Several species prey on juvenile sockeye salm- 

 on at different stages in the Naknek system. 

 Dolly Varden are abundant and in several areas 

 feed actively on sockeye salmon fry during their 

 emergence and migration. Juvenile coho .salmon, 

 present throughout the system, have been ob- 

 served feeding on sockeye salmon fry in Brooks 

 Lake. Presmolt sockeye salmon migrate out of 

 Coville Lake in late summer day and night; at 

 that time predation by lake trout, Arctic Terns, 

 and Bonaparte's Gulls is very obvious. 



Dolly Varden and rainbow trout prey on mi- 

 grating smolts in the Naknek River. Other pos- 

 sible fish predators in the Naknek system are 

 Arctic char, northern pike, Arctic grayling 

 (Thumalliis arcticus), and sculpins {Cottus 

 spp.). 



Dolly Varden and juvenile coho salmon both 

 prey on juvenile sockeye salmon in the Chignik 

 system (Roos, 1960), but the predation by coho 

 salmon seems to be the more significant (Roos, 

 1959). 



The Karluk system has large numbers of Dolly 

 Varden and Arctic char (DeLacy and Moi'ton, 

 1943) , which may be sei'ious predators of juve- 

 nile sockeye salmon. 



The effectiveness of predators in keeping com- 

 petitors of sockeye salmon in check is unknown 

 but may be important. In the Wood system, for 

 example, threespine sticklebacks are an im- 

 portant item in the diet of Arctic char. 



The destruction of a constant number of juve- 

 nile sockeye salmon by predators each year 

 would help sustain and perhaps reinforce the 

 cyclic nature of some runs to the Bristol Bay 

 systems. A constant annual demand would have 

 a greater effect in years of low abundance of 

 juvenile sockeye salmon. Such an effect is sug- 

 gested for the sockeye salmon of Shuswap Lake 

 of the Fraser system (Ward and Larkin, 1964) . 



REPRODUCTIVE RELATIONS OF 

 POPULATIONS IN FRESH 

 AND SALT WATER 



The relations between numbers of smolts pro- 

 duced and numbers of adults in the parent 

 escapements (described by the reproductive 

 curve) are highly variable, even within a system, 

 because they are determined by the interaction 

 of many physical and biological factors. In this 

 section we consider the relations between num- 

 bers of adults in the parent escapement and their 



progeny in terms of numbers, size, and age of 

 smolts migrating to sea, and numbers of adults 

 returning from the sea (catch and escapement) . 



RELATION BETWEEN NUMBERS OF SMOLTS .4ND 

 NUMBERS OF ADULTS IN PARENT ESCAPEMENT 



The major purpose of studies of the numbers 

 of smolts produced by various munbers of 

 spawners is to define the relation and interpret 

 it to make estimates of optimum escapement (in 

 terms of numbers of smolts) either directly or 

 with the aid of ancillary information. 



Smolt migrations were sampled with fyke nets 

 in the trunk rivers. The abundance is expressed 

 in index points for the Wood and Kvichak Rivers 

 and in numbers of fish for the Naknek, Ugashik, 

 and Karluk Rivers. The methods of counting the 

 escapement of adults were described earlier in 

 this report. 



The index method of estimating production of 

 smolts requires the use of standard fishing gear, 

 site, and effort. The numbers of smolts captured 

 are then given value in relation to the catches of 

 a base year which has an arbitrary value of 100 

 index points. This method is used in the Kvichak 

 .system (Kerns, 1961) and the Wood system 

 (Burgner, 1962). It yields data comparable be- 

 tween years within a system but not between 

 systems. 



The estimates of numbers of smolts produced 

 by a system were based on a regular system of 

 sampling. Many fishing sites were set up across 

 the width of the river, and fyke nets were fished 

 for a fixed period of time once a day in each site 

 in random order. The sequence of fishing sites 

 and times of day was a modified Latin-square 

 design (Cochran and Cox, 1957). Catches were 

 classified according to site, time of day, and date 

 — factors considered to have the greatest influ- 

 ence on variability of the individual catches. 

 E.stimates of the total numbers of smolts in the 

 annual migrations were obtained by expanding 

 the mean catch per standard fishing period ac- 

 cording to the number of potential fi.shing sites 

 in the width of the river and the number of fish- 

 ing periods in the season. We used this method 

 to estimate the numbers of smolts migrating in 

 the Naknek and Ugashik systems. 



The age composition of the smolts in each mi- 

 gration was estimated from length-frequency 

 measurements or scale samples taken during the 

 season and weighted by magnitude of fyke net 



SOCKEYE SALMON IN MAJOR RIVER SYSTEMS IN SOUTHWESTERN ALASKA 



43; i 



