In each of the five subsequent years, a field crew 

 sampled each location at weekly intervals, from late 

 April to October. 



Juvenile coho salmon were collected with beach 

 seines and minnow traps, and chars with beach 

 seines, floating gill nets, and hook and line. Only 

 coho salmon larger than 80 mm were found to be 

 predators of sockeye salmon during the April- 

 October sampling period. Since char were captured 

 with hook and line or in gill nets, our samples con- 

 tained only fish that were large enough to consume 

 sockeye salmon fry. Coho salmon were preserved 

 in formalin for examination later, when the contents 

 of the stomachs were removed and the sockeye 

 salmon fry and fingerlings were counted. The chars 

 were tagged and released after the contents were 

 flushed from their stomachs and preserved in 

 formalin. 



Results 



Juvenile coho salmon were aggregated around 

 tributary outlets and in the littoral areas of Karluk 

 Lake. The chars were found almost exclusively 

 around the tributary outlets and in the Karluk River. 

 Few coho salmon or chars were captured by sein- 

 ing in the pelagic areas of the lake. We did not 

 distinguish between the two chars. 



Predation rates for the chars did not increase as 

 the abundance of sockeye salmon increased (Table 

 1), but the average number of sockeye salmon fry 

 consumed by each coho salmon did increase and ap- 



peared to be depensatory (Fig. 3). A general equa- 

 tion (Real 1979) was used to describe the relation 

 between the predation rate (Y) by coho salmon and 

 the index of prey abundance (X), 



Y = bX'-lil + aJC-). 



(1) 



A value of c = 1 provides a classic. Type II func- 

 tional response curve, while values of c exceeding 

 1.0 provide a sigmoid, depensatory shape. A value 

 of c = 2 gives the classic Type III curve. 



Best (least squares) fit values for a, b, and c were 

 obtained by transforming Equation (1) into the form 



ln(l/y - a/b) = In(i-i) - c\n(X). 



(2) 



Using trial values of a/b, we regressed the left side 

 of Equation (2) on \nX until we identified the value 

 of a/b giving the lowest residual variance. The 



Table 1 . — Predation on juvenile sockeye salmon by predatory coho 

 salmon (i.e., juvenile coho salmon longer than 80 mm) and chars 

 in Karluk Lake, AK during June and July from 1982 to 1986. 



o 

 ■a 



5n 



0.2 0.4 0.6 0.8 



Index of Prey Density (Millions of Spowners) 



1.0 



Figure 3.— Functional response curve for coho salmon greater than 80 mm 

 (predators) and sockeye salmon (prey) in Karluk Lake. The index of prey abun- 

 dance was divided by 1,000,000 (e.g. index 0.8 = 800,000 adults in the 

 escapement). 



614 



