SALMON RUNS OF THE COLUMBIA RIVER IN 1938 



133 



greatly increased on account of the closed season from August 25 to September 10. 

 The explanation is obvious. A very large part of the total catch of chinooks in the 

 river below Bonneville is made by means of gill nets, and this type of gear is selec- 

 tive — taking more of the larger fish and permitting most of the smaller ones to pass 

 through. During the closed period this selection is not operating, and both large 



o 

 z 



Figure 5.— Percentage of grilse in the Bonneville count for 1938, by weeks. 



and small fish pass through the lower river and arrive at Bonneville with their pro- 

 portions practically unmodified. 



The grilse passing Bonneville during the 2 weeks ending September 10 and 17 

 (the weeks in which the run has been least affected by the intensive fishery in the 

 lower river) form approximately 10 percent of the total count, so that it seems prob- 

 able that this figure is not far from the correct one for the fall run as a whole. 

 This is approximately half of the percentage of grilse found both earlier and later in 

 the season — a fact which supports a previous conclusion based on quite different 



449668 — 42 5 



