FISHEKY BULLETIN; VOL. 86, NO. 1 



ing the three stocks, provided sufficient evidence 

 to reject the null hypothesis. In each of the 20 

 years analyzed, the main effects of tonnage class 

 and statistical area represented highly signifi- 

 cant (P < 0.01) sources of variation. The tonnage 

 class-area interaction term was also highly sig- 

 nificant in all cases, implying that vessels of var- 

 ious tonnage classes exhibit different CPUE 

 trends relative to each other in different areas. 



The initial results established the basis for fur- 

 ther investigations. In subsequent analyses, the 

 data were grouped to test the null hypothesis that 

 no significant differences in CPUE existed among 

 the three traditionally accepted stock definitions 

 (Georges Bank, Southern New England, and 

 Cape Cod). The highly significant differences ob- 

 tained in 19 out of 20 years indicate that catch 

 rates differ among the three stocks. The resulting 

 highly significant tonnage class-stock area inter- 

 action term obtained from the ANOVAs in all 

 years suggests that standardization of CPUE 

 among tonnage classes should be performed sepa- 

 rately for each stock. 



Analysis of variance within each stock provided 

 the final basis for performing the standardized 

 CPUE calculations. In these tests, the rejection of 

 the null hypothesis for tonnage class main effects 

 in all years for each stock suggests that separate 

 fishing power coefficients must be calculated for 

 each tonnage class even though the coefficients 

 are similar in many cases. The ANOVA results 

 also indicated that differences in CPUE among 

 statistical areas within each stock were highly 

 significant in SO^f or more of the years implying 

 that, within each stock region, yellowtail abun- 

 dance is not homogeneous. This is not surprising 

 since yellowtail flounder are prevalent only on 

 certain grounds within each geographic region. 

 Further analyses of the data by depth indicated 

 no overall significant differences in CPUE be- 

 tween the two primary depth zones (1-55 m and 

 56-110 m) where yellowtail flounder are consis- 

 tently caught. 



The infrequent number of significant interac- 

 tions on Georges Bank and Cape Cod grounds 

 relative to Southern New England (Table 4) sug- 

 gests a greater independence of the tonnage class 

 and area main effects with respect to CPUE, i.e. 

 both large and small vessels exhibited relatively 

 similar changes in mean CPUE among statistical 

 areas within each stock. In choosing data sets for 

 computing fishing power coefficients we sought to 

 minimize the amount of interaction among the 

 vessel tonnage classes and geographic areas in- 



volved. This criterion was met to a greater extent 

 for the Cape Cod and Georges Bank stocks than 

 for the Southern New England stocks. It appears 

 that yellowtail flounder inhabiting this region 

 are subject to a more complex set of interactions 

 perhaps due to temperature and bottom type. We 

 decided, however, to accept the results for each of 

 the three stocks and proceed with the calculations 

 of fishing power coefficients. 



Annual fishing power coefficients computed for 

 each vessel tonnage class fishing on Georges 

 Bank, Southern New England, and Cape Cod 

 grounds provided a basis for examining the con- 

 sistency in relative fishing power of individual 

 tonnage classes over time. Annual deviations for 

 Georges Bank and Southern New England 

 grounds indicated a gradual change in relative 

 fishing power of most tonnage classes between 

 1964 and 1983, and tests for autocorrelation of 

 residuals indicated significant time effects. On 

 these grounds, larger vessels exhibited higher 

 catch rates relative to the standard in the later 

 years as compared with the earlier years. Since 

 many of the larger vessels have been replaced in 

 recent years by newer vessels which are, pre- 

 sumably, equipped with more sophisticated elec- 

 tronics, any attempt to relate CPUE to stock 

 abundance must account for such technological 

 advances. 



Similarly, changes in seasonal availability are 

 often great enough to mask interannual variation 

 in stock abundance. Thus, the presence of signifi- 

 cant tonnage class-season interactions may be ex- 

 plained by the ability of certain vessel classes to 

 effectively target seasonal concentrations. Since 

 peak spawning of yellowtail flounder occurs dur- 

 ing late spring (Lux 1964), the presence of high 

 seasonal coefficients during the second and third 

 quarters is not surprising. 



By specifying the model to include tonnage 

 class, annual, and seasonal components, we have 

 attempted to account for technological and sea- 

 sonal availability factors which interact with 

 temporal changes in abundance. Although other 

 factors could be incorporated in the model to ac- 

 count for a larger portion of the variation in 

 CPUE, analyses of historical commercial fishing 

 operations of this type are often limited to those 

 attributes which can be directly linked to land- 

 ings records (Kimura 1981; Westrheim and 

 Foucher 1985). An alternate approach adopted by 

 Stern and Hennemuth (1975) involved the use of 

 a study fleet of selected vessels whose characteris- 

 tics and fishing practices were closely monitored. 



106 



