702 



Fishery Bulletin 89(4), 1991 



Since equations for both the California and the Baja 

 California samples were highly significant, the ob- 

 served differences cannot be explained by measure- 

 ment errors or data variation, but probably reflect 

 different hake species. 



Fish weight was regressed against otolith and fish 

 lengths (Tables 3 and 4). When otolith size was used 

 to estimate fish weight directly, the correlation coeffi- 

 cient, and consequently percent r -squared for some 

 species (Table 3), was slightly smaller than the respec- 

 tive values of fish weight estimated from fish length 

 (Table 4). Regression lines and scatter plots for eight 

 fish species studied are shown in Figures 1-8. 



Coefficients of fish weight on fish length were the 

 highest (Table 4). The value of r 2 was greater than 

 90% in all cases. Thus, when comparing prey impor- 

 tance based on biomass in feeding-habit studies, both 

 equations (Tables 3 and 4) should be used to estimate 

 weight. 



Acknowledgments 



Thanks to various persons who participated in collec- 

 ting data for different species: Cuahutemoc Alonso, 

 Luz Elena Rizo D., Jacobo Schmitter, and Marco A. 



