26 



BULLETIN OF THE BUREAU OF FISHERIES. 



stitutes a slight element of unavoidable uncertainty. It is plain also that the more 

 idle bulls there are the more single cow harems may be expected. The maximum 

 development of the different rookeries is not strictly contemporaneous, and this also 

 adds a variable feature to the harem counts. Such irregularities are probably com- 

 pensated in the results from year to year, and in any event the total number of harems 

 and idle bulls is not affected. 



The method of counting is simple and reliable. The rookeries are mostly extended 

 along the shore in Hnear formation frequently beneath low cliffs from which the observer 

 can look over them with ease. In the present condition of the herd the number of 

 bulls in tier formation between the shore and the back of the rookery does not often 

 exceed five, and marked rocks and natural prominences are sufficient for all necessary 

 subdivision of rookery space into areas for successive counting. A few of the massed 

 sections, as the flat under Hutchinson Hill and certain parts of Reef Rookery, offer 

 difficulties which will mcrease as the herd grows and which could be overcome by simple 

 devices. In such places repeated counts were made by four individuals until complete 

 agreement was reached. The large relative size of the bull makes him conspicuous 

 even at a considerable distance, and except when fully recumbent in a heavily massed 

 area, he can not possibly be overlooked. 



Preliminary counts. — In order to overcome the lack of previous experience and to 

 make general preliminary observations, numerous counts of harems and various classes 

 of seals were made before the height of the season. In this way counts were made at 

 least once for every rookery on St. Paul Island and some rookeries were counted from 

 three to six times. In addition, weekly counts were made of all the rookeries on St. 

 George Island in late June and early July by Mr. G. Dallas Hanna. Therefore, when 

 the height of the season arrived those engaging in the count were familiar with the 

 peculiarities of each rookery and all were agreed as to the method to be employed. 

 As early as June 29, the total number of bulls in position on St. Paul Island was 1,060. 



Owing to exigencies of transportation, it was necessary to make the count of 

 harems on St. George Island on July 13-14, 1914, a few days earlier than desirable. 

 On these dates, 219 harems and 12 idle and young bulls were found as follows: 



Early count of harems, Si. George Island. 



Height 0} season counts. — The height of season counts on St. Paul Island were made 

 July 17, 18, and 19. Sea Lion Rock, or Sivutch, was counted July 20. The counts 

 were made jointly by Messrs. Osgood, Parker, Preble, and Harmon except that of Sea 

 Lion Rock which was counted by Parker and Harmon only. The early count made 

 on St. George Island being obviously unsatisfactory, arrangements were made for a 



