FISHERY BULLETIN VOL 77, NO 3 



Temperature in 1975 ranged 3.1-11.2" C and was 

 increasing during the run. Greatest fry movement 

 in this year occurred when river temperature was 

 6°-9° C (Figure 3). In 1976 discharge ranged 18- 

 91m% and was higher early in the season than in 

 1975. Increases in the 1976 fry run often preceeded 

 increases in discharge (Figure 3). River tempera- 

 ture ranged 5.0°-13.3° C and greatest fry move- 

 ment was when temperature was 8^-11" C (Figure 

 3). 



In addition to temperature and discharge, the 

 catch of chinook salmon in the traps was probably 

 influenced by tide. The traps were set very near 

 the river mouth and at high tide flow past the traps 

 was often negligible. To examine the potential 

 contribution of discharge, river temperature, and 

 tide height to variations in trajD catch, I performed 

 a stepwise multiple regression analysis on the 

 data. The dependent variable was trap catch and 

 the independent variables were river discharge, 

 river temperature (morning and evening mea- 

 surements averaged), average tide height during 

 three periods of the "trapping day" (0800-1800 h, 

 1800-0000 h, 0000-0800 h), and Julian day of cap- 

 ture. I performed separate analyses on catches 

 preceding and following the peak catch each year. 

 The hypotheses tested were; 1) catch is positively 

 correlated with discharge and temperature and 

 negatively correlated with tide height for all data 

 sets; 2) catch is positively correlated with day of 

 capture prior to peak catch and negatively corre- 

 lated after peak catch. 



The regression analysis failed to confirm or re- 

 ject either of these hypotheses unequivocally. Dis- 

 charge was positively correlated with trap catch 

 while catches were increasing, but was not corre- 

 lated while catches were decreasing (Table 1). 

 Temp)erature was not significantly correlated with 

 catch in any of the analyses. Tide height was nega- 

 tively correlated with trap catch while catches 

 were increasing as predicted. While catches were 

 decreasing, however, tide height was uncorrelated 

 with trap catch in 1975 and positively correlated 

 in 1976 (Table 1). The correlation of trap catch 

 with Julian day was positive while catches were 

 increasing and negative while catches were de- 

 creasing, as predicted, except that the correlation 

 with increasing catch was not significant in 1976 

 (Table 1). The multiple correlation coefficients 

 were highly significant and explained 50-79% of 

 the variation in trap catch {R^. Table 1). Some of 

 the results, like the positive correlations between 

 trap catch and tide height, were counterintuitive, 

 however, and cast doubt on any interpretation of 

 the regression analysis. In spite of these difficul- 

 ties the regression analysis suggests that dis- 

 charge and tide height may have influenced trap 

 catch, while temperature probably did not. 



Recaptures of marked fry in the traps in 1975 

 ranged 0-16. 6*^^ of the daily estimate of marks 

 migrating. The ratio of recaptures to marks mi- 

 grating for the whole run was 0.0175, indicating 

 an overall trap efficiency of 1.757( (Table 2). 



Peterson estimates of total daily run were made 



Table l . —Results of stepwise multiple regression analysis of fry trap catch of juvenile chinook salmon 

 regressed on river discharge, river temperature, average tidal height during three daily time periods 

 (0800-1800 h, 1800-2400 h, 2400-0800 h) and Julian day of capture. Only the regression coefficients for 

 the variables that made a significant (P<0.05) contribution to the multiple regression are shown. 



0.50 



658 



