BRISTOL BAY AND ALASKA PENINSULA SALMON STATISTICS 65 



between all possible pairs of districts is presented in Table 4. It is apparent from a 

 consideration of those values that there is a high degree of association between the 

 catches in the three districts on the eastern shore of Bristol Bay at Kvichak, Egegik, 

 and Ugashik, but the correlation between any of these districts and Nnshagak is 

 distinctly lower; in fact, so low that the correlation of Nushagak with Kvichak and 

 Egegik is Avithout significance. The correlation between Nushagak and Ugashik is 

 significant of some slight degree of association but is less than four times its probable 

 error. On the whole, it appears that the fluctuations at Nushagak are independent 

 of those on the eastern side of the bay. The correlation between the catches at 

 Kvichak, Egegik, and Ugashik is so marked, however, as to indicate some causal 

 relationship, and it seems more than probable that it is due to the catching in the 

 Egegik and Ugashik districts of fish bound for the Kvichak district. This possibility 

 was pointed out by Gilbert and O'Malley in their report on the salmon fishery in 

 central and western Alaska.* It may also be due, in part at least, to fish being re- 

 ported as taken in one district when actually they were caught in another and were 

 brought into the district from which reported for canning ; or it is possible that there 

 is enough "straying" from the parent stream to cause the catch in near-by streams to 

 fluctuate together. This last possibility does not seem likely, however, as one would 

 suppose that any such straying might afl'ect the correlation between the catches at 

 Nushagak and the other streams as well as between the other three streams. 



We have mentioned above that an increase in the size of the fluctuations may be 

 an indication that the fishery has been developed to the danger point or that depletion 

 has occurred already. This hypothesis has also been advanced by Gilbert and O'Mal- 

 ley (loc. cit.), who, in discussing the situation in the Kvichak region of Bristol Bay, 

 say: "Other river basins have been watched during the progress of depletion. The 

 sequence of events is always the same. Decreased production is [accompanied] by 

 increase of gear. Fluctuations in the seasons become more pronounced. Good 

 seasons still appear in which nearly maximum packs are made. But the poor seasons 

 become more numerous. When poor seasons appear, no attempt is made to com- 

 pensate by fishing less closely. On the contrary, efforts are redoubled to put up the 

 full pack. The poorer years strike constantly lower levels, until it is apparent to all 

 that serious depletion has occurred." Figure 8 shows, with great clearness, that the 

 amplitude of the fluctuations in all districts of Bristol Bay has been increasing with 

 considerable regularity, thus corroborating the evidence given by the trends of 

 general depletion throughout Bristol Bay. 



Such cyclic fluctuations in the abundance of salmon are extremely interesting 

 biological phenomena, and a knowledge of them is of great practical importance to 

 the industry and to an adequate conservation program. It may not be out of place 

 here, therefore, to speculate briefly upon some of the characteristics of such fluctua- 

 tions. It seems safe to assume that in a state of nature the abundance of any race 

 of salmon would be constant from year to year, except as modified by environmental 

 conditions, and that the level of abundance will be at the maximum capacity of the 

 waters occupied by the race. If under these conditions unusually favorable circum- 



< Special Investigation of Salmon Fishery in Central and Western Alaska. By C. H. Gilbert and Henry O'MsIley. In Alaska 

 Fisheries and Fur Industries in 1919. by Ward T. Bower. Appendix IX. Report, U. S. Commissioner of Fishei Its for 1919 (1921). 

 Bureau of Fisheries Document No. 891, pp. 143-160. Washington. 



