718 



Fishery Bulletin 93(4), 1995 



catch-at-age analysis in which the stock-synthesis 

 model and other age-structured methods are em- 

 ployed, is typically less than 1.00 ( 3c =0.84; Table 1). 

 The "Entry to fishery" CVs presented in Table 1 were 

 based on time series of recruitments ranging from 

 10 (Sebastes goodei and Sebastes paucispinis) to 25 

 years (Sebastes flavidus), although they did not all 

 span the same time period. 



A compensatory mortality source acting from 

 postsettlement until recruitment to the fishery would 

 tend to ameliorate year-class differences observed in 

 the pelagic and recently settled juvenile stages. Two 

 possible agents of compensatory mortality at this 

 stage in life are 1) intraspecific competition for food 

 (Shepherd and Cushing, 1980) or 2) a type-Ill func- 

 tional response (sensu Holling, 1959) by predators 

 (i.e. predator switching). Likewise, a rapid numeri- 

 cal response by predators could also lead to compen- 

 satory mortality. 



There are other data to support this interpreta- 

 tion. Adams and Howard 4 provide data showing that 

 blue rockfish experience compensatory mortality 

 from the time they settle (August-September) until 

 the following spring. They attributed the increased 

 mortality rate experienced by strong year classes to 

 predator switching. Hobson et al. 5 also describe pre- 

 dation on recently settled blue rockfish by Hexa- 

 grammus decagrammus, Sebastes melanops, and 

 Sebastes mystinus (i.e. cannibalism), but only in years 

 when settlement was particularly strong (see above). 

 Likewise, Hallacher and Roberts ( 1985) showed low 

 dietary overlap during the nonupwelling season 

 among a kelp-forest assemblage of six Sebastes spp. 

 However, when newly settled young-of-the-year rock- 

 fish became abundant during the upwelling season, 

 these species fed heavily on the juveniles, and di- 

 etary overlap indices rose sharply. In addition, these 

 authors observed other kelp-forest predator species 

 feeding heavily on juvenile rockfish (Ophiodon 

 elongatus, Anarrhichthys ocellatus, and Scorpae- 

 nichthys marmoratus). That a broad suite of preda- 

 tors capable of switching onto juvenile rockfish has 

 been described provides a plausible compensatory 

 mortality mechanism. Similar conclusions regarding 

 the importance of compensatory mortality in recently 

 settled plaice, Pleuronectes platessa, living in the 

 North and Wadden Seas have been drawn by 

 Lockwood (1980), Zijlstra et al. (1982), and van der 

 Veer (1986). 



It is a widely held precept that stage-specific mor- 

 tality rates generally decrease with ontogeny, while 

 stage duration increases concomitantly (Miller et al., 

 1988; Bradford, 1992). Because the total stage mor- 

 tality is the product of the instantaneous mortality 

 rate and the stage duration (M-t), an increase in du- 



ration may more than offset a decrease in rate (Shep- 

 herd and Cushing, 1980). This has led some to sug- 

 gest that population regulation may occur during the 

 juvenile phase of the life history (Beverton, 1984; 

 Sissenwine, 1984; Sissenwine et al., 1984). Our find- 

 ings indicate that rockfish year-class strength is prob- 

 ably determined at some point in the larval phase, 

 which lasts at least 50 days in Sebastes jordani 

 (Laidig et al., 1991). Upon completion of the larval 

 stage, however, compensatory density dependence in 

 the settled juvenile phase seems to reduce cohort 

 variability. 



We have argued that recruitment success in these 

 species of rockfish is governed by large-scale oceano- 

 graphic processes (see also Mearns et al., 1980). 

 There is evidence that year-class strength depends 

 on the thermal environment at the time of spawning 

 (Fig. 4), although we use SST only as a simple proxy 

 for some complex set of covarying physical variables. 

 However, the association between these variables is 

 nonlinear; apparent year-class failures occur at the 

 extremes of the continuum. This conclusion is sup- 

 ported by the findings of Ainley et al. (1993), who 

 showed that the early summer occurrence of pelagic 

 juvenile rockfish in the diet of a seabird (common 

 murre, Uria aalge) was parabolically related to up- 

 welling in January and February. Too little or too 

 much upwelling during the rockfish spawning sea- 

 son had a negative impact on the availability of pe- 

 lagic juvenile rockfish six months later. Notably, up- 

 welling and SST in that study were inversely corre- 

 lated. Cury and Roy (1989) have also argued that 

 recruitment success of pelagic fish stocks in up- 

 welling systems is greatest at an intermediate point 

 along the environmental continuum, although they 

 argued that wind speed is the forcing mechanism. 



Summary 



We argue that events occurring in the larval period 

 are primarily responsible for determining the suc- 

 cess or failure of rockfish year classes in central Cali- 

 fornia. Moreover, on the basis of a consideration of 

 intra- and interspecific synchrony and oceanographic 

 scale, physical factors seem to have the greatest im- 

 pact on larval survival. Consequently, to the extent 

 that recruitment limits stock size in rockfish, popu- 

 lation regulation is based on larval dynamics. How- 

 ever, compensatory mortality in the juvenile phase 

 may ultimately limit population growth. Like Myers 

 and Cadigan (1993), we believe that the interplay 

 between stochasticity in the larval period and com- 

 pensation in the juvenile phase plays a seminal role 

 in structuring rockfish population dynamics. 



