FUR-SEAL CENSUS, PEIBILOF ISLANDS, 



1919. 



109 



As the number of bulls is reduced in the herd the counting of pups 

 ^vill become easier and the work can be midertaken somewhat earlier. 

 If the force of men recommended is supplied in 1920 doubtless some 

 of the larger rookeries can be counted again. But unless a thoroughly 

 accurate and satisfactory count of harems can be made on a rookery 

 the count of pups is not so valuable as it otherwise would be. 



PUPS. 

 Distribution of Pups in 1919. 



a Based on cstimatetl average harem. 



The number of dr:u\ pups was as('(>rtainod on each rookery counted 

 at the time tlie count was made. On those rookeries on which the 

 cows and pups were estimated the number of dead was computed 

 from tlie percentage of dead used in tlie 1918 report. The small 

 increase in the average liarcrn, especially on St. George, should have 

 redu('ed the percentiige of dead pups to a slight extent. On St. Paul, 

 con<litions were so little changed in this respect, however, that the 

 difference would not overbalance variations due to the laws of 

 chance. Therefore until average harems become as large as they 

 have been in some former year (as 1916 for instance) when the death 

 rate was determined for all of the rookeries it seems best to make no 

 change in the percentages. The rookeries which were counted in 

 1919 gave no data upon which to base a reduction. In fact there 

 was in some cases evidence of a slight increase in the number of dead 

 pupa. 



