Hannah and Jones: Fishery-induced population changes in Pandalus jordani 



45 



where W = Weight (g), and 

 L = Length (mm), 



this represents a 28.4% increase in mean weight at age. 

 This is an approximate figure scaled for Area 22 data; 

 however, other areas yield similar results. 



The exploratory correlation analysis (Table 2) shows 

 that sea surface temperature at Charleston, Oregon, 

 and sea level at Newport, Oregon (Fig. 5), are positively 

 correlated with the age-1 growth index. However, 

 CPUE (Fig. 6) is negatively correlated with the age-1 

 growth index suggesting density-dependent growth. 



Adjusted sea level at Newport and sea surface tem- 

 perature at Charleston displayed evidence of a curvi- 

 linear relationship, of decreasing slope, with the age-1 

 growth index. However, the regression coefficients of 

 the second-order polynomial regressions were only 

 significantly different from zero (t-test, P>0.05) for 



adjusted sea level in year t - 1 . The curvature of this 

 relationship is strongly influenced by a single outlier, 

 the 1983 sea level, data point. With this point deleted, 

 the coefficients of the polynomial regression are not 

 significantly different from zero, while the adjusted 

 r-squared value for the simple linear regression in- 

 creases to 0.528 (Table 2). This suggests that bottom 

 temperature is also influencing shrimp growth. 



Graphs of mean length at age for areas 19, 21, 22 

 and 26 (Fig. 7) show that much of the difference in 

 growth between the 1965-76 and 1977-86 broods is 

 already apparent at age-13 months when the fishery 

 first catches age-1 shrimp. This is not surprising, since 

 density-dependent growth has been demonstrated 

 more frequently for species during the immature phase, 

 with density-dependent changes in fecundity more 

 prevalent in the adult phase (Bailey and Almatar 1989). 

 With the exception of area 19, the curves for the 



