268 THE ASIATIC PUR-SEAL ISLANDS. 



This table is graphically illustrated by the broken line in diagram h on plate 110. 



It will be noted that the average for 1892 is only based upon the catches of 8 

 vessels and that their average is somewhat less than the average in 1893. But -the 

 latter circumstance is easily explained by the fact that it was the first serious attempt 

 at pelagic sealing in Japanese waters. When we consider that the sealers were almost 

 entirely inexperienced it is simply remarkable that they could average 1,508 skins per 

 vessel, and it becomes quite certain that if they had been as well posted that year on 

 the movements of the seals as they became afterwards the average would have been 

 considerably higher than in 1893. 



Apart from this apparent increase in 1893 we find a large and steady decrease in 

 the average catch per vessel, in spite of the fact that the sealers became more and 

 more experienced both as to the movements of the seals, the handling of the schooners, 

 and the use of the guns. 



I wish to call attention to and emphasize the exceedingly large decrease as early 

 as 1894, which, on the other hand, was the most successful sealing year so far as the 

 total number of seals are concerned, showing as it does that the very first year of 

 wholesale killing of pelagic seals, viz, 1893, was followed by a startling decrease in the 

 average catch, which has continued ever since. 



ANALYSIS OF PELAGIC SEAL CATCHES OFF THE COAST OF JAPAN. 



In order to fully understand and illustrate various phases of the pelagic sealing I 

 have spent a great deal of labor and care in constructing a series of tables and 

 diagrams (pis. 110-112) which are intended to illustrate graphically the progress of 

 the catch audits surrounding conditions for each week of the spring season. 



These tables and diagrams are based upon the log entries of 39 schooners, which 

 are found rendered in detail at the end of the chapter on pelagic sealing. These 39 

 logs are all which I have been able to collect, but the number is large enough to secure 

 permanent averages from which reliable conclusions can be drawn. 



The diagram showing the total number of seals taken by the 39 schooners whose 

 logs we possess (pi. 112, fig. a) demonstrates several interesting facts: First, that the 

 great bulk of the skins were taken between the middle of March and the middle of 

 June; second, that the season of 1893 was from one to two weeks behind that of the 

 other years, and third, we observe that, whether the season be early or late, there 

 are four distinct maxima, the first being normally about April 1, the second about 

 April 20, the third near the end of the first week of May, and the fourth toward the 

 end of May and the beginning of June. 



Now, turning to diagram h, on the same plate, giving the total number of days 

 during which the same 39 schooners have been hunting, we are struck by the general 

 similarity between these lines and those of diagram' a. We note here also four distinct 

 maxima, which in time correspond pretty well with the maxima of catches, and that 

 the hunting day maxima of 1803 are one to two weeks behind those of the years 

 1894-1896. This correspondence between the lines in the two diagrams is evidently 

 more than a coincidence, and as the number of hunting days per week is chiefly due 

 to the weather it is tolerably clear that the cause of the lateness of the seal catch 

 and the seals in 1893 was due chiefly to some backwardness of the meteorologic or 

 hydrographic conditions which influence seal lite. 

 On plate 1 11 is given a diagram showing the average catch per hunting day for 



