 age 

 classes (Fleet, 1974; Diamond, 1975a). which, together with 

 the 21 eggs, provided an indication of laying phenology for 

 140 days prior to our arrival (Fig. 3). Eggs in surviving nests 

 had been laid at a fairly even rate from early May (possibly 

 starting earlier) through September. The finding of only two 

 additional pairs on territory, and only one courtship flight, 

 indicated that laying was ending. On 24 May 1 990. many nests 

 contained eggs and downy chicks (75% nests with chicks), and 

 pairs were still courting. 



On Christmas Island, peak laying generally occurs from 

 June to October ( Schreiber & Ashmole. 1 970), later than those 

 parts of the 1988 and 1990 breeding seasons we observed on 

 Caroline. 



Masked Booby (Sula dactylatra) (Figs. 4,5; PI. 2) 



The masked booby is widely distributed in the Atlantic, 

 Indian, and Pacific Oceans. Clapp (1967) estimated that 

 19,100 masked boobies bred in the Line and Phoenix Islands, 

 with about 13,000 of them in the Line Islands, mostly 

 (ca. 9,000) on Jarvis. 



Distribution and Habitat Preference : Eighty-four percent 

 of masked booby nests (159) were on the windward, rubbly 

 shores of Long and Nake Islands, extending to the north end of 

 the atoll. Fifteen additional nests were scattered along the 

 lagoon edges of five South Nake Islets (Table 1 ). Nests 

 consisted of bare scrapes with exposed sand, usually within a 

 sparse ground cover of Portidaca and Heliotropium (PI. 2). 

 Over half the nests were amassed in one open colony on Nake 

 that extended nearly 1,000 m, beginning approximately 150 m 

 south of Tr. 2 and extending about 50 m north of Tr. 4 

 (Subchapter 1.1, Fig. 8). Here a nearly unbroken Heliotropium 

 mat 30-80 m wide, with patches of Tournefortia, occupied the 

 area between the leading edge of the Tournefortia scrub and the 

 beach crest. Nests were 20-30 m apart in the densest section 

 (nearTr. 3). All nests were exposed to the sun, unlike those of 

 the brown boobies. Some adults and juveniles roosted under 

 the scrub; guano deposits indicated regular occupancy. 



A loose group of 7 breeding pairs was scattered on a broad 

 plain of low herbs along a partially filled old interislet channel 

 370 m south of the north end of Long Island (Tr. C. 

 Subchapter 1.1, Figs. 8,40). Four more pairs nested in coral 

 rubble along the channel separating Nake and Long, one pair 

 with a downy young only 2-3 cm above high-tide flow on an 

 "islet" between fingers of the channel, a precarious location 

 where nesting surely must fail in stormy periods. No birds were 

 seen there in March and May 1 990, following a severe storm in 

 February 1990. Four pairs nested singly along a leeward 

 1 ,000-m stretch of lagoon shore on the northern end of Long 

 (Fig. 4); hardpan was the primary substrate. 



141 



