FISHERY BULLETIN: VOL. 71, NO. 3 



A measure of the small importance attached to the 

 Karluk [River] spawners is that all reports including 

 Gilbert and Rich (1927) do not include the Karluk 

 River on their maps of the watershed. 



To this one might reply that neither do the 

 available manuscript reports of the junior author 

 (Bevan and Walker, 1955; Bevan, 1957) or the 

 unpublished manuscript reports of W. F. Thomp- 

 son (Thompson, 1951; Thompson, Bevan, and 

 Thorsteinson, 1954; Thompson and Bevan, 

 1954) include the river on their maps. The map 

 in my report (Rounsefell, 1958) does show the 

 river, just as does that of Van Cleve and Bevan. 

 As a matter of far more importance, I do not 

 find any attempt to measure the number of 

 sockeye spawning in the Karluk River in any 

 of the manuscripts just mentioned. My report, 

 showing that in some years a number of mid- 

 to late-season sockeye salmon spawned in the 

 river below the lake, apparently stimulated the 

 authors into hypothesizing one cause for the 

 decline of the runs. In this connection I should 

 like to mention that Bevan and Walker (1955) 

 tabulate the results of their Karluk sockeye 

 salmon spawning ground observations in 1954. 

 They total 138 observations (101 by Bevan 

 and Walker themselves) between 18 May and 

 29 September. Very surprisingly not a single 

 observation is recorded for the Karluk River 

 despite the fact that they camped at the weir 

 site at the foot of the lake. 



Van Cleve and Bevan attempt to explain the 

 importance of the Karluk River spawners by 

 saying that the river spawners were predomi- 

 nately of age 53 and that their decline in numbers 

 could explain the increase in relative numbers 

 of 4-yr smolts in the total runs. An examination 

 of the basic data on which they base their con- 

 clusions (Barnaby, 1944) finds (Figure 1) that 

 for the brood years 1922 and 1924 through 

 1929 that the rate of decrease of the 53 age 

 group is similar in both the spring and summer- 

 fall groups of spawners. The latter group of 

 spawners, with the slightly greater decrease in 

 the 53 age group shows a somewhat higher rate 

 of return, quite the contrary of the theory ad- 

 vanced by Van Cleve and Bevan concerning 

 the superiority of the 53 spawners. 



The assumption of two seasonal modes in 

 the Karluk run (or merely one mode depleted 



80n 



70- 



60- 



50- 



40' 



r8 



-7 



-6 



-5 



Figure 1. — Relation of age fa fish to brood years and re- 

 turn date (Barnaby, 1944). Showing the geometric mean 

 (left hand scale) of the percent 53 age fish in the spring 

 and fall runs (solid lines). Dots are for the spawners, open 

 circles for the returns. Geometric mean of number 

 (-00000) of spawners (dot-dash lines) in broo.d years (open 

 circles) and geometric mean of number in returns (dots). 



in the center as these authors claim) is not 

 borne out by the weekly data on runs accumu- 

 lated since 1921. Thus in the 30-yr period from 

 1921 through 1950 (Rounsefell, 1958; see also 

 Figure 2) three modes are evident. The first 

 mode, peaking in mid-June, falls off rapidly 

 with a low point in the week ending 12 July. 

 The second mode peaks from the first to the 

 ninth of August. The second low point is not 

 as distinct as the first because of overlapping 

 between the second and third modes but it is 

 about August 16th. The third mode peaks in 



10 24 7 



Moy Jun Jul 



Figure 2. — Percentage seasonal occurrence of the Karluk 

 sockeye salmon runs 1921-50 (open circles) and percentage 

 seasonal occurrence of age 53 fish 1922 and 1924-49 (dots). 



652 



