SALMON RUNS OF THE COLUMBIA RIVER IN 1938 



111 



1928). Fairly good data are available as to the aggregate troll catch of chinooks and 

 silvers in Alaska, Oregon, and Washington. The percentage of Columbia River fish 

 in this catch, however, undoubtedly varies greatly during the season. There are no 

 satisfactory data on this latter point. Even though we knew the proportions of 

 Columbia River fish in the catch at different times and in different localities, it would 

 be impossible to allocate these to the seasonal runs of the Columbia and thus, eventually, 

 to determine the element in the troll catch derived from the runs to the Columbia 

 River above Rock Island Dam. Likewise, we have no data on the catch of the sport 

 fishery or on that part of the Indian catch that is not sold. All of these elements 

 increase to some unknown extent the economic importance of the salmon runs with 

 which we are here concerned. 



Table 7. — Estimates and counts of fish passiiig Bonneville, 1938 



[The figures up to and including May 7 are estimates based on partial counts only. Differences between the figures given hero 

 and those in the report by Calkins, Durand, and Rich are due to the fact that this table includes the final figures as given by the 

 Army Engineers, in which minor corrections were made of the figures submitted weekly.] 



1 Grilse, locally designated as "jack" salmon, are precocious males. These are included in the preceding column headed 

 "Chinooks," the figures in which are, therefore, the totals for this species. 



In table 7 are given the counts and estimates of the Dumber of salmon and steelhead 

 passing Bonneville Dam during 1938. Actual counting did not begin until May 7, 

 but estimates could be made from partial counts — the so-called "spot" counts — 

 covering the period from the middle of February to and including May 6. These 

 partial counts were made by observers stationed for portions of the day at the several 

 fish ladders. The records consisted of (1) the length of time during which the obser- 

 vations were continued, and (2) the number of fish of each species observed. This is 

 essentially a sampling method, and it is known that the fish do not run uniformly 

 during the entire 2-1 hours, or even during the daylight hours. A fairly good estimate 

 can be made from such records, however, if the hours during which the fish run are 

 determined with care, and if the periods during which the counts are made are suitably 

 distributed. The method adopted here for estimating the total number for the day 

 from the partial counts is to multiply by 12 the average hourly count as determined 

 from the records. This is the method recommended and used by Fred Morton, who 



