recruitment, which usually is measured as the 

 number of adults returning, and the parent es- 

 capement. A simplified example of this relation, 

 developed partly from inference and partly from 

 observation, is a curve that starts at zero and 

 rises to the left of the replacement line (the 45° 

 line representing replacement reproduction at 

 various levels of escapement) , so that over a con- 

 siderable range of escapements the expected 

 return is greater than the parent escapement 

 (fig. 1) . The distance between the curve and the 

 replacement line represents surplus production. 

 The curve, however, either levels off or declines 

 because the capacities of the spawning area, the 

 nursery area, or the ocean feeding ground have 

 been exceeded; and an increased number of 

 spawners will not increase the expected return. 



In practice, a satisfactory curve is difficult to 

 obtain. First, observations over many years that 

 include values for the entire range of escape- 

 ment sizes are needed for each run. These ob- 

 servations are not available for most runs in 

 southwestern Alaska. Attempts to estimate the 

 escapement on the basis of the catch per unit 

 of effort of the fishery (International North 

 Pacific Fisheries Commission, 1962) yielded un- 

 certain results because the relation between the 

 catch per unit of effort and the level of the 

 escapement varied with changing gear, chang- 

 ing management practices, and changing effects 

 of competition between units of gear. 



Data on the catches of sockeye salmon from 

 runs to Bristol Bay since 1955 present a special 



an 



3 



LiJ 



Figure 1. — Theoretical retui-n-escapement relation of 

 sockeye salmon. 



problem. The Japanese fishing fleet has inter- 

 cepted parts of these runs at sea, and its catches 

 in some years have nearly equaled those of 

 United States fishermen in Bristol Bay. The total 

 Japanese catch of Bristol Bay sockeye salmon 

 is not known with reasonable accuracy, and we 

 have no breakdown of their catch by river sys- 

 tem. 



Even when accurate catch and escapement 

 data have been available, the relation between 

 return and escapement has not been clearly de- 

 fined — partly because of the cyclic nature of 

 many sockeye salmon runs and partly because 

 survival rates are greatly influenced by environ- 

 mental conditions which differ from year to 

 year. Obviously, it is desirable to measure abun- 

 dance before the adults enter the fishery, partic- 

 ularly after completion of any critical part of 

 the life cycle or residence in an environment that 

 may limit production. The mortality of sockeye 

 salmon in fresh water, from potential egg depo- 

 sition to seaward-migrating smolt, is high and 

 extremely variable ; considerable evidence indi- 

 cates that the fresh water rather than the ma- 

 rine environment usually places upper limits on 

 the numbers of returning adults. 



Thus, an important question is : What factors 

 in the fresh-water environment are limiting the 

 size of the sockeye salmon population in each 

 system? First, the adult salmon must have ac- 

 cess to spawning areas, and in some systems 

 lack of access may be the limiting factor. Next, 

 the spawning area must have gravel and water 

 suitable for spawning and incubation of the 

 eggs and for sheltering the fry until they 

 emerge. Then, the young fish must find food and 

 shelter for 1 or 2 years in the lake while they 

 grow from a length of about 2.5 cm. to 7 to 13 

 cm. — smolt size. Finally, the smolts must be able 

 to migrate successfully to sea. 



In the sockeye salmon systems of southwest- 

 ern Alaska, neither the adult's access to the 

 spawning area nor the smolt's path of migration 

 to the sea is restricted. It seems most logical, 

 therefore, to seek limiting factors either during 

 the salmon's life on the spawning grounds as 

 adult, egg, larva, or fry or during its life in the 

 lake as a growing juvenile. Our studies were 

 planned on this assumption, and our task has 

 been to determine approximately the number of 

 spawners that can be expected to produce 

 enough young to fully use the environment. 



406 



U.S. FISH AND WILDLIFE SERVICE 



