462 APPENDIX. 



totally different causes, occurs prior to August 10, and lias probably 

 occurred for centuries, must be kept constantly in mind in reading all 

 early discussions of dead pups. 



Murray, 1894: Page 27. 



The figures for tlie pelagic catch here given include also the seals 

 taken on the Asiatic side, a fact which is not made clear. 



Murra)', 1895: Page 452. 



This detailed estimate of seals for 1895 is doubtless the most elabo- 

 rate and accurate which Mr. Murray has made. It, however, contains 

 manifest inconsistencies, as for example, Lagoon Rookery is estimated 

 at 50 harems and 2,000 cows. This rookery was counted in the same 

 season both by Mr. True and by Mr. Townsend. The latter found 80 

 harems and 1,21(5 cows, the former 82 harems and 1,264 cows. Again, 

 on Kitovi Eookery 200 harems and 8,000 cows are found, whereas Messrs. 

 True and Townsend in the same season found 145 harems and 2,640 

 cows. Moreover, the figures themselves show that no account is taken 

 of numbers less than 50 in the enumeration of harems. But the most 

 serious defect in the enumeration arises from the date at which it was 

 made. Mr. Murray assumed that the rookeries were at their height by 

 the 20th of July and, beginning his enumeration at this time, completed 

 it on August 14. Our investigations for the past two seasons show that 

 the height of rookery development falls about the 15th of July; that 

 by the 20th the harems are beginning to break up, and that the mating 

 season for adult seals is ^practically over by August 1. Counts and 

 observations made after the 20th of July give no true idea of condi- 

 tions in the height of the season, and those made during the first half 

 of August show wholly different conditions. Then the-original harems 

 are broken ui3. The regular bulls are gone, and their places are tilled 

 with young and idle bulls controlling transient harems of virgin cows. 

 This enumeration of the seals, therefore, has only the value of a per- 

 sonal estimate made at an unfavorable time and under a misapprehen- 

 sion of the facts of rookery development. 



We may h6re contrast the various estimates offered for the season of 

 1895, and express our regret that such variant and contradictory 

 results should be reached and published by duly accredited agents of 

 the Government: 



Agent. Harems. Cows. 



True 4,402 70,423 



Crowlpy 5,552 99,936 



Murray 5,000 200,000 



It may, however, be observed that all this work was conscientiously 

 and intelligently done. The trouble lay in the methods employed. It 

 is a curious fact that the estimate most carefully and accurately worked 

 out is farthest from the truth. This resulted chiefly from the vitiating 

 assumption that practically all the cows were present on the rookeries 

 at the height of the season. 



Correspondence : Page 357. 



The estimates of starved pups here given include also pups which 

 died of natural causes prior to the beginning of pelagic sealing. Eef- 

 erence should be made to notes upon this subject aj)pended to the 

 reports of Messrs. Hamlin, Crowley, and Murray, where the subject is 

 discussed at length. 



