Birds in this category are distributed at ran- 

 dom in the study area from the standpoint of 

 miles from land. Albatrosses were recorded 

 mainly in the northern half of the area, which 

 suggests that albatross distribution shows a 

 closer correlation with latitude-dependent fac- 

 tors--probably temperature and salinity--than 

 land-dependent factors. 



5. Species that demonstrated a bimodal dis- 

 tribution, i.e., high densities within 100 

 miles of land and more than 400 miles 

 from land, but with low density between 

 100 and 400 miles: 



Wedge-tailed Shearwater 

 Newell's Shearwater 

 Sooty Tern 



Wedge-tailed Shearwaters and Sooty Terns 

 are the two most abundant species in the study 

 area. Their distribution is complex. During 

 the breeding season high densities of Sooty 

 Terns and light-phase Wedge-tailed Shear- 

 waters are found within 100 miles of the Ha- 

 waiian Islands. The peaks disappear during 

 the nonbreeding season when these populations 

 migrate to their wintering grounds. A second 

 density center, at the southern end of the area, 

 contains mainly dark-phase Wedge-tailed Shear- 

 waters which come from unknown Southern 

 Hemisphere breeding populations. The cycle of 

 this southern density center is roughly in syn- 

 chrony with the one in the north. If birds from 

 the Southern Hemisphere have breeding cycles 

 6 months before and after the cycle of northern 

 birds--which probably holds true for colonies 

 in subtropical areas but probably not for those 

 in the tropics where breeding cycles may be of 

 extended length — then these birds represent a 

 wintering population rather than foragers from 

 nearby colonies. The cycle of the Sooty Terns 

 in the south is not sufficiently regular to war- 

 rant an analysis. It is clear that at least two 

 populations of each species are encountered, a 

 nearby breeding population and a foreign win- 

 tering population, of which only the former 

 shows orientation to land. Newell's Shear- 

 waters at the southern end are mainly sub- 

 adults or nonbreeding adults. 



Analysis of Density 



Density was analyzed after consideration was 

 given to the maximum distances at which var- 

 ious types of birds could be identified and 



counted. To test the data for patterns of rela- 

 tive density, a further study was made of the 

 distribution of birds in subareas. 



Maximum distances of observation . --Dis- 

 tances at sea are notoriously difficult to es- 

 timate. Even when observers become familiar 

 with the shape, build, behavior, and pattern of 

 markings of sea birds, they can often mistake a 

 species for a similar species of different size 

 because no scale is afforded by the ocean sur- 

 face. The distance at which a given species will 

 be observed most of the time, or overlooked 

 most of the time, is an essential element in the 

 computation of population estimates based on 

 samples made along transects at sea. Several 

 factors interact to produce a maximum obser- 

 vation distance. This maximum observation 

 distance changes from species to species, from 

 day to day, from ship to ship, and from obser- 

 ver to observer. Species-variable factors in- 

 clude size of bird, height at which a bird is 

 flying, flight speed and behavior, color or pat- 

 tern, tendency to flock, and how much a ship 

 attracts or repels the birds. Wind speed and 

 direction, sea state, atmospheric visibility, and 

 amount and direction of glare are among the 

 factors which change from day to day, or even 

 from hour to hour. The size of the ship is im- 

 portant; small ships tend to repel birds less 

 than large ones, but large ships can be seen at 

 a greater distance by birds that are attracted to 

 ships. The height of the observation platform 

 and the extent to which observations are im- 

 paired by ships' structures are important. An 

 observer can see high-flying birds, such as 

 terns, more readily from a high platform and 

 can see low-flying birds, such as storm petrels, 

 more readily from a low platform. Observers 

 differ in the degree and length of their concen- 

 tration, visual acuity, and familiarity with the 

 techniques of observation and identification. 



It is impossible to take all these factors into 

 consideration in the determination of maximum 

 observation distances. To do so would be to 

 reappraise a list of many variables almost con- 

 tinuously. Some of these factors cancel out 

 others, however; some can be disregarded un- 

 der normal conditions; and the effects of some 

 can be minimized by standardization of obser- 

 vation techniques, by the repetition of one 

 cruise track over a long timespan, and by use 

 of experienced observers only. 



Our experience suggests that the following 

 maximum ranges of observations from the 



69 



