waiian island group, the Wedge-tailed Shear- 

 water and Sooty Tern, as well as the two that 

 winter in large numbers in the South Pacific, 

 the Juan Fernandez and Black-winged Petrel. 

 The "winterers" reach peak density in the fall 

 at the southern end of the study area, whereas 

 the breeders reach their peak density in the 

 spring at the northern end. 



Red-tailed Tropicbirds were associated with 

 slightly warmer air on the average than were 

 White-tailed Tropicbirds. Leach's Storm 

 Petrels showed increasing densities at the 

 lower temperatures and Bulwer's Petrels at 

 the higher temperatures. 



As would be expected for a direct migrant, 

 Sooty Shearwaters showed no clearcut associa- 

 tion with a particular temperature. The high 

 density of Slender-billed Shearwaters at tem- 

 peratures from 24° to 26° C. is undoubtedly a 

 coincidence, reflecting the "wave" migration 

 of this species (96 percent of all birds were 

 observed in a 5-day period). 



The relative randomness in the abundance of 

 Great Frigatebirds and White Terns at dif- 

 ferent air temperatures is more difficult to 

 explain. Possibly the sample size of species 

 that show a spotty distribution and are in low 

 abundance would have to be larger to show any 

 clearcut association. A second possibility is 

 that other conditions, environmental or other- 

 wise, have an overriding effect on the distribu- 

 tion of these species. 



The marked association that some species 

 have with certain air temperatures suggest that 

 air temperature may be a contributing factor in 

 the north-south distribution of birds in the 

 study area. 



Surface water temperature . --Many species 

 showed clear-cut associations with certain 

 water temperatures (fig. 34). Albatrosses were 

 seldom found over water warmer than 23° C, 

 whereas Juan Fernandez Petrels, Black-winged 

 Petrels, and dark-phase Wedge-tailed Shear- 

 waters were seldom found over water colder 

 than 23° C. Sooty Terns were most dense at 

 24° and 25°, and also at 27°, a reflection of their 

 bimodal distribution discussed in the species 

 accounts. In general, the species that were 

 distributed predominantly in the north or the 

 south of the study area showed associations 

 with colder or warmer water, whereas the spe- 

 cies that were distributed at random showed no 

 particular temperature associations. These 

 patterns would be expected since water tem- 



perature is directly related to latitude in the 

 study area. Brown Noddy distribution was 

 unchanged regardless of seasonal change of 

 water temperature. 



Surface salinity . --Sea bird densities at var- 

 ious salinity categories in increments of 0.30 

 p.p.t. (parts per thousand) showed several 

 poorly defined patterns (table 8). The most 

 notable were in the albatrosses, which tended 

 to be most numerous over high-salinity water, 

 and the wintering southern petrels — Juan Fer- 

 nandez and Black-winged--which were gener- 

 ally in less saline water. These patterns un- 

 doubtedly reflect the difference in geographical 

 ranges of these species, however, because the 

 water was less saline toward the southern end 

 of the study area than at the northern end. A 

 much clearer picture is presented in the sea 

 surface-temperature analysis, which is more 

 directly a factor of latitude. From the present 

 arbitrarily chosen set of increments it appears 

 unlikely that surface salinity is a significant 

 limiting factor in the distribution of sea birds 

 in the study area. 



From the data presented here it cannot be 

 said that temperature actually limits distribu- 

 tion, but it is clear that relative densities of 

 many sea birds are associated much more 

 closely with air and surface water tempera- 

 tures than with wind direction, wind speed, and 

 surface salinity. 



Flock Analysis 



For this analysis a flock is regarded as a 

 group of five or more birds acting as a unit. 

 Flocks of direct migrants, e.g.. Sooty Shear- 

 waters or Golden Plovers, are excluded from 

 consideration because they use airspace only. 

 Exceptions are the few Sooty Shearwaters (53 

 of 3,725, or 1.4 percent) observed feeding in 

 flocks. Included in the analysis are flocks of 

 birds not observed feeding but which probably 

 fed in the study area during the 24-hour period 

 in which observations took place. This analysis 

 deals with the abundance, distribution, compo- 

 sition, species participation and tendency to 

 flock of 893 flocks. 



Flock abundance . --Birds in flocks accounted 

 for 69.5 percent of the total birds observed 

 (table 9). The 3 months which fell substantially 

 below this percentage — March 1964, November 

 1964, and April 1965--were months of heavy 



91 



