APPEI^DIX. 



Murray, 1894: Page 11. 



That no dead pups Aveie fouud upon the rookeries in 1894 in the early- 

 part of August was due, not to their absence, but to the fact that no 

 close inspection was made. It is impossible without actually going on 

 the breeding grounds and driving oflt' the living cows and pups to get 

 an idea of the number of dead pups. Such an inspection was not made 

 in 1894 nor in any year prior to 189G; consequently the facts regarding 

 the phenomenon of dead pups were never known until that time. The 

 dead pups seen on Tolstoi Rookery in 1891 and 1892 belonged, in the 

 latter year wholly and in the former partly, to this early mortality, 

 which occurs before pelagic sealing begins. What has heretofore been 

 said regarding this estimate of starved pups in connection with the 

 reports of Messrs. Hamlin and Crowley applies here also. In the quota- 

 tion here ascribed to Mr. Crowley appears the statement that in the 

 count of dead pups an effort was made to distinguish the recently dead 

 from those long dead. If this is true, it vv^ould increase the value of 

 the figures as a measure of starvation; but this statement does not 

 occur in Mr. Crowley's report, and in any event, granting that the 

 figures included only starved pups, they still fall short of the facts. 



Murray, 1894: Page 15. 



The several estimates by Mr. Elliott and others here quoted or men- 

 tioned will be discussed in connection with the reports from which they 

 are taken, which appear in later volumes of this series. 



Mr. Murray's estimate for 1891, here given in detail for St. Paul 

 Island, represents a broad and general personal impression rather than 

 an accurate enumeration, as undoubtedly does also that for 1894, the 

 details of which are not given. The elements of weakness in these esti- 

 mates lie in the assumed average size of harem and in the arbitrary 

 doubling of the number of bulls seen in order to account for others 

 supposed to exist but not seen. The size of harem assumed (40) is 

 more than double that of the average number of aninuils ever seen at 

 one time in a harem (17) and is one-fourth larger than the actual num- 

 ber (30) of cows, including absent ones, which the investigations of 

 1896-97 show to belong to the average harem. These figures therefore 

 must be taken with a good deal of allowance and can be held only to 

 represent in a very general way the relative condition of the herd. It 

 may be noted that no higher accuracy was claimed by Mr. Murray for 

 these and subsequent estimates made by him. 



Murray, 1894: Page 23. 



The discussion of dead pups on this and subsequent pages of this 

 report has but little value, because built on the assumption tliat all had 

 died of starvation. This was the common belief until the investigations 

 of 1896 were made. The fact that a large natural mortality, due to 



461 



