ALASKA FISHERY AND FUR-SEAL INDUSTRIES, 1920. 105 



The chief object m the counting of pups is the determination of 

 the average number of cows to each bull, or the average harem. 

 Since birth is given to but one young each year it follows that the 

 determination of the number of 3'oung will give the number of 

 breeding females, and by dividing this by the number of harem 

 bulls found earlier in the season the average harem is determined. 



Obviously the greater the number of pups counted the greater the 

 accuracy of the census as a whole. Up to and including 1916 it was 

 possible to count this class on all of the rookeries. In 1917, however, 

 the greatly increased number of bulls prevented pup counting until 

 such a late date that breeding areas could not all be gone over. The 

 increase in size of the herd has further complicated matters. While 

 it is not physically impossible to make a complete pup count when 

 bulls are present in no greater numbers than in 1920, still a much 

 larger force of counters would be required for the work than has 

 heretofore been available. 



Since 1917 a comparatively small proportion of the pups has been 

 counted. Eookeries have been chosen, however, which were believed 

 to represent the herd as a whole with regard to growth. They have 

 been, in the main, the smaller ones, in order to reduce to a mini- 

 mum the danger of loss from trampling bj'^ bulls and from smoth- 

 ering. 



Neither of these factors enters largely into the operations if a 

 competent force of white men is performing the woik, because 

 there are expedients which can be used in emergencies which effec- 

 tually j)revent deaths. For instance, if pui)s have piled up in the 

 counting and are in danger of smothering, they can be scattered 

 by a man wading into the mass. There is no other known means 

 whereby the animals can be spread out quickly enough to prevent 

 loss on a warm day. It so happens that the natives are entirely too 

 irresponsible to be depended upon in surh an emergency. In seven 

 years of counting I have never seen one offer to do the scattering 

 in such an emergenc}'. It is obviously impossible for the j)erson 

 who does tiie actual counting to keep constant watch of the " pod- 

 ding" ahead of him, and if there should be no one else looking after 

 that part of the work danger is sure to result; this is particularly 

 true on the larger rookeries. I^nfortunately the natives allowed 

 eight pups to smother on Zapadni Keef rookery during the work on 

 St. Paul in 1920. They have been included among the live pups on 

 that rookery in the table because it is desirable to have the number 

 of dead represent the natural loss only. 



In selecting rookeries for counting in 1920 it seemed desirable in 

 many ways to choose those which were counted in 1919. The Avork- 

 ing out of the average harem for those uncounted I'ookeries would 

 then give results whic-h were more strictly comparable thati if new 

 territory were .selected. 



While it seems reasonable to suppose that any one rookery would 

 grow at the same rate as the herd, this is unfoitunately far from 

 the c-ase. There is great variation among the several rookeries and 

 on an}' one from year to year. This makes the estimating of the 

 average harem on rookeries where pups have not been counted more 

 or less uncertain. But uidess all rookeries are counted — a manifestly 

 impracticable task with a limited force and a large herd — no way 

 to avoid the difliculty is known, and the matter mu.st devolve upon 

 the best judgment of those who have it in hand. 



