Consequently, stream flow is not considered further 

 in our analysis. 



It was immediately apparent, on the other hmnl, 

 from our first observations in 1002 that the water 

 temperature of the Okanogan River greatly influ- 

 ences the migration of sockeye salmon bound for 

 Lake O.soyoos. To understand the effects of tem- 

 perature better, we have assembled migration-route 

 water temperatures dating i)ack to 1937. Tem- 

 jjerature was originally recorded l>y thermographs 

 or by hand-hckl thermometers. Daily averages 

 have been computed from the highs and lows on the 

 thermograph charts or from the 8 a.m. and 4 i).m. 

 thermometer readings. 



LIMITATIONS AND ADJUSTMENTS OF TUF. DATA 



Before proceeding, it is appropriate to review 

 certain limitations and adjustments of (he data that 

 are potential sources of error. 



Reliability of the Fish Counts and 

 Tag Observations at Zosel Dam 



After 1948, varying proixutions of the Okanogan 

 run passed Zosel Dam by means other than the 

 fishways, and the reliability of the counts and tag 

 observations recorded at Zosel Dam since 1948 

 varies accordingly. We used two criteria to deter- 

 mine which data were adetiuate for this study. 

 First, b}^ comparing the covmts at Zo.-<el Dam with 

 estimates of numbers of fish on the spawning 

 grounds, we calculated the percentage trapping 

 efficiency for 19.52-57 and 1962 — years for which 

 both fish counts and spawning ground estimates 

 were available. The.se percentages, however crude, 

 are our best estimates of trapping efficiency at Zosel 

 Dam. Second, we determined fiom the detail(>d 

 notes maintained by the fish coimters at Zosel Dam 

 whether the modal counts correspontled to the 

 modal numbers estimated to be passing the dam. 



Estimates of trapping eflTiciency are shown in 

 table 2. The overall accuracy of the fish counts in 

 1955-57 was severely limited. Equally important, 

 the modal counts for those years did not correspond 

 to the modal luimbers estimated to be passing the 

 dam. We have accordingly deleted the 1955-57 

 data from our analysis. Although trapping effici- 

 ency was not much higher in 1952, 1954, and 19(12, 

 the modal counts agreed closely with the numbers 

 estimated by the counters to be available below the 

 dam. The data for 19.52, 19.54, and 19(i2 have been 

 retained, therefore, and, together with the data of 



high-efliciency years, 1937, 1944, 1953, and 1963, 

 constitute the l)asis of our evaluation of the effect 

 of water temperature on the migration of sockeye 

 between Rock Island and Zosel Dams. Although 

 no spawning ground estimate was made in 1963, our 

 regular observations indicated a relatively high 

 trapping efficiency at Zosel Dam for that j-ear. A 

 fish-tight weir enabled the counters to make complete 

 fish counts in 1935-37 and 1944. Although the 

 counts of fish were comi)lete in 1935 and 1936, we 

 cannot use them here because there are no water 

 temperature records for those years. 



Table 2. — I nfonnalion on the efficiency of the trajiping system 

 at Zosel Dam 



' Tufts, Dennis F . ami Donovan R. Craddock. 19G3. Spawnins 

 esrapenient of Coluiuhia River sockeye salmon (O. nerka). 1902. U.S. Bur. 

 of Conini. Fish Biol. I.ab.. Seattle, Wasli., 17 pp , Jan. 1963 (Processed). 



Adjustment of the Tagging Data 



French and Wahle (1900) estimated that, before 

 the construction of Rocky Reach Dam, 10.7 days 

 were required for sockeye salmon to migrate from 

 the forebay of Rock Island Dam to Zosel Dam. 

 Their estimate was based on a small sample of 30 

 tagged .sockeye salmon that had been released just 

 above Rock Island Dam in 19.54 and 1955 and later 

 observed at Zosel Dam. 



To estimate migration time, we have used the 

 large numliers of tagged sockeye salmon (334 and 

 215) that were released just brlow Rock Island Dam 

 in 19.53 and 1951, and later observed at Zosel Dam. 

 Tagged .sockeye salmon were relea.sed above Rock 

 Island Dam in 1962 and 1963. An adjustment was 

 oln iously necessary before the travel time to Zosel 

 Dam of tagged fish that had been released lielow 

 Rock Island Dam could be compared with the travel 

 time of tagged fish that had been released above 

 Pvock Island Dam. We adjusted the data according 

 to the day or days when the number of tagged fish 

 which had been released below Rock Island Dam 

 peaked at the counting stations of Rock Island- Dam. 

 The dates of these i)eaks were treatetl as dates of 

 release in the forebay. For examjile, sockeye that 



136 



U.S. FISH ASD WII.OI.IFIC SKRVHi; 



