21 



element in their estimate by acreage. In tliis estimate, Kitovi is giveu credit for 3| 

 per cent of tlie total number of seals (70,4li3). This figure can not be far from the 

 truth. ]?ut the fact that, iu 189(5, in spite of some shrinkage, Kitovi shows 6,049 

 pups, demonstrates that the figures based on counts made at the height of the season 

 are far from comi)l('te. 6,049 is 3J per cent of 161,060. 



ACREAGE ESTIMATES AND PHOTOGRAPHS UNSATISFACTORY. 



For reasons above indicated, we have been unable to accept estimates of acreage 

 of roolveries or the testimony of maps as indicating with exactness the amount of 

 decrease. Series of idiqtographs taken from year to year show clearly the fact of de- 

 crease, but give no exact data as to its amount. The photographs examined by us 

 go back only to 1892. Those taken later in the year than .July 20 are of little value 

 for cdm])arison, for after that date the outlines of the rookeries change from day to 

 day. Moreover, there is no assurance that photograi)hs taken the same date on 

 successive years show the same or relative conditions, as the arrival of the seals and 

 doubtless their movements on the rookeries are affected by the state of the weather 

 and the advancement of the season. 



TOWNSEND'S CROSSES. 



The most definite evidence of decline since 1895 lies in th.is fact that in that year 

 Mr. Townseud painted a white cross on a rock at the head of each of the large masses 

 of breeding seals as located at the hi'ight of the season. In no case was any of these 

 white crosses reached at the corresponding jieriod in 1896. This shows the fact of 

 the diminution, but not the amount. 



DECREASE OF HAREMS. 



Another item of importance is this: In 1895 Mr. Murray made a careful count of 

 the number of harems of the two islands, finding 5,000 in all. At the same period in 

 1896 he found that the number of harems was reduci'd to 4,853, a loss of 3' per cent, 

 the number of bulls without harems having increased 7 percent. As the same infiu- 

 ence which lowered the number of harems would tend to make them also individu- 

 ally smaller, this might mean a decrease of 3^, times 3' per cent, or nearly 10 percent 

 since 1S95. But as to this there can be no certainty. An estimate of the decrease in 

 the number of pups born in 1896 as compared with 1895 may lie between 3 and 10 per 

 cent, but can not l)e made more exact by any data in our possession. 



ELEMENTS OF CHANGE IN THE HERD. 



The changes iu the breeding herd i'rom year to year lie in the following elements: 



1. The addition of the 3-year-old females who bear their first pups. 



2. The loss of females killed at sea in the previous fall, those killed in the spring, 

 those killed by the bulls, and those destroyed by natural causes. Keeping these 

 elements iu mind, we may attempt a comparison between the breeding herds of 1895 

 and those of 1896. 



In 1896, 30,000 killable males were taken, 22,000 of these, to the best of our infor- 

 mation, being 3-year-olds. Some 2,000 or 3,000 more of these are left, while nearly 

 as many nujre were killed in 1895 as "long'" 2-year-olds. Of male pups born in 1893, 

 therefore, about 26,000 survived till they became killable. As many 3-year-old 

 females entered the rookeries to bear their first pups in 1896. These numbers repre- 

 sent something less than a third of the pups born in 1893. 



A tabulation of the gains and losses from 1895 to 1896 would then run as follows: 



Gains : 3-year-old females bearing their first pup (about) 26, 000 



Losses : 



a Females killed at sea, August, 1895 (about) 23, 000 



b Females killed at sea, April, 1896 (about) 8, 000 



c Females shot at sea and lost, unrecorded (al)out) 5, 000 



^Females killed by bulls, 1896 130 



e Females lost at sea by natural causes, unrecorded (about) 10, 000 



Total 46,130 



Estimated net loss for 1896, 12^^ per cent of total breeding cows 20, 130 



The uncertainty of each of these elements, and esi^ecially of c and e, leaves the cal- 

 culation almost valueless. 



