Table 19. — Estimates of the pup population based on marked -to-unmarked ratios obtained by saJH)llng 

 live pups after shearing, year class 1965, St. Paul Island 



■'■ Based on sanples of 25 pups, 3-4 August. 

 ^ Based on sanples of 25 pups, 12-13 August. 

 ■^ Based on sanples of 100 pups, 2-11 August. 



St. Paul Island at the time of shearing in 1965. 

 The uncorrected estimates for year classes 

 1961-65 for St. Paul Island are given in table 

 20. 



An estimate of the total number of pups born 

 on both Islands may be obtained by adding 

 247,500 to 30,000 (estimated mortality prior to 

 the time of shearing), and extending the result 

 (277,500) by dividing by 0.8, the proportion of 

 Pribilof Islands harem bulls counted on St. Paul 

 Island. This calculation gives 347,000 as the 

 total number of pups born on both Islands in 

 1965. 



Discussion of population estimates .- - The 

 nunnber of pups born, as estimated from tag 

 recoveries, has been decreasing steadily since 

 1959. Estimates based on tag recoveries, how- 

 ever, are about 150,000 higher than the esti- 

 mates based on marking and sampling of live 

 pups in the year of birth. The variability in the 

 estimates based on tag recoveries by 5-day 

 periods was not excessivefor 3-year-old males 

 but was considerable for 4-year-old males. In 



recent years estimates based on tag recoveries 

 have shown less year-to-year variability than 

 in the mid-1950's. Apparently, the factors that 

 caused the highly variable and inflated esti- 

 mates of the middle- 1950's (Roppel, Johnson, 

 and Chapman, 1965) are being eliminated. 



There is some year-to-year variability in 

 the nunnber of pups born, as estimated from 

 marking and sampling of live pups in the year 

 of birth,' but little variability within a year. The 

 accuracy of the estimates, as checked by com- 

 parison with total counts, is satisfactory. Gen- 

 erally, the estimates are probably slightly low 

 even though corrected for the difference be- 

 tween counts and estimates. This bias comes 

 from the disproportionately large represen- 

 tation of small areas where total counts of 

 live pups are made. The total estimate of 

 347,000 born in 1965 nnay, therefore, be 

 slightly low. 



The estimated number of yearling males for 

 the 1961 and 1962 year classes are of the same 

 general magnitude and fit reasonably well with 

 other data. 



20 



