STATISTICS OF STARVED PUPS. 



169 



the accuracy of which can not be questioned. The following table will give in detail 

 the statistics of the estimate: 



Pup statistics nummary. 



a These are the figures of the original census of 1896 as published in the preliminary report of 

 that year. 



b The figures herein given for starved pups on the rookeries of St. George are estimates based 

 u]K>n the conditions of St. Paul. 



STARVED PUPS IN 1897. 



The investigations of the season of 1897 have made the count of dead pups in 

 1896 seem still less satisfactory. It may be said that the dead pups lie concealed 

 among the rocks, and as they quickly wear away under the action of the elements and 

 the trampling of the living animals they are not easily seen in the hasty inspection, 

 which alone is possible in counting them. A good illustration of the probable 

 inadequacy of the counts of dead bodies is shown by the results on Kitovi rookery. 

 A count of this breeding ground, made with a good deal of care on the 3d of August, 

 disclosed 12C dead bodies. When a week or ten days later the carcasses were actually 

 gathered up and removed from the beach, the closer inspection disclosed 202 dead. 

 This inadequacy of the counts of dead pups in 1896 is, however, of such a nature as 

 to make the case all the stronger, because it leaves an under, rather than an over, 

 estimate. 



