Red-foots were absent from South Island, which was 

 primarily covered with Cocos (Subchapter 1.1, Figs. 50,51). 

 Even though Tournefortia occurred on all its coastlines, no 

 boobies nested in them. Ana-Ana was also unoccupied: the 

 presence of a family of four people, a cat (removed in 1990), 

 and a dog undoubtedly discouraged nesting attempts. Red- 

 foots also avoided the mixed forests of south Nake, which 

 contained much Cocos and Pandanus (Subchapter 1.1, 

 Fig. 37 ). Red-footed boobies were thus found only in Caroline' s 

 indigenous woodlands, primarily in Tournefortia >2 m tall; 

 they avoided anthropogenic plant communities and man. 



Red-foots used a wider range of habitats for roosting. 

 Nonbreeding birds were found throughout the taller indigenous 

 trees, even in leeward situations where Pisonia and Cordia 

 overhung the lagoon (as on Long Island). 



Numbers : The POBSP(Clapp& Sibley, 197 la) estimated 

 5,000 ± 25% Red-foots on Caroline in June 1965, with about 

 2,000 + 25% nesting pairs. In 1988, we sampled systematically 

 more than 7% of the available habitat on all motus except 

 Crescent (4.6% sampled) and North Arundel, and estimated 

 that 2,22 1 pairs of red-footed boobies nested on 27 of Caroline' s 

 islets (Table 1). We found an additional 1,234 roosting, 

 nonbreeding birds. We know (Kepler, 1969; Nelson. 1978) 

 that fewer boobies remain in their colonies during the day than 

 at night. Thus, an unknown fraction of the population was at 

 sea when we conducted our counts. Impressive flights of red- 

 footed boobies returned each evening: 3^4 birds arrived for 

 each one that had remained behind, many undoubtedly mates 

 of incubating birds. To approximate the number of returning 

 nonbreeding birds, we doubled the number of roosting adults 

 to allow for an additional 1,234 adults and juveniles. Thus, our 

 conservative estimate was at least 7,000 individuals. 



Because red-footed boobies were so dependent upon 

 Tournefortia, we determined the nesting population on each 

 islet by multiplying the number of nests found on transects by 

 the ratio of sampled to total Tournefortia area. Perimeter 

 counts (Subchapter 1.1, Fig. 9) were used if the number of red- 

 foots observed exceeded the number calculated from the cross- 

 island transects. 



Long Island held the greatest number of nests (659), 

 mostly in the leeward Tournefortia and Tournefortia— Pisonia 

 edge. Bird densities were typically highest on the largest islets: 

 Windward and Tridacna, the largest Windward Islets, held 163 

 and 1 1 1 nests, respectively; and Mannikiba. the biggest leeward 

 islet, harbored the largest population ( 1 84) of the entire leeward 

 side. There were exceptions, however: Pandanus, with four 

 times the area of Tournefortia of any of the South Nake Islets, 

 held fewer birds than three much smaller islets (Table 1 ). 



Tournefortia scrub and forest covered approximately 

 125.25 ha (Subchapter l.l,Table9). Overall, there were 1.75 

 red-footed booby nests/ 1 .000 nr of Tournefortia forest. Nest 

 densities for occupied islets by island groups (Table 4) showed 

 that red-foots favored areas less exposed to the trade winds: 

 most nests on the windward motus were protected by well- 

 developed Pisonia forests. The exposed Central Leeward 

 Islets held the lowest nest densities ( 1 . 2 nests/ 1 .000 m 2 ), far less 

 than on the South Nake Islets (5.3/1,000 nr), which are 



protected by the northern edge of Long. The greatest densities 

 (7.8 nests/ 1 ,000 nr) occurred on the South Nake Islets south of 

 Pandanus. 



Broadly speaking, red-foots breed in well-dispersed 

 colonies. A record density of 600 nests/ 1 ,000 m 2 on Tromelin 

 Island (Indian Ocean) is exceptional. Elsewhere, 

 53 pairs/1,000 nr on Tower Island (Galapagos), 40/1,000 nr 

 on Moku Manu (Oahu. Hawaii), and 27/1,000 m 2 on Half 

 Moon Cay (Honduras ) are more consistent high-density colonies 

 (Nelson, 1978). Only on tiny Motu Kota (Subchapter 1.1, 

 Fig. 52), with 1 2 nests in 303 nr of Tournefortia (40/1 ,000 nr ). 

 did we find such density, and for this reason we named the islet 

 "Kota" (Gilbertese for red-footed booby). 



Phenology : In September 1988, we located 339 nests. Of 

 the 152 whose contents could be determined, 87 were empty, 

 63 contained eggs, and 2 held downy chicks. We saw dozens 

 of flying juveniles along the windward coasts. Most pairs were 

 building or guarding their nests during a pre-laying stage that 

 lasts from 1 1-35 days (Nelson, 1969). Of the pairs with nests, 

 57.2% had yet to lay and 41.4% had laid their eggs between 

 mid-August and late September (Fig. 7). Applied to the total 

 breeding population, approximately 1,270 nests were in the 

 prelaying stage and would be expected to produce eggs 

 throughout October. An additional 9 1 9 nests had a mean laying 

 date in early September (Fig. 7). Red-footed boobies were 

 synchronous with brown boobies but delayed relative to masked 

 boobies. 



In June 1965. nests containing prelaying adults, eggs, and 

 young in all stages indicated that the birds were in the midst of 

 a protracted breeding season extending from January to June. 

 Our data reveal that no successful nesting occurred in May- 

 June 1988. Data from March and May 1990 indicate that nest- 

 building began in January (or earlier), with eggs laid from 

 January to May. Red-footed boobies in other tropical locations 

 have variable, opportunistic breeding seasons that depend 

 upon food availability (Nelson. 1978); our data suggest that 

 similar pressures could be operating at Caroline. 



Color Morphs : Red-footed boobies are polymorphic 

 (Nelson, 1978). The basic plumages are brown or white, with 

 brown morphs having many combinations of tail, back, scapular, 

 foot, and bill colors. A variety of brown forms and white forms 

 occurred on Caroline, with a ratio of 9:1 (337 brown to 

 35 white), which contrasted sharply with Nelson's (1978) 

 statement that "in the Line and Phoenix Islands all birds are 

 white morphs." Most of the dark morphs were the "white- 

 tailed" form (see Nelson, 1978. pp. 660-661 ). The variations 

 and proportions of plumage types show clinal change in the 

 Line and Phoenix Islands ( F. Sibley, personal communication), 

 and the question of plumage morphology needs much more 

 study in the central Pacific. 



Great Frigatebird (Fregata minor) (Subchapter 1.1. Figs. 7,8 

 and PL 42) 



The great frigatebird breeds at widely scattered locations 

 throughout tropical waters in the Atlantic, Pacific, and Indian 

 Oceans. It is known to breed on all of the Line Islands except 

 Starbuck (Perry. 1980). 



143 



