that stocks of sockeye salmon bound for these river 

 systems are still abundant in the offshore waters. 

 Although the differences are slight, the recovery data 

 indicate that Nushagak Bay stocks are more abun- 

 dant offshore (station 13) than inshore (station 12). 



Tagging stations 15 and 16 are closer to the head of 

 Bristol Bay (Fig. 12), and Ugashik and Egegik 

 recoveries from station 15 (nearer shore) were sub- 

 stantially greater than expected (Table 1). Ugashik 

 and Egegik recoveries from station 16 (offshore), on 

 the other hand, were almost lacking. This is the same 

 pattern that resulted from taggings at stations 12 and 

 13 farther seaward. Likewise the returns of Naknek- 

 Kvichak and Nushagak stocks tagged at stations 15 

 and 16 are like those of taggings at stations 12 and 13, 

 i.e., these stocks are still mostly offshore (station 16) 

 rather than inshore (station 15). 



Finally, comparisons of the tag recoveries from the 

 east districts (Naknek-Kvichak, Egegik, and Ugashik 

 combined) with those from the west district 

 (Nushagak) further suggest that segregation of stocks 

 is taking place as the fish progress toward the inner 

 bay. The recoveries from nearshore taggings (stations 

 12 and 15) in the combined three east districts and the 

 single west district were near the expected number 

 (Table 1). Recoveries from the offshore taggings 

 (stations 13 and 16) were greater than the number ex- 

 pected in the Nushagak district (west side) and less 

 than the number expected in the combined east side 

 districts (Table 1). These differences, however, were 

 not great enough to result in rejection of the 

 hypothesis of like distributions for sockeye salmon 

 stocks destined for fishing districts on each side of 

 Bristol Bay. 



All the tagging in 1965 was done within Bristol Bay 

 in three general areas: (1) lat. 57°N and long. 160°W 

 (most seaward); (2) lat. 57°48'N and long. 158°30'W 

 (nearest the head of the bay); and (3) opposite Port 

 Heiden — actually seven closely adjacent locations 

 midway between areas one and two (Fig. 12, Table 1). 



Again, the results of chi-square tests of the recovery 

 distributions from the taggings at lat. 57°N and long. 

 160°W and from the taggings opposite Port Heiden 

 (Table 2) did not result in rejection of the hypothesis 

 of like distributions of all stocks. Recoveries in the 

 Nushagak district were less than expected, but 

 recoveries in the Naknek-Kvichak district occurred in 

 about the numbers expected. Recoveries in the Egegik 

 district were about as expected from tagging at lat. 

 57°N and long. 160°W but were considerably more 

 abundant than expected from the seven grouped 

 taggings. Differences from the expected number of 

 recoveries in the Ugashik district were not great. A 

 comparison between east and west side fishing dis- 

 tricts shows that recoveries from the west side 

 (Nushagak Bay) were fewer than expected. These 

 results and those from the 1964 taggings farther 

 offshore (stations 13 and 16 in Table 1) indicate that 

 sockeye salmon stocks bound for the Nushagak dis- 

 trict (west side) were more abundant at the offshore 



tagging sites toward the west side of the bay than at 

 the inshore sites toward the east side. 



The distribution of the recoveries from fish tagged 

 at lat. 57°48'N and long. 158°30'W (near Cape Greig, 

 the most easterly of the offshore tagging sites) differs 

 from the distribution of recoveries from fish tagged 

 more seaward in 1965 (Table 1). Significant chi- 

 squares were obtained for comparisons both between 

 individual fishing districts and between east and west 

 side districts (Table 2), and are cause for rejection of 

 the hypothesis of like distribution for all stocks in the 

 area of tagging. The high chi-square values are due to 

 the much greater-than-expected number of recoveries 

 from the Nushagak Bay district and the substantially 

 smaller-than-expected number of recoveries from the 

 Naknek-Kvichak district. This distribution of 

 recoveries suggests that stock segregation is occurring 

 toward the head of the bay, and appears consistent 

 with the results farther offshore. The Cape Greig tag- 

 ging site (lat. 57°48'N and long. 158°30'W) is offshore 

 and comparable to sites 13 and 16 — Fig. 12). My inter- 

 pretation of this pattern is that by the time migrating 

 salmon have progressed this far up the bay, most of 

 the Egegik and Ugashik fish have gone inshore, as 

 have many fish of the Naknek-Kvichak stocks, and 

 many of the remaining offshore fish are bound for the 

 Nushagak district. 



Synopsis: Segregation of Sockeye Salmon Stocks 

 in the Eastern Bering Sea and Outer Bristol 

 Bay. — The distribution of recoveries of sockeye 

 salmon tagged in the offshore area between long. 165° 

 and 170°W and south of lat. 57°N (Fig. 12) were con- 

 sistent in showing that fish bound for the four major 

 fishing districts were present in the tagging area in 

 proportion to their abundance in the total run to 

 Bristol Bay. In only one instance (1958) was the 

 hypothesis of like offshore distributions for all stocks 

 rejected. In most cases, the results also showed that 

 stocks of sockeye salmon bound for the Nushagak 

 fishing district occurred at tagging locations in 

 numbers less than expected. Although exploratory 

 fishing has shown substantial numbers of sockeye 

 salmon occur north of lat. 57°N, only limited tagging 

 was carried out in this area and too few tags were 

 recovered to make an analysis. 



The recoveries from the tagging at sites inside 

 Bristol Bay (between long. 158° and 161°W— Fig. 12) 

 showed that Nushagak district stocks were present in 

 expected or greater-than-expected numbers at the 

 northernmost tagging sites. Sockeye salmon bound for 

 the Nushagak district may also be more abundant 

 than expected in the eastern Bering Sea. Apparently, 

 the more northerly portion of the area encompasses 

 their migration route, but data to substantiate this 

 point are lacking. For the present, the stock composi- 

 tion of sockeye salmon in the northern portion of the 

 approach to Bristol Bay remains open to question. 



The recoveries of sockeye salmon tagged between 

 long. 158° and 161 °W showed an increase in the 



11 



