indicate that the population level is controlled by 

 the capacity of the nursery areas in each lake. 



The production of juvenile sockeye salmon in 

 the lakes of the Wood system increases linearly 

 as the escapement increases, up to a density of 

 5,000 spawners per square kilometer of lake. 

 Above this density the relation is no longer lin- 

 ear, and the production of juveniles per spawner 

 declines when escapements exceed 6,000 per 

 square kilometer. To have 5,000 spawners per 

 square kilometer in all lakes of the system would 

 require an escapement of 2,125,000 fish. At this 

 density, however, the growth of juvenile salmon 

 is considerably reduced. The mean length of age 

 fish in September is inversely related to the 

 size of the parent escapement in each of the lakes 

 of the Wood system. The mean length of age I 

 smolts leaving the system also is inversely re- 

 lated to numbers in the parent escapement. 



The combined effect of large populations of 

 juveniles is to increase total mortality (fresh- 

 water and ocean) by causing many juveniles to 

 remain an extra year in fresh water and to be 

 of smaller average size for their age when they 

 reach the ocean. This increase in mortality tends 

 to make the population self limiting. 



Threespine sticklebacks and Arctic char are 

 the species of fish thought to have the most im- 

 portant effects on juvenile sockeye salmon in the 

 Igushik, Snake, and Wood systems. In the Wood 

 system, threespine sticklebacks and juvenile 

 salmon use the same food supply, and, therefore, 

 the growth of sticklebacks and age I salmon is 

 reduced when large numbers of age salmon are 

 present. In all three systems Arctic char is the 

 chief predator on small fish and probably eats 

 substantial numbers of sticklebacks and juvenile 

 salmon. The role of sticklebacks as a competitor 

 for food and as a bufi'er between Arctic char and 

 sockeye salmon and the effect of the Arctic char 

 in keeping stickleback populations in check are 

 not understood. 



The recent average escapement of sockeye 

 salmon per unit area of lake has been higher in 

 the Igushik than in the Wood system. Juveniles 

 from the larger escapements in the Igushik sys- 

 tem have not grown as well (the smolts have not 

 been as long) as those from the smaller escape- 

 ments, so that here also the food supply ap- 

 parently limits growth. Preliminary limnological 

 studies show no marked difference in richness 

 between the waters of the Igushik and Wood 



Lakes, and the greater concentration of spawn- 

 ers in the Igushik system appears to be caused by 

 differences in the intensity of fishing on the two 

 stocks. 



Although primary productivity is low in the 

 Nuyakuk system, escapement densities are prob- 

 ably too low to produce enough juvenile salmon 

 to tax the food supply. Spawning areas appear 

 to be far in excess of the present need. Relatively 

 heavy fishing pressure probably has contributed 

 to the present low level of the runs. Lake trout 

 are more plentiful than Arctic char in the Nuya- 

 kuk system ; the lake trout may be the primary 

 predator on young salmon and sticklebacks. 



The return-escapement relations of sockeye 

 .salmon in the Nushagak district in 1946-62 indi- 

 cate that the optimum escapement is about 1 mil- 

 lion fish, but that escapements between 600,000 

 and 1.2 million will produce near-optimum 

 results. 



If spawners were properly distributed, pro- 

 duction in the Nushagak fishery could probably 

 be increased by increasing ^he escapements. In 

 an escapement of 1 million, for instance, if 70 

 percent of the fish went to the Wood system the 

 lakes of the system would have an average 

 density of 1,650 spawners per square kilometer. 

 If distribution of spawners among the lakes were 

 even, this density could be profitably exceeded. 

 In years of uneven distribution of spawners, es- 

 capements should be held at a lower level to avoid 

 overstocking of some rearing areas. Further, the 

 mesh size of the gill nets used in the fishery 

 could be regulated so as to obtain escapements of 

 the two primary ocean-age groups in proportions 

 that would more fully use all of the lakes of the 

 Wood system and would give more protection to 

 the populations of the Nuyakuk system. All the 

 consequences of changes in the size of the mesh 

 must be carefully studied, however, before this 

 change is initiated. 



Kvichak System 



The Kvichak, the largest sockeye salmon sys- 

 tem in Alaska, consists principally of two lakes, 

 Iliamna and Clark, which total 2,889 km.= The 

 sizes of the runs in this system are cyclic. The 

 range of the escapement in 1955-63 was 250,000 

 to 14.6 million. In 1931-61 the annual catch of 

 the Naknek-Kvichak fishery averaged about 7.5 

 million fish. During the years of very large re- 

 turns the Bristol Bay catch is dominated by 

 salmon returning to the Kvichak system. 



SOCKEYE SALMON IN MAJOR RIVER SYSTEMS IN SOUTHWESTERN ALASKA 



451 



