THE FUR SEALS OF THE PRIBILOF ISLANDS. 



all this the total results as here given are close enough to the actual conditions for all 

 practical purposes. The total of 157.405 breeding seals means between 150,000 and 

 160,000. No closer accuracy is claimed for the figures, and none is needed. The margin 

 of error can not be greater than this. 



THE CENSUS OF 1897. 



In making the census of 1897 no such difficulty as that encountered in 1896 was 

 met with. From the start we had a clear idea of the problem and were guided by the 

 experience obtained in the work of the former season. The first thing to be done \\ as 

 to make a full enumeration of the breeding harems on all the rookeries at the height 

 of the season. The remaining step was to obtain an average size of harem by making 

 a count of live pups on some typical rookery space. Accordingly between the period 

 of July 13 and 20 the count of harems was made. For purposes of comparison the 

 cows actually present on rookeries counted last year were recounted this year on 

 the same dates. 



THE COUNT OF PUPS IN 1897. 



The count of pups was made between the 28th of July and the 3d of August. 

 A comparison of the number of pups with the number of cows counted showed clearly 

 the inadequacy of our correction of 75 per cent for absent cows in the original census 

 of 1896. These counts of cows and pups for 1897 are as follows: 



Comparison of counts of cows and pups, 1897. 



THE PROPORTION OF COWS TO PUPS. 



These counts seem to show that the cows belonging to any rookery for the season 

 are about 2.61 times as numerous as the cows found present at any one time in the 

 height of the season. But an examination of the figures for Zapadni Reef and for 

 Polovina cliffs indicates an abnormal condition of things on these rookeries. In both 

 cases the pups outnumber the cows nearly three to one. It would seem likely that 

 the weather or some other cause had on these rookeries produced an unusual effect 

 for the day when the counts were made. 



In view of this apparently abnormal condition in part of the rookeries counted, 

 it has been thought best in the census of 1897 to waive the general average results 

 and take the simple conditions of Kitovi rookery as again typical. We find from a 

 comparison of this rookery for the two seasons that its condition is apparently 

 normal. In 1896 there were 182 harems, 3,152 cows present on July 13, in the 

 height of the season, and 0,049 pups in August; in 1897 there were 179 harems, 2,436 

 cows actually present on the same date, and 5,289 pups in August. 



