TABLE 1 



Relative abundance of seubirds in each sector of the R/V Akademik Korolev cruise track. 1 5 September-3 1 October 1 988. 



Each hour is the combined total of two people simultaneously viewing opposite sides of the ship, and thus equals 2 hours of observation 

 in some other papers. 



relevant seabird records. Most of the previous information on 

 breeding and at-sea distribution of tropical seabirds was gathered 

 from 1963 to 1969 (Humphrey, 1965) by the US National 

 Museum's (Smithsonian Institution) Pacific Ocean Biological 

 Survey Program (POBSP). 



Methods 



Daily at-sea observations were conducted from the flying 

 bridge of the R/V Akademik Korolev, 12 m above the sea 

 surface and within a viewing arc of approximately 180°. 

 During the first week we honed our observation skills, using 

 Harrison ( 1 985 ) as a major reference. Thereafter, we maintained 

 almost constant watch during daylight hours (29 days during 

 9 September-3 1 October). Methods were based on those of the 

 POBSP (King, 1970; Gould. 1974), modified by techniques 

 utilized elsewhere (Tasker <?/«/., 1984;Haney, 1985;Gould& 

 Forsell. 1989). The watch rotated between three observers 

 (AKK, CBK, and DHE), with two observers on watch at all 

 times. A change of one observer took place every hour on the 

 hour, each person alternating 2-hour watches with a 1-hour 

 break to reduce fatigue. Observations began 10-15 min before 

 sunrise and terminated 1 0-1 5 min after sunset. Because we did 

 not have dedicated ship time, observations were interrupted by 

 periodic oceanographic sampling stations, during which time 

 no seabird counts were made. Counts were not conducted 

 within 10 km of oceanic islands where we landed (Christmas 

 Island, Caroline Atoll, and Tarawa), but were made close to 

 land in the Bohol, Sulu. and South China Seas. 



Our ship speed averaged 15 knots during observations, 

 higher than the 10 knots that Gould & Forsell ( 1989) consider 

 ideal. At this time of year in the tropical Pacific, our higher 



speed did not cause identification problems because few species 

 were found in large numbers. When approaching or leaving a 

 sampling station, we interrupted observations if the ship was 

 moving less than 5 knots. 



Bird counts were contained within rectangular strips 

 extending 400 m to each side of the ship. Observers stood left 

 and right of the midline of the ship, counting all birds seen 

 400 m or less ahead of the ship on their side, to a line 

 perpendicular to the ship's direction at their position. Thus 

 each hour of observation represents the pooled records of two 

 observers watching a combined strip 800 m wide. Data were 

 recorded on standardized field forms using local time. The 

 following information was included: identification (to species 

 or subspecies when possible), number of birds per sighting, 

 feeding flocks and other associations, flight direction, plumage 

 (adult, juvenile, sex), and weather. Due to poor lighting or 

 weather conditions, some birds were identified only to genus or 

 family. All birds sitting, flying, or flushed within the transect 

 were counted, and their different behaviors noted. No birds 

 were collected. The ship's position (latitude/longitude) and 

 speed were recorded at the beginning and end of each hourly 

 observation period. Ship-following species were noted during 

 position checks and intermittently during the watch and were 

 recorded when first seen. 



Because the visibility of different species at sea varies 

 greatly, we subdivided the maximum 400 m transect width into 

 3 bands corresponding to the approximate detection distances 

 of each species: /. 1 00-m band — small species, ordinarily seen 

 relatively close to the ship, includes shorebirds, storm-petrels, 

 and Bulwer's petrel (Bulweria bulweri); 2. 30-m band — gadfly 

 petrels, shearwaters, most larids, and anatids, and 3. 400 m 

 band, large or conspicuous species such as boobies, frigatebirds. 



227 



