268 



Fishery Bulletin 101(2) 



1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 



Year 



Figure 4 



Estimated recreational, commercial, and total cowcod landings during 1951-97. 



Age, growth, and maturity 



The youngest fish in the 263 cowcod sampled for age deter- 

 mination was age 1, and the oldest was age 55. Average 

 percent error (Beamish and Fournier, 1981) for readings by 

 three (or four) readers was 0.09 and the index of precision 

 (Chang, 1982) was 0.08. 



Von Bertalanffy parameter estimates for male and fe- 

 male cowcod size at age were L_ = 91.5 and 91.8 cm FL, ^ = 

 0.0459 and 0.0447, and t^ = -2.41 and -1.88/yr. Growth pa- 

 rameters were not significantly different; therefore we com- 

 bined data from both sexes and incl uded specimens for which 

 sex was not determined. Von Bertalanffy parameter esti- 

 mates for growth in length with sexes combined were L^ = 

 86.9 cm FL, k = 0.0524, and /„ = -1.94/yr (Fig. 5). The corre- 

 sponding von Bertalanffy parameter estimates for growth 

 in weight were W_ = 35.1 kg, K = 0.00605, t^ = 4.78 (Fig. 

 6). The estimate W^ = 35.1 kg appears to be an imprecise 

 estimate of maximum body mass because the largest cow- 

 cod reported in the recreational fishery are about 10-15 kg 

 (Wertz^). Male and female cowcod appear to reach sexual 

 maturity at about the same ages and lengths (Table 2). 



clear, but the age at full recruitment appeared to fall some- 

 where between age 10 and age 20. Age 17 was used in the 

 catch curve analysis because it gave the highest coefficient 

 of determination (r^) in catch curve regressions. The mean 

 of four estimates of mortality based on age data (Table 3) 

 wasZ=0.071/yr. 



The natural mortality rate (M) for cowcod by Jensen's 

 (1997) method was 0.069/yr. In modeling and reference 

 point calculations for cowcod, we used the lowest estimate 

 (Z=0.055/y ) to approximate M. This estimate is crude, impre- 

 cise, and may be biased high because total mortality Z 

 includes both natural mortality (M) and fishing mortality 

 (F). 



Yield per recuit and spawning biomass per recruit 



Biological reference points for cowcod rockfish from yield- 

 per-recruit and spawning-biomass-per-recruit calcula- 

 tions were relatively low because of low natural mortality, 

 prolonged growth, and recruitment to the fishery prior to 

 full maturity. In particular, with M=0.055/yr, F^,.^;^~0.11, 

 F„ 1=0.048, and Fj„,, =0.039/yr. 



Natural mortality 



Slopes of catch curve regressions were similar for male and 

 female cowcod and for samples from the commercial and 

 recreational fisheries (Butler et al., 1999). We therefore 

 combined data for males, females, and unsexed samples to 

 increase sample size and reduce variance. The best choice 

 for age at full recruitment in catch curve analysis was not 



Abundance index data 



The best logistic model for CalCOFI data included terms 

 for season, month, line, and station effects and all 2- and 

 3-way interactions. Residual plots showed no evidence of 

 lack of fit. Larval presence for cowcod in the SCB (Table 4 

 and Fig. 7) varied without trend during 1950-67, was 

 elevated during 1968-74, and was then low beginning in 



