sturgeon sport fishermen, who seek salmon car- 

 casses for bait. 



Until our study, only one extensive effort had 

 been made to evaluate the significance of dead 

 fish near the dam. In 1946 FWS (U.S. Fish and 

 Wildlife Service) and the U.S. Army Corps of 

 Engineers investigated the causes of injuries and 

 deaths of fish (Hanson et al., 1950). The major 

 emphasis of that study was on counting, from 

 boats, the floating dead fish near the dam during 

 the main part of the chinook salmon fall migration. 

 The study did not provide the information needed 

 to determine the relation between numbers of 

 floating dead salmon, numbers of dead salmon 

 not observed, and numbers of salmon surviving to 

 continue their upstream spawning migration. 



Sporadic observations in other years by biolo- 

 gists and fishermen also failed to provide informa- 

 tion that could be used to evaluate the significance 

 of floating dead salmon. To illustrate the difficulty 

 of interpreting such information, we cite two ex- 

 amples of observations made at a time when un- 

 usually large numbers of floating dead salmon 

 were in the Columbia River. 



The first observation was on September 9, 

 1943. Arnie J. Suomela, Fishery Biologist with 

 the Washington State Department of Fisheries, 

 spent 8 hours searching from a boat for dead 

 salmon between Bonneville Dam and Multnomah 

 Falls. He found 146 floating chinook salmon, nine 

 steelhead trout, and one sockeye salmon; the 

 following statement is from his report. "Checked 

 Oregon shore and river on trip down to Mult- 

 nomah Falls. Only two salmon were found on this 

 part of trip. The first floating dead fish were found 

 at Butler's Eddy, at 3:05 p.m.; and floaters were 

 found from that point to below the spillway at 

 the dam. This observation definitely traces the 

 mortality to the north spillway channel and it is 

 reasonable to believe that the mortality is occur- 

 ring at the spillway."^ 



The second observation was in the spring of 

 1952. The Oregonian newspaper for June 12, 1952, 

 reported: "A commercial fisherman . . . recently 

 . . . found 'thousands' of dead chinook salmon 

 between Martins Slough [102 km. downstream 

 from Bonneville Dam] and the mouth of the 

 Lewis River. [He] said the beaches were littered 

 with spring chinook. 'You could find more as you 

 get closer to the Dam.' " 



'Unpublished field notes, Washington State Department of Fisheries. 



The inadequacy of the simple observational 

 method in assessing the true magnitude of mortal- 

 ity is exemplified further in a classic study by the 

 International Pacific Salmon Fisheries Commis- 

 sion. At Hell's Gate on the Eraser River, salmon 

 were blocked by turbulent water at certain river- 

 flows, and an annual loss of thousands (even 

 millions) of salmon there has been well docu- 

 mented (Thompson, 1945; Talbot, 1950; Jackson, 

 1950). Although a great mortality was suspected, 

 only a relatively small number of moribund or 

 dead fish were sighted on extensive searches down- 

 stream from Hell's Gate over a period of many 

 years. Thompson (1945: 96) described the situa- 

 tion at Hell's Gate in 1941: "It could be said that 

 numbers of them [sockeye salmon] were observed 

 approaching death, having reached a condition 

 which obviously precluded their passage through 

 any difficult currents; yet simple observation 

 could not prove that death actually occurred nor 

 that the percentage dying was very high. To find 

 even hundreds of fish near death along the riflSes 

 in the river, or in the creeks did not necessarily 

 prove that a great part of the run perished below. 

 Some form of evidence more conclusive was 

 necessary." 



Thus, on neither the Columbia nor Eraser 

 Rivers did simple observations of dead fish pro- 

 vide a basis for estimating the true mortality. 

 For this reason, we devised a different method. 



Our aims were (1) to estimate the mortality of 

 adult chinook salmon near Bonneville Dam during 

 a period when large numbers of salmon were 

 passing the dam and (2) to evaluate factors con- 

 tributing to or associated with these deaths, such 

 as streamflow, temperature, turbidity, commercial 

 fishing, fish passage facilities, and disease. 

 Throughout this paper, data on counts of fish at 

 Bonneville Dam and on flow, turbidity, and 

 temperature of the Columbia River are from U.S. 

 Army Corps of Engineers (1943-56). 



We first examined records of observations of 

 dead fish and counts of chinook salmon through 

 the ladders at the dam to determine when maxi- 

 mum mortality had occurred. Biologists, com- 

 mercial fishermen, and boat moorage operators 

 near the dam generally reported that the number 

 of floating dead salmon was greatest in the spring, 

 coincident with high riverflows and large numbers 

 of chinook salmon in the river. We, therefore, 

 selected the spring period of high riverflows for 



464 



U.S. FISH AND WILDLIFE SERVICE 



