ALASKA FISHERY AND FUE-SEAL INDUSTRIES, 1921. 79 



of the prevalence of rough weather. A sufficient number of tripods 

 with markers separating the areas to be counted from each would 

 make the harem count very nearly, if not entirely, correct. Tripods 

 would not be necessary where natural elevations occur from which 

 the whole harem area can be plainly seen. 



Several concrete markers were made on St. Paul Island, but it 

 was impossible to place them for use this year. They are similar 

 to a concrete fence post in shape, 1 foot in diameter at the base and 

 6 inches at the top, with an iron rod protruding. Markers should 

 be placed in lines dividing the areas to be counted from separate 

 tripods or natural elevations. 



DATES OF COUNTS. 



The count of bulls was made at the height of the season, July 16 

 to 23, inclusive. The pup count was made this year several days 

 earlier than has been the practice. To be as accurate as possible, 

 the pup count must be made after all are born and before they begin 

 to enter the water. This period is short. Observations on 

 St. George Island rookeries showed that pups could be counted 

 there on July 30. A close watch of the herd, together with the 

 breaking up of the harems and the scattering of the pups, led to the 

 conclusion that few, if any, pups would be born after that date. At 

 the time the count was completed, August 7 on St. Paul Island, it 

 was seen that if it had been delayed a few days many pups would 

 have been in the water where they could not have been counted. 

 On August 7 over 100 pups were in the water at Little Zapadni 

 rookerj^. On August 6, when Morjovi rookery was counted, a few 

 pups were in the water, but as the water was not of a sufficient depth 

 to cover their bodies, none were overlooked in the count. 



PUPS. 



Pups were counted upon the same rookeries that were selected in 

 1920. With the average harem obtained on these rookeries, together 

 with field observations, the number of pups upon each of the other 

 rookeries has been computed. This method is very unsatisfactory 

 in many ways, but it is the best that can be devised under the cir- 

 cumstances. 



The importance of the pup count and the necessity for making it as 

 accurate as possible are recognized from the fact that the average harem, 

 and number of breeding cows is derived directly from such figures. 

 The increasing size of the herd has caused the count to be made on 

 those smaller rookeries which were believed to represent conditions as 

 applying to the herd as a whole. It is known though that the 

 topography of the rookery plays some part in determining the size 

 of the average harem. Field observations have been used in connec- 

 tion with the average harem of the rookeries where pups were counted 

 in computing the average harem of other rookeries. At least one 

 more complete pup count should be made before the herd becomes 

 too large to make such a count prohibitive on several of the larger 

 rookeries. It would bring to light any errors in computation that 

 m^ay have been made in the last few years and would furnish an 

 accurate basis for computations in the future. 



4502°— 22 G 



