and right pectoral fins. The adipose-left 

 pectoral group was released as unfed fry, and 

 the adipose-right pectoral group was released 

 after a 3- month rearing period. Preliminary 

 tabulations show the following recoveries of 

 marked fish: 33 adipose- right pectoral and 3 

 adipose-left pectoral in the Washington-Oregon 

 ocean troll fishery, 22 adipose-right pectoral 

 and 1 adipose-left pectoral in the Columbia 

 River gill net fishery, and 124 adipose-right 

 pectoral and 6 adipose-left pectoral at hatch- 

 eries. From the combined recoveries the in- 

 dicated survival rates for fingerling and fry 

 releases are in the ratio of 18 to 1 (179 to 

 10). Some qualification of the results of the 

 regression study as well as the marking ex- 

 periment is necessary, however, since in 1948 

 and 1949 when very large releases of fry were 

 made and at the time of the release of the 

 marked fry of the 1956 brood, water tempera- 

 tures in Columbia River were near freezing 

 in comparison with the constant temperature 

 of 46° F. for Spring Creek. Possible effects 

 of such a temperature differential will be dis- 

 cussed in a later section. All studies that have 

 been considered here give evidence favoring 

 fingerling survival under existing conditions, 

 although a precise evaluation of the relative 

 survival rates should await the returns of 

 marked releases of the 1957 and 1958 broods. 



Two graphs have been prepared to observe 

 the possible effects of local straying on the 

 correlations between pounds of fingerlings 

 released and adjusted returns of females, and 

 to study whether or not the use of Spring 

 Creek Hatchery data as a unit is justified 

 (fig. 5). In figure 5a releases and returns at 

 Spring Creek Hatchery are compared only, 

 and in figure 5b combined releases and re- 

 turns at Spring Creek Hatchery and Big White 

 Salmon River are plotted. These graphs in- 

 dicate that Spring Creek Hatchery alone rep- 

 sents a better unit of study than the combined 

 river systems. The correlation coefficient 

 using Spring Creek Hatchery data only is 

 highly significant (P<0.002) and actually much 

 greater than the correlation for the combined 

 data, which is not even significant. 



A consideration of the probable factors af- 

 fecting the widely divergent points for 1943 



2 4 6 8 10 12 



POUNDS Of FINGERLINGS RELEASED AT SPRING CREEK (THOUSANDS) 



14 



So 



2 4 6 8 10 1^ 



b POUNDS OF FINGERLINGS RELEASED AT SPRING CREEK AND 

 BIG WHITE SALMON (THOUSANDS) 



Figure 5,— Correlation between fingerling releases in 

 pounds and adjusted renirn of females by brood year 

 (1941-49). 



and 1949 broods in figure 5b is of interest. 

 In 1943, due to high water, almost the entire 

 run to Big White Salmon River was permitted 

 to spawn naturally (24,000 eggs taken), and 

 the large return to Big White Salmon River 

 in 1947 (actually much larger than indicated, 

 because a large but unknown number again 

 spawned naturally) probably reflects the pro- 

 ductivity of natural spawn. The large natural 

 spawning in 1947 also was followed by a large 

 return to Big White Salmon River in 1951. 

 Good natural spawning in 1950 and 1951, how- 

 ever, did not produce high returns in 1954 and 

 1955, although a localized gill net fishery may 

 have greatly reduced returns in these years. 

 The observed intensity of this fishery in 1955 

 resulted in a closure below Hood River Bridge 

 in 1956. 



For the 1949 brood year relatively large 

 numbers of fingerlings of Spring Creek 



10 



