10 
BULLETIN OF THE BUREAU OF FISHERIES 
in this way arrived at the approximate number of cows on the given rookery." It 
turned out that his figures " approached sufficiently near the autoptic estimate of 
the rookery guard" on Copper Island. With regard to the Bering Island rookeries, 
where his " area method" had proved a failure, he had to devise a different method, 
and as his figures on Copper Island had " approached sufficiently near" those of the 
watchmen, he assumed that those of the Bering Island guard were equally reliable. 
He consequently adopted the latter, applying to them, however, a "correction 
D is t vibutioTt of Seals J910 
Fig. 16. — Karabelnoye Rookery, Copper Island, showing distribution of seals in 1910, according to Suvorof 
coefficient" deduced from a comparison of his own count on Copper Island with the 
eye determination of the watch. In this way he obtained the figures of the following 
table, which in his opinion approximately express the number of animals on the 
rookeries, the number of bachelors present being practically negligible. He remarks 
specificially that the estimate relates to the number of seals left after 1,017 cows and 
500 bachelors had been killed. 
