BOEHLERT ET AL.: TIME SERIES OF GROWTH IN SEBASTES 



tween deep and shallow water over the con- 

 tinental shelf off Oregon (Kruse and Huyer 

 1983); thus, differences between gi'owth rates 

 for age group 1 and later age groups are not 

 surprising. 



The dominant signal apparent in all of the time 

 series of growth, except age group 1 of S. dip- 

 loproa (Figs. 3, 4), and in the fii'st principal com- 

 ponent for both species (Figs. 6, 7) was that of 

 increased growth after about 1972. Many of the 

 fish used to calculate the gi'owth anomaly values 

 for this period would have been captured before 

 the age at full recruitment to the fishery, which 

 is 12 years for S. pinniger (Wilson 1985) and 14 

 years for S. diplopwa (Boehlert 1980). A poten- 

 tial concern wdth the use of fish younger than the 

 age at full recruitment is that the gear will be 

 selective for larger, faster gi-owing individuals; 

 thus, fish younger than the age at full recruit- 

 ment might conceivably be characterized by 

 moi'e rapid gi'owth rates, resulting in the in- 

 creased growth rates observed late in the time 

 series (Figs. 3. 4). If true, then the size of 

 growth increments for fish born in a given year 

 should decrease with time from first recruitment 

 until the age at full recruitment (as is true in 

 Lee's phenomenon). Our samples, however, 

 were taken with sampling gear of much smaller 

 mesh than used in the commercial fishery 

 (Gunderson and Sample 1980). Nonetheless, to 

 test for more rapid growth of younger age fish, 

 we compared the six gi-owth indexes of S. pin- 

 niger born in 1973-77 and collected in 1980 (A^ = 

 41, ages 3-7) with those collected in 1984 (A'^ = 

 51, ages 7-11). This timespan covered the most 

 rapid increase in gi-owth (Figs. 3, 4). The gi'owth 

 indexes did not significantly differ (paired ^test, 

 P > 0.10); this suggests that the size of gi'owth 

 increments did not change as fish born in 1973- 

 77 were collected 4 years closer to the age at full 

 recruitment (1984 versus 1980). 



An alternative explanation of increased 

 growth in the 1970's was density dependence. 

 Density-dependent gi*owth has been observed in 

 a variety of fish stocks, generally in association 

 with exploitation (Margetts and Holt 1948) or 

 strong year classes (Jones 1983), and is most 

 evident in immature fish (see summary in Ware 

 (1980)). There is little question that stocks of 

 several Sebastes spp. have declined under the 

 influence of fisheries (Gunderson 1984; Bracken 

 1987; Lenarz 1987; Westrheim 1987). Ito et al. 

 (1987) have suggested that stocks of S. alnfiis off 

 Oregon, Washington, and British Columbia de- 

 clined from a vu'gin biomass of about 144,000 



metric tons (t) to about 13,500 t in the early 

 1970's. No direct work documents density- 

 dependent changes in gi-owth in this species, 

 although Gunderson's (1977) model of the stock 

 used increased growth as a compensatory 

 mechanism at low stock density. The best evi- 

 dence for density-dependent growth changes in 

 the genus is for S. mentella (Sorokin et al. 1986). 

 For S. diplopwa, length frequencies from 1977 

 (Boehlert 1980) were shifted to much smaller 

 sizes than those from 1961 to 1962 (Alverson et 

 al. 1964); this decrease may have been a result of 

 fishing pressure (Boehlert 1980). Faster growth 

 in the region north of California has been de- 

 scribed for this species (Boehlert and Kappen- 

 man 1980); density-dependent growth increase 

 (since the major stock reduction occurred in the 

 north) was suggested as one of the factors re- 

 sponsible for the geographical gi'owth differ- 

 ences. 



The decrease in stock size of many deepwater 

 Sebastes spp. along the west coast of North 

 America and the Gulf of Alaska may be related 

 to the growth increase after about 1972 for both 

 species. The major removals of this group by 

 foreign fisheries occurred in the mid-1960's to 

 1970's (Bracken 1987; Ito et al. 1987; Westrheim 

 1987). Time series of reliable biomass estimates 

 for S. pinniger and S. diploproa are not avail- 

 able, but estimates for S. alutus have been made 

 using the stock reduction analysis method (Ito 

 et al. 1987). All three species inhabit similar 

 environments and depend upon similai" food re- 

 sources, so biomass estimates for S. alutus can 

 be used as a proxy of biomass for the other two 

 species. A comparison of the first principal com- 

 ponent time series of S. pinniger and S. dip- 

 loproa with the stock size of S. alutus on the 

 west coast (Ito et al. 1987) shows that the in- 

 crease in growth begins slightly after the major 

 stock decline (Fig. 8). The stock size of S. alutus 

 is negatively correlated with the first principal 

 component (and thus growth indexes) for both 

 species {P < 0.01). 



If decreased stock size is responsible for the 

 increased growth evident in both species, it ap- 

 parently occurs in age groups 1-6 for S. pinniger 

 (Figs. 3, 5A) and in age groups 2-6 for S. dip- 

 lopwa (Figs. 4, 5B). The ecological differences 

 between age gi'oup 1 and older 5. diploproa (de- 

 scribed above) may explain why density-depen- 

 dent growth does not occur in the young fish. 

 Density-dependent gi-owth in the first year of 

 life has been described for several fishes (van der 

 Veer 1986; Peterman and Bradford 1987) that 



801 



