FEMALES 



The kill of female seals on the Pribilof 

 Islands in 1966 was 391; 61 were taken for 



research and 330 were killed accidentally 

 during the kill of males. Canine teeth for 

 deternnining age were not collected from 

 females. 



SURVEY DATA 



Data collected in 1966 to followthe response 

 of population to changes in its size included: 

 (1) Counts of dead adults and pups; (2) counts of 

 living adult males; (3) major causes of mor- 

 tality of pups; (4) reproductive condition of 

 females; and (5) weights of live pups. 



MORTALITY 



This section includes data on pups that died 

 on the Pribilof Islands during most years 

 since 1941, and adults that died on St. Paul 

 Island in 1965-66 and on St. George Island in 

 1966. 



Pups 



Biologists have counted dead pups on the 

 Pribilof Islands about mid-August nearly every 

 year since 1941 (table A-5). Until 1966, only 

 the total numbers of dead pups were recorded 

 for each rookery. In 1966, we subdivided the 

 rookeries on St. Paul Island into sections that 

 contained about 100 harem males each in 

 mid-July, and counted the dead pups in each 

 section (table A-6). In future years, the mor- 

 tality of pups within each section will be 

 compared with the density of harem males and 

 other characteristics to determine if they are 

 related to mortality. 



The causes of mortality of pups each day on 

 a selected sampling area of Reef Rookery were 

 also recorded. 



Counts of dead pups . --The count of dead 

 pups in 1966 was 22,485 on St. Paul Island and 

 4,907 on St. George Island--or a total of 27,392 

 (the figures include a 5 percent addition for 

 overlooked pups). The count for St. Paul Island 

 was the lowest since 1941, but only slightly 

 lower than the count of 22,651 in 1964. The 

 count in 1966 on St. George Island was nnore 

 than double the number observed in 1964. 



Estimates of mortality . --Counts of dead 

 pups are useful for showing changes in mor- 

 tality on land over short periods of time when 

 it can be assumed that the same number of pups 

 are born each year. Comparisons of mortality 

 over several years or between pup populations 

 on different rookeries, however, cannot be 

 made unless the fractions of the population 

 represented by the counts are known. 



Estimates of the number of pups born based 

 on tag recoveries from 1956 to I960 cannot 



be used with counts of dead pups to calculate 

 mortality by rookery, because the estimates 

 were inaccurate and highly variable from year 

 to year. 



Estimates based on sampling for marked-to- 

 unmarked ratios in the year of birth became 

 available beginning in I960, By 1963, this 

 method was sufficiently improved to make the 

 estimates, in conjunction with complete counts 

 of live pups, reasonably accurate. To overcome 

 year-to-year variability due to sampling, the 

 means of the estimates from 4years (1963-66) 

 of marking and sampling were used to develop 

 estimates of mortality for St. Paul Island 

 (table 2). Live pups were first sheared and 

 sampled on St. George Island in 1966, and we 

 used the estimate of the number of pups born 

 to calculate mortality by rookery in 1966, and 

 also in 1963, 1964, and 1965 (table 2). We as- 

 sumed that the population was constant during 

 the 4-year period. Though the populations of 

 both islands rniay have changed somewhat, 

 sampling for marked-to-unnnarked ratios prob- 

 ably provides the best estimate of the pup 

 population on which to calculate annual mor- 

 tality by rookery. 



Total mortality of pups varies considerably 

 fronn year to year, and the loss, in proportion 

 to the nunnber born, tends to be higher on 

 certain rookeries than on others. In addition, 

 the mortality of pups in 1963-65 was much 

 lower on St. George Island than on St. Paul 

 Island. A correlation coefficient of 0.90 be- 

 tween islands shows that even though mortality 

 on each followed the same trend (whether up 

 or down), the nnagnitude of the changes has 

 varied widely (fig. 5). This variation also held 

 for individual rookeries on each island. 



Causes of pup mortality . --To determine the 

 relative importance of the nnajor causes of 

 pup mortality and relate these causes to 

 variations in the size of the total population, 

 an annual survey of pup mortality was begun 

 in 1966. This program was based on a study 

 of pup miortality made in 1964 when all dead 

 pups that could be reached from a catwalk 

 were removed daily with a 5 -mi. gaff and 

 autopsied (Keyes, 1965). The catwalk used in 

 1964 and 1966 is on the east shore of Reef 

 Rookery between tripods 5 and 6-7 (fig. 6) 

 and provides access to about 1,672 m.^ of 

 breeding space. The survey in 1966 was begun 

 on 28 June when the first dead pups were 

 found, and ended on 22 August, after which 

 very few pups die. 



