terns, bristle-thighed curlews, tattlers (Heteroscelus incanum 

 and H. brevipes), and "a great number of small pigeons" with 

 white heads (certainly noddy terns, perhaps both A. minimis 

 and A. stolidus). The "shoal birds" that greeted him were 

 probably sooty terns. His most unusual contribution was 

 mention of a possible flightless rail: "The other birds of the 

 coast were a kind resembling a cool..." (p. 372). 



The 1883 Solar Eclipse Party (Subchapter 1.1. History of 

 Caroline Atoll section) published a few sketchy notes, adding 

 lesser golden-plover (Pluvialis dominicd), reef heron (Egretta 

 sacra), and masked booby (Sula dactylatra) ("gannet") to the 

 bird list. Of dubious identity were two species of "seagull" and 

 a "snipe" (Dixon. 1884). Holden. one of the astronomers, 

 heard "the notes of a singing bird." which prompted us to add 

 mist nets to our equipment in the hopes of capturing an 

 Acrocephalus warbler. This resulted in our discovery of the 

 long-tailed cuckoo (Eudynamistaitensis) (Ellis et al., 1990) and 

 piqued our curiosity about what Holden might really have 

 heard. 



The POBSP expedition spent 3 days on Caroline in June 

 1965. They found 10 species of seabirds (9 breeders). 

 4 migrant shorebirds, and a reef heron (Clapp& Sibley. 1971a), 

 providing rough population estimates for each species. This 

 w ork laid the foundation for later expeditions. Brief visits to 

 Caroline by the Kiribati government in 1974 and Roger Perry 

 in 1977 (Garnett, 1983) added no further information. 



The 1988 expedition to Caroline was longer and more 

 extensive than all former visits. We found three new island 

 records: a breeding seabird ( red-tailed tropicbird ). a shorebird 

 (Sanderling [Crocethiaalba]), and a migratory land bird (long- 

 tailed cuckoo), and determined islet-by-islet distributions for 

 each species. Our population estimates, calculated from field 

 work, aerial photographs, and detailed vegetation analysis. 

 indicate that Caroline's avifauna is far more important than had 

 previously been suspected (King, 1973; Garnett. 1983). In 

 March and May 1 990. the ICBP 1990 Line and Phoenix Islands 

 Expedition (Subchapter 1.1. Methods section) filled in minor 

 gaps in our knowledge. Caroline's residents added another 

 breeding seabird, the blue-gray noddy, in summer 1990. 



Methods 



Distribution and Habitat Preference 



We described se\ en major plant communities on Caroline 

 Atoll (Subchapter 1 . 1 i. With the use of aerial photos and the 

 transect data, we mapped the communities found on each islet. 

 Bird distribution was determined and plotted using these islet 

 vegetation maps. If a species nested within a particular plant 

 community, it was plotted on the distribution maps as occurring 

 throughout that community unless determined otherwise. 



Population Sizes and Breeding Phenology 



We measured transect distances for each islet using a hip- 

 chain and biodegradable cotton thread. We recorded all birds 

 seen within the 30-m-wide strips: transect width was estimated 

 visually. We assigned birds to one of several mutual I v exclusive 

 categories: adults present, adults on territory, adults on nests 

 (contents unknown ). eggs, naked chicks, dow ny chicks, chicks 

 w ith remiges erupting, chicks with scapular feathers, or chicks 

 in juvenile plumage. We created a range of possible laying 

 dates for each egg and chick using known growth parameters 

 for each species (Kepler. 1978; Kepler & Kepler. 1978). This 

 enabled us not only to estimate seabird populations, but also to 

 determine and plot a rough breeding phenology for each 

 species (Figs. 3.5.7.9.10.12). In these figures, the height of the 

 bar for each category ("downy," "scapulars." etc.) represents 

 the number of nests found or estimated with that development 

 stage in September 1988. The bar width represents the 

 approximate time span over which eggs could have been laid 

 to produce that stage, while the "no. day s" is a count back from 

 the survey dates to accommodate growth and development that 

 had occurred. Thus, while each figure shows what breeding 

 stages we found, we extend those nests back in time to show- 

 roughly when they would have begun. The number of clutches 

 begun per day is determined by dividing the number of nests 

 per stage by the time span in days over which those eggs were 

 laid. 



Sooty terns nested in dense colonies. Each colon) was 

 mapped, and its total si/e (m') w as calculated. A minimum of 

 10 plots (3x3 mor.3 x 6 m). within which all eggs and chicks 

 were counted, were randomly located along a compass line in 

 each colony. The population size of each colon) w as estimated 

 from these plot densities. 



From 22 2 l > September I9S8, C. B. Kepler. A. K. Kepler, 

 D. H. Ellis, and K. Teeb'aki surveyed all of Caroline's 39 islets 

 except North Arundel Islet, naming most of them (Fig. 1 ; see 

 Subchapter 1.1, Methods section). We established 50 linear 

 transects, extending 13,300 x 30 m. laid out to ensure that at 

 least y< of each islet was sampled lor birds and plants (see 

 Subchapter I.I. Methods section and Fig. 8). Sampling was 

 increased with 19.300 m of perimeter surveys along the 

 windward and leeward coasts of 21 islets (Subchapter 1.1. 

 Fig. 9). On Noddv Rock. Skull. Atibu. Bo'sun Bird. Coral. 

 Reef-flat, and Fishball (Fig. 1). we made total counts of the 

 breeding seabirds. All sur\ cv s w ere conducted during daylight 

 hours. Some incidental data have been added from the 1990 

 ICBP expedition. 



Mist Nets 



We operated 4 ATX 4-shelf 36-mm mesh mist nets 



(2.6 x 12 m) 



43.5 net hours, according to the following 



schedule: 14.5 net hours (daylight) beneath a 10-15 m Cocos 

 canopj on South. 27.5 net hours (day and night i in Pisonia- 

 Cocos interface 1 12 m tall) near Tr. 10 on Long, and 1.5 net 

 hours m Pisonia-Tournefortia within a 4-6 m canopj onTr.4, 



Long. One cuckoo was collected (USNM 607 19 1 1. 



Collecting Other Vertebrates 



Lizards that were active and conspicuous were collected at 

 base camps on South and Long, either bv hand or with a 

 blowgun firing steel darts. No attempt was made to search lor 

 reptiles under coral, litter, or in other concealed locations 



140 



