Fish released below Rock Island Dam on July 20-23 

 peaked at Rock Island Dam on July 25, and at 

 Zosel Dam on August 2-3, 8 to 9 days later. From 

 these records we infer that most of the tagged fish 

 reached the Okanogan River on July 29, when the 

 temperature was just beginning to increase, and that 

 this increase did not suppress entry to the stream. 

 Similarly, the collective releases of July 27-30 

 peaked at Rock Island Dam on July 31 to August 1 

 and at Zosel Dam on August 8. On the assumption 

 of 4 days for travel, we estimate that most of these 

 fish arrived at the mouth of the Okanogan on 

 August 5-6 when the temperature began a new rise. 

 Again, however, migration was not affected. Final- 

 ly, the August 3-6 releases of tagged fish peaked at 

 Rock Island Dam on August 7-8 and probaljly 

 arrived at the mouth of the Okanogan River on 

 August 11-12, when temperatures were fairly 

 stable. These fish also appear to have migrated 

 freely through the Okanogan River. A possible 

 explanation of the normal progress of the migration 

 despite rising temperatures is given below in our 

 discussion of the 1963 migrations. 



Temperatures of the Okanogan River were not 

 recorded in 1962 until August 1. High air tem- 

 peratures indicate, however, that the water tem- 

 perature almost surely had been rising prior to 

 August 1. Rising temperature would account for 

 the July 26 to August 5 lull in the count at Zosel 

 Dam (figs. 6 and 10). On the basis of counts at 

 Rocky Reach Dam, we would have expected the 

 arrival of sockeye salmon, both tagged and untagged, 

 at Zosel Dam during this interval. The arrival of 

 the fish at the mouth of the Okanogan River coin- 

 cided, however, with rising water temperatures, a 

 condition which apparently blocked their entry. 



The 1963 migration progressed from Rocky Reach 

 Dam to Zosel Dam without major delay (figs. 6 and 

 10). Tag recoveries were orderly in contrast to 

 those of 19()2; peaks at Zosel Dam followed com- 

 parable peaks at Rocky Reach Dam by 6, 6 to 8, 

 6, 7 to 9, 6 to 9, and 6 to 10 days. Thus, migration 

 was normal in 1963, despite generally rising water 

 temperatures. 



The ab.sence of delay in 1954 and 19(53, despite 

 rising water temperatures, focuses attention on the 

 importance of the level at which the temperature is 

 changing. For example, migration was unimpeded 

 by rising temperatures in the 62° to 69° F. range in 

 1963, but was halted by rises in the 75° to 78°, 70° to 



77°, and 74° to 78° F. ranges in 1952.= Furthermore, 

 a temjierature rise in the 73° to 78° F. range at 

 Oroville interrupted the migration in 1937, a j'ear 

 in which occasional tcmperatin-e readings taken at 

 Monse on the lower Okanogan River were even 

 higher than those at Oroville (Chapman, 1941). 

 On the other hand, the 1954 migration was ap- 

 parently' unaffected l\v rises in the 68° to 70° F. 

 range. The dependability of the latter example is 

 subject to some question, however, because the 

 temperature readings were recorded at Oroville, 

 not at Monse. 



These sevei-al examples suggest a threshokl tem- 

 perature of about 70° F., below which migration is 

 not affected, but above which rising or stable tem- 

 peratures inhibit migration — a condition which 

 endures until a sharp drop allows the migration to 

 resume. 



We have not considered here a situation in which 

 fish enter the Okanogan River under favorable con- 

 ditions only to be confronted enroute by sharply 

 rising water temperatures. We have no data on 

 this aspect of the problem, but suspect that the 

 behavior and survival of the fish depend on several 

 factors, including: (1) their location at the time 

 they are confronted by rising temperatures; (2) 

 their ability to acclimate: (3) their size, general 

 health, and stage of maturity; and (4) the level to 

 which the water temperature rises. 



RESULTS OF OTHER STUDIES AND 



THEIR POSSIBLE BEARING ON THE 



PROBLEM IN THE OKANOGAN RIVER 



The environmental factors that control the migra- 

 tions of adult Pacific salmon have long been of 

 practical and theoretical interest to fishery biologists. 

 The literature gives many examples of environ- 

 mental influences that affect different populations 

 in different ways. Rather than present another 

 review of the extensive literature on this broad 

 subject, a matter so capably handled by Hoar (1953) 

 and Allen (1956), we refer here only to the more 

 important papers that deal with the environmental 

 factors that influence the sudden mass movement 

 of migrating salmon. 



Several investigators have found that rainfall and 

 streamflow afi'ect the migration of adult salmon. 

 Pritchard (1936), and Davidson, Vaughan, Hutch- 

 inson, and Pritchard (1943), who studied pink 



■^ These temperatures, recorded at Monse. are not subject to the possible 

 error of estimating temperatures in the lower river from actual readings 

 at C)roville. 



MIGRATION OF SOCKEYE SALMON 



145 



