FISHERY BULLETIN: VOL. 85, NO. 2 



Pups were counted by driving off the adult animals 

 just ahead of the counting team. 



Counts of northern sea lions were made daily by 

 observers on Ugamak Island between 1 June and 

 3 July 1985 and between 16 June and 26 July 1986. 

 Hourly counts were made between the hours of 0700 

 and 2400 ADT for 6 days in 1986. Animals were 

 counted from the cliffs above the sites, using 7 x 

 35 binoculars, a 15-60 power spotting scope, and 

 unassisted vision. Counts were made of animals ac- 

 cording to five types: adult territorial male, other 

 adult male, adult female, juvenile, and pup (Merrick 

 1984). Animals in the water were excluded from 

 counts. Freshly dead pups were recorded when seen, 

 with pup mortality estimated as the total number 

 of dead pups divided by the maximal number of pups 

 counted (living and dead). 



Data Analysis 



The number of northern sea lions counted in the 

 1984-86 surveys were compared with counts from 

 surveys in 1956-79 conducted by Kenyon and Rice 

 (1961), Kenyon (fn. 3), Mathisen and Lopp (1963), 

 Braham et al. (1980), Fiscus et al. (1981), and 

 Calkins and Pitcher (fn. 5). Differences in survey 

 areas complicated comparisons for the entire region, 

 so comparisons were generally performed by area. 



Comparisons also were complicated by the differ- 

 ences in the counting methods used. Some were 

 counts from land (Fiscus and Johnson 1968^; Fiscus 

 et al. 1981; Withrow 1986io), while others were 

 estimates from ships (Fiscus and Johnson fn. 9; 

 Calkins and Pitcher fn. 5), and counts from aerial 

 photographs (Mathisen and Lopp 1963; Braham et 

 al. 1980; Calkins and Pitcher fn. 5). The most ac- 

 curate were visual counts from land and from aerial 

 photos (Withrow 1982); these were the methods 

 used in the 1984-85 aerial surveys. 



Several assumptions were made in the analysis of 

 these data. The first was that all sites with more 

 than a few animals were surveyed in 1984-85. Sec- 

 ond, the dates and times of peak seasonal and daily 

 abundance were considered to have remained con- 

 stant throughout the 30-yr period. The 1984-85 



surveys and those conducted by Braham et al. (1980) 

 were scheduled to coincide with these peaks; where- 

 as, those of Kenyon and Rice (1961) and Kenyon (fn. 

 3) were conducted in the spring and without regard 

 to time of day. A smaller proportion of animals were 

 probably onshore in the spring than in the summer 

 (Mathisen and Lopp 1963; Braham et al. 1980). 

 Third, the proportion of the population onshore was 

 assumed to have remained unchanged in the 30 

 years of counting. Finally, double counting was con- 

 sidered to be negligible in all the surveys because 

 large areas were surveyed in a single day. 



Counts presented here are indices of population 

 size because they exclude animals at sea and because 

 it is difficult to count at the exact time peak num- 

 bers are ashore. There are few data on the propor- 

 tion of animals that are at sea at the time the peak 

 number is onshore. Consequently, it was necessary 

 to assume that the proportion had not changed over 

 time. Even during the period when maximal num- 

 bers of animals were expected onshore there was 

 variation due to weather and tidal affects (Withrow 

 1982; Merrick 1984), so that it was unlikely that a 

 survey would occur on the day and time of peak 

 numbers. However, because the sites on Ugamak 

 Island were counted daily, the maximal number 

 counted there was a closer approximation than the 

 aerial survey counts of the actual peak number of 

 animals onshore during the breeding season. Thus 

 the Ugamak Island data were used to determine if 

 seasonal and daily variation in northern sea lion 

 hauling patterns had changed and to assess the 

 potential amount of error (due to counting at the 

 wrong time) in the aerial photo counts. 



Rates of decline between two points in time were 

 calculated using the formula 



where A^o 

 t 



d = 



N, = Nod< 



count in base year 

 count in future year t 

 number of years between the base 

 year and year t 



rate of change, with the percent an- 

 nual change calculated as {d - 1) x 

 100. 



'Fiscus, C. H., and A. M. Johnson. 1968. Site for research on 

 the Steller sea Hon, June-July 1968. Processed Rep., U.S. Fish 

 Wild). Serv., Bur. Commer. Fish., Mar. Mammal Biol. Lab., 

 Seattle, 33 p. Northwest and Alaska Fish. Cent. Natl. Mar. Mam- 

 mal Lab., NMFS, NOAA, 7600 Sand Point Way N.E., Seattle, WA 

 98115. 



'"D. Withrow, Northwest and Alaska Fisheries Center National 

 Marine Mammal Laboratory, National Marine Fisheries Service, 

 NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115, pars, 

 comm. January 1986. 



Area counts were regressed as a linear function of 

 time to determine if trends in population sizes ex- 

 isted. Student's ^test was used to assess the signif- 

 icance of the regressions. Wilcoxon's signed rank 

 test was used for between year comparisons of 

 paired site counts within an area (Hollander and 

 Wolfe 1973). 



354 



