15 30 45 60 75 90 



NUMBER OF FlSH/50 FATHOMS OF NET / I 2 HR FISHING 



Figure 9. — Gill net catches of sockeye salmon along approxi- 

 mately long. 161°30 W, 24 June-6 July 1965 and 1966. (See Ap- 

 pendix Table 1 for source of data.) 



that shown by catches along the transect between 

 Cape Seniavin and Cape Newenham in 1939 (Fig. 10). 

 Here, the distribution remains essentially the same 

 from early to late July. The greatest abundance of 

 sockeye salmon along both transects was offshore but 

 in the southern half of Bristol Bay. The pattern of 

 abundance of sockeye salmon in gill nets fished by the 

 Japanese training ship Oshoro Maru in 1967 and 1969 

 (Hokkaido University, The Faculty of Fisheries 1968, 

 1970) in the western part of Bristol Bay substantiated 

 their high abundance in this region during the spawn- 

 ing migration. 



The abundance of sockeye salmon along all 

 transects was greatest in offshore waters of the eastern 

 Bering Sea and in Bristol Bay itself. 



20 40 60 60 too 20 40 60 80 100 



(JUNE 27- JULY 7) (JULY 10-14) 



NUMBER OF FISH/(,030 FATHOMS OF NET 



20 40 60 80 100 120 140 

 (JULY (7-23) 



Figure 10. — Gill net catches of sockeye salmon between Cape 

 Seniavin and Cape Newenham, 27 June-23 July 1939. Note: 

 Fishing time was not specified, but I assumed it was the same 

 for each set. (See Appendix Table 1 for source of data.) 



Synopsis: Distribution of Sockeye Salmon in the 

 Eastern Bering Sea and Outer Bristol Bay during 

 Spawning Migration. — From the foregoing results, it 

 is clear that sockeye salmon bound for Bristol Bay do 

 not follow a route close inshore, at least not westward 

 of Cape Seniavin. After entering the Bering Sea via 

 Aleutian Islands passes, the salmon apparently move 

 north for a considerable distance before moving east 

 into Bristol Bay. This northward movement into the 

 offshore waters of the Bering Sea was reported by 

 Hartt (1966). Early tagging experiments in the vicini- 

 ty of Cape Seniavin by Gilbert (1923) showed that 

 even here Bristol Bay sockeye salmon were not pres- 

 ent in the coastal waters. These results were con- 

 firmed through additional tagging studies carried out 

 in the same area in 1925 by Rich (1926). 



The above results have been summarized in Figure 

 11 to provide a synopsis of the distribution of sockeye 

 salmon during migration through the eastern Bering 

 Sea and outer Bristol Bay. 



Offshore Tagging. — To determine if individual 

 sockeye stocks or groups of stocks comprising the run 

 to Bristol Bay were segregated according to river of 

 origin while in the offshore area, I analyzed the 

 published results of 6 yr of salmon tagging carried out 

 at various locations in the eastern Bering Sea and out- 

 er Bristol Bay (Fig. 12, Appendix Table 1). These 

 years were selected because the tagging sites were 

 widely distributed over the approaches to Bristol Bay; 

 many of the sites of tagging were east of long. 170° W, 

 and tagged fish from most sites were recaptured in the 

 Bristol Bay fishery. 



Tag recoveries are grouped by regulatory fishing 

 district, i.e., Nushagak, Naknek-Kvichak, Egegik, 

 and Ugashik. The sockeye salmon migrating to each 

 district are treated as individual populations, even 

 though several major river systems discharge into 

 some of the districts. The assumption is made in the 

 analysis that tagged fish recaptured in a particular 

 fishing district were actually bound for the major river 

 systems of that district. 



In analyzing the recovery data, I adopted the 

 hypothesis of like distribution for all sockeye salmon 

 stocks of Bristol Bay origin, i.e., that sockeye salmon 

 tagged at the various offshore sites occurred in the 

 same proportion as they did in the total run to Bristol 

 Bay. To test this hypothesis, I used chi-square 

 analysis to compare the actual tag recoveries with the 

 expected recoveries in individual fishing districts and 

 also when the districts were grouped according to 

 location, i.e. whether fishing districts were located on 

 the west side (Nushagak) or east side (Naknek- 

 Kvichak, Egegik, and Ugashik) of Bristol Bay. 

 Probability values of less than P = 0.05 were con- 

 sidered to indicate unlike distribution (segregation) of 

 sockeye salmon stocks offshore. 



The expected number of tag recoveries from in- 

 dividual or grouped fishing districts were computed 

 from the following: 



