CENSUS OF ROOKERIES IN 1896. 



93 



CORRECTION FOR ABSENT COWS. 



Averaging the various counts it seemed that the proportion of live pups to the 

 apparent number of cows present in the breeding season was 1.75. Our experience 

 during the season of 1897 shows that this was too low, and that it would have been 

 better had the simple conditions of Kitovi rookery been taken as typical of the other 

 rookeries, using its proportion of 1.91 to 1. 



On the basis of our count of live pups the previous census was corrected by the 

 addition of 75 per cent to allow for absent cows in the height of the season. In this 

 way the completed preliminary census put forward in 1898 was made. It is as follows : 



Summary of breeding seals (counts and estimates) 1896. 



Bobrovi (Otter Island) had 1 harem, containing 5 cows and 5 pups. 

 REVISION OF CENSUS OF 1896. 



As already intimated, the estimate for St. George Island for 1896 is not satisfactory. 

 That Little East rookery was considerably underestimated is evident from a closer 

 inspection of this breeding ground during the season of 1897. The original count 

 gave only 27 harems, but probably included no more than two-thirds of the rookery. 

 As there were but few idle bulls about this rookery, the count of bulls made late in 

 July can not be far out of the way here, and in the absence of a better estimate it 

 may be allowed to stand. 



CORRECTIONS ON ST. GEORGE. 



On Zapadni and North rookeries, however, the number of harems counted in the 

 latter part of July can not be accurate. When the rigid harem system breaks up and 

 the original harem masters begin to leave, the idle bulls (of which there were many 

 about these rookeries) enter the breeding grounds, increasing the number of families, 



