34 



BULLETIN OF THE BUREAU OF FISHERIES. 



should have been 40,992 of them as yearlings in 1913. As none of them were killed, 

 they should have returned the next year in numbers undiminished except from natural 

 causes. As a matter of fact they appeared in 191 4 in large numbers, constituting in 

 the latter part of the season at least two-fifths of the bachelors found on the hauling 

 grounds. Exact enumeration of them is impossible since all are not present at any one 

 time. In 191 2, 5,529 of these seals were branded as pups and a considerable number 

 of these were found throughout the season of 1914, but this furnishes no criterion of the 

 total number of surviving 2-year-olds. The only feasible method of estimating them 

 is by subtracting a fixed percentage from the number estimated as yearlings the pre- 

 ceding year. This percentage has been rather arbitrarily determined as 15 per cent, 

 but from experience during commercial killing in past years it is e\'ident that the result 

 obtained in this way is a consers^ative one. That is, in former years with the herd 

 approximately the same size as now and in spite of the drain of both land and pelagic 

 killing, the lessees found it possible to obtain a quota of 2 -year-olds as large or larger 

 than the number estimated in this way. Deducting 15 per cent from 40,992, the number 

 of yearlings estimated for 1913, gives 34,844 as the number of 2-year-olds in 1914, half 

 of these being males and half females. 



Three-year-old males. — These were bom in 191 1, a year for which only very incom- 

 plete data are available. No count of pups was made in that year, nor any determination 

 of the average harem even for a single rookery. The count of harems was made, how- 

 ever, and this combined with knowledge of the conditions in 1910 and 191 2 furnish prac- 

 tically the only data for estimating the n umber born in 1 9 1 1 . There are two methods of 

 making such an estimate, one by deductions drawn from the average harem on a single 

 rookery known for 1910 and 191 2, the other from the count of pups in 191 2 and the 

 relative effect of pelagic sealing. 



The average harem method may be considered first. During pelagic seaUng or in 

 all years pre\'ious to 191 2, the birth rate for a given year was estimated by counting 

 the pups on one or several rookeries only and determining the average number of pups 

 to a harem for these rookeries, after which this average harem was multiplied by the 

 total number of harems, the result being the supposed total number of pups, and by 

 inference, the number of cows. Applying this method to the years 1912, 1913, and 

 1914, for which we have actual counts, it is apparent that the estimates for former 

 years must be greatly below the facts. This is shown by the following tabulation : 



Coml><iriion of actual counts of pups with estimates based on an average harem. 



Total number of harems 



Pups counted on Kitovi 



Average harem on Kitovi 



Estimate of pups, entire herd, under average 



harem method 



Actual count of pups. 



Percentage of underestimate 



1,387 



1,979 



36.0 



1,381 

 1, 966 



3I-7 



1,369 



<'34.S 

 47,230 



i,3S8 

 1,975 

 37-3 



5°, 6S3 

 81,984 



1,403 

 I,8SS 

 43. 3 



59,206 

 92, 369 

 35-8 



1,559 

 2, 119 

 36. 5 



56.903 



93,250 



38.9 



O' Mean between figures for 1910 and 1912. 



The inference is thus very strong that the estimates for 1909 and 1910 are less than 

 two-thirds of the actual number and that a similar estimate for 1911 would be propor- 

 tionatelv small. Since the total number of harems (actually counted) for 191 1 falls 



