Abstract. — We attempt to deter- 

 mine at what point in the early life his- 

 tory year-class strength is established 

 in two species of rockfish, Sebastes 

 mystinus and Sebastes flavidus. We com- 

 pare abundance estimates of young-of- 

 the-year rockfish before and after settle- 

 ment to determine whether this life 

 history transition alters relative year- 

 class strength. Estimates of pelagic ju- 

 venile abundance obtained in midwater 

 trawl surveys and indices derived from 

 direct underwater observations of 

 settled juveniles over a common 10- 

 year period (1983-92) were in good 

 agreement (r=0.58-0.86). Trends in 

 rockfish year-class strength were simi- 

 lar, in spite of substantial spatial sepa- 

 ration between the trawl and nearshore 

 study areas (50-350 km) and differ- 

 ences in the timing of the surveys (2-4 

 months). Thus, settlement seems to 

 have little effect on relative year-class 

 strength. Estimates of stage-specific 

 interannual cohort variability show 

 that coefficients of variation (CVs) for 

 five species of late-stage pelagic juve- 

 nile rockfish ranged from 0.96 to 2.25. 

 Cohort variability measured at recruit- 

 ment to the fishery (2-7 years) is much 

 less (CVs=0.60-1.39), suggesting that 

 year-class strength has been deter- 

 mined by the late pelagic juvenile 

 phase. Because cohort variability de- 

 clines from the recently settled juvenile 

 stage to the age at recruitment to the 

 fishery, some form of compensatory 

 mortality may ameliorate interannual 

 differences in reproductive success. 



On the development of year-class 

 strength and cohort variability in 

 two northern California rockfishes 



Stephen Ralston 

 Daniel F. Howard 



Southwest Fisheries Science Center 



National Marine Fisheries Service. NOAA 



3 1 50 Paradise Drive. Tiburon, California 94920 



Manuscript accepted 11 May 1995. 

 Fishery Bulletin 93:710-720 (1995). 



Interannual fluctuations in repro- 

 ductive success lead to substantial 

 recruitment variability in fisheries. 

 Understanding the causes of this 

 variability is an important manage- 

 ment issue. However, the problem 

 has yet to be solved despite pro- 

 longed, intensive study (Cushing, 

 1973; Sissenwine, 1984; Rothschild, 

 1986). A number of investigators 

 have argued that the greatest po- 

 tential for regulation of year-class 

 size occurs during the larval stage 

 (e.g. Shepherd and Cushing, 1980; 

 Houde, 1987). Likewise, Smith 

 (1985) has shown that in northern 

 anchovy, Engraulis mordax, domi- 

 nant year classes arise only when 

 larval mortality rates are low. Oth- 

 ers have suggested that population 

 regulation occurs later in the juve- 

 nile period (Beverton, 1984; Sissen- 

 wine, 1984; Sissenwine et al., 1984). 

 To identify at what point during 

 the early life history of fish year- 

 class strength is established, inves- 

 tigators have correlated young-of- 

 the-year abundance measures with 

 later stage recruitment indices. 

 Dementjeva (1964) demonstrated a 

 strong positive correlation between 

 the catch rate (CPUE) of "one sum- 

 mer-old" Caspian bream, Archo- 

 sargus rhomboidalis, in fishery-in- 

 dependent surveys and recruitment 

 indices derived from virtual popu- 

 lation analysis (VPA) of the catch. 

 Peterman et al. ( 1988) tested Lasker's 

 (1975) hypothesis that the annual 



abundance of northern anchovy re- 

 cruits is fixed at an early life his- 

 tory stage by comparing abun- 

 dances of eggs, 4.5-day-old yolk-sac 

 larvae, and 19-day-old larvae, with 

 estimates of age- 1 recruits derived 

 from an age-structured analysis of 

 the catch. They found very low cor- 

 relations (r=-0.09 to 0.07, n=13) and 

 concluded that year-class strength 

 is determined at some point after 

 age 19-day. Similarly, Bailey and 

 Spring (1992) examined the rela- 

 tionship between survey estimates 

 of walleye pollock, Theragra chalco- 

 gramma, larvae (15 mm SL) and 

 young-of-the-year juveniles (50-130 

 mm FL) with the numbers of age-2 

 fish that had recruited to the fish- 

 ery from a tuned VPA. They re- 

 ported a good correlation between 

 young-of-the-year juveniles and 

 age-2 fish (r=0.69), but the associa- 

 tion between larval abundance and 

 age-2 fish was not significant 

 (r=0.36). Like Smith (1985) they 

 concluded that strong year classes 

 are the result of good larval sur- 

 vival. 



Bradford (1992) modeled early 

 life history dynamics and the re- 

 cruitment process using informa- 

 tion assembled from the literature. 

 Owing to poor correlations between 

 the abundance of small larvae and 

 subsequent recruitment, he con- 

 cluded that year classes are fixed 

 after the early larval period. He also 

 showed that precise predictions of 



710 



