21 )S 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



Table 2. — Average calculated lengths (1) of year classes al 

 4-ring and older, sampled in the commercial catch of 

 pilchard, British Columbia and San Pedro 



I Parentheses indicate average based on 13 5-ring fish from Washington 

 brackets, fish from Oregon; asterisks, 9 fish sampled] 



CONSISTENCY OF GROWTH CURVES 



Year classes 1937 through 1942 provide, suffi- 

 cient data for construction of curves of growth 

 similar to those of the 19.39 class (see tables 1 and 

 2). For the year classes 1937 through 1910, the 



estimates of means have greater reliability owing 

 to more extensive sampling during seasons 1941-42 

 through 1944-45. A decline in catches coincided 

 with curtailment of the sampling program so that 

 from seasons 1945-46 through 1949-50 the esti- 

 mates of mean lengths are based on fewer fish. 



The transformations for year classes 1937, 1938, 

 1940, 1941, and 1942 are similar in growth char- 

 acteristics to those of the 1939 year class (see 

 table 3). From analysis of covariance, no signifi- 

 cant difference was apparent in the mean slopes, 

 k, of the transformations between each of these 

 year classes in the two areas of catch, Canada and 

 San Pedro. That slope, or rate of deceleration of 

 growth, is the more stable of the two growth char- 

 acteristics, and that, with relatively constant en- 

 vironments, slope is a physiological character of 

 genetic meaning, is suggested from an experi- 

 mental study of growth in Platypoecilus maculatus 

 (Felin 1951). 



For each of the year classes tested, two distinct 

 y-intercepts, or levels, for transformations of mean 

 calculated and observed lengths are maintained in 

 the northern and southern areas. From the co- 

 variance tests, the differences are significant at 

 the 1-percent level. Further translated into pre- 

 ^/-intercept_ 



dieted ultimate size 



--}„ h the range 



l-jfc 



of the means for San Pedro is 220-236 millimeters 

 standard length, and for Canada, 249-258 milli- 

 meters, with no overlap for those year classes 

 tested (table 3). 



Table 3. — Growth characteristics in northern and southern 

 areas of catch derired from regressions of means l n+ i on 1 



BIOLOGICAL SIGNIFICANCE OF DIFFERENCE IN 

 GROWTH CHARACTERISTICS 



Significant difference in level, or the growth 

 characteristic /„, may represent phenotypic re- 

 sponse of a plastic genotype to varying hydro- 

 graphic environments. It may be useful as an 



