FUR-SEAL CENSUS, PRIBILOF ISLANDS, 1919. 



By G. Dallas Hanna. 



A census of the Alaska fur-seal herd was taken during the summer 

 of 1919 by employing the methods which have been in practice for 

 several previous years. These, together with the habits of the ani- 

 mals and the history of the herd have been fully described in the 

 annual reports from 1914 to 1918, so that in order to avoid needless 

 repetition these subjects will not be treated herein, except in so far 

 .'♦6 seems necessary to properly present the new data. Some new 

 facts in regard to the natural history of the animals, learned in 1919, 

 will be mentioned, however, in order to keep this phase of the subject 

 up to date. 



A complete count of bulls was made at the height of the breeding 

 season. To facilitate the count, driftwood and paint markers were 

 placed on some of the rookeries which are more difficult to count. 

 While these were very helpful they only emphasized the need of 

 permanent work of this character carried on with the erection of 

 overhead observation positions. 



After the breaking up of the harems, pups were counted on as many 

 rookeries as possible in order to determine the average harem. 



Preliminary harem counts (not tabulated herein) were made on St. 

 George Island by C. E. Crompton and on St. Paul by the author. 

 These proved helpful in the final counts and afforded a fair check on 

 the work. 



In the harem counts, H. D. Aller assisted on St. Paul and C. E. 

 Crompton on St. George. The same men assisted in the pup counts 

 on the same islands. Besides this Mr. Crompton made the pup count 

 of Stara3'a Artil Rookery on St. George and Mr. Allor counted 

 Polovina Cliffs Rookery on St. Paul. 



DIFFICULTIES ENCOUNTERED. 



As the seal herd grows a census taken by present means becomes 

 less and less accurate. This is necessarily the case because of the 

 large masses of animals dealt with, and unless an elaborate system of 

 markers and elevated viewpoints be established it will not be many 

 years until the count of harems will give no more accurate an enumera- 

 tion than an estimation by the area method. The construction of 

 the necessary apparatus, however, is believed to be feasible, and in 

 view of the accurate knowledge which can be gained in regard to the 

 state of the herd it would seem to be very profitable. Under present 

 methods of calculation the harem count is fundamental. If it be far 

 wrong the entire census becomes an estimate with figures which look 

 mathematically exact. When the herd was small the harems could 

 be counted exactly, but this is not true to-day. Therefore the read ere 

 of this report should bear in mind that although figures for the different 

 classes are given exactly, only round immbers are intended. A com- 

 plete census is an absolute impossibility. A close approximation to 

 106 



