FISHERY BULLETIN: VOL. 87, NO. 3, 1989 



4.0 5.0 6.0 7.0 8.0 9.0 



LENGTH (mm) 



Figure 4. — Length distribution of nehu larvae by 0.5 mm intervals before correc- 

 tion for extrusion and avoidance (upper panel) and after correction (lower panel). 



100.0 



O 80.0 

 o 

 o 

 o 



'3 60.0 



o 

 < 



< 



> 



40.0 



20.0 



0.0 



596.77 



2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 

 LENGTH (mm) 



Figure 5. — Bootstrap estimate of variance of the corrected number of larvae 

 by 0.5 mm length intervals. 



especially when the corrected length-frequency 

 distributions will be subsequently used to esti- 

 mate mortahty rates. Whether mortality is esti- 

 mated by using methods that require converting 

 length-frequency distributions to age-frequency 

 distributions (Leak and Houde 1987) or by using 

 some form of length-based method (Wetherall et 

 al. 1987), precision of the mortality estimates 

 will depend on the precision of corrected length- 



frequency distributions. Thus, procedures used 

 to estimate mortality from larval length-fre- 

 quency distributions should include weighting 

 factors that incorporate the size-specific vari- 

 ances of the corrected length-frequency distribu- 

 tions. 



Most of the problems associated with viola- 

 tions in the assumptions P,,, = P,-,- and P, „ = P,,. 

 could be eliminated with proper attention to the 



454 



