228 
Fishery Bulletin 95(2), 1997 
maturation stage, i.e. “fertilized.” In contrast, the fish 
used in constructing the length-maturity curve for 
1968-72 were collected during February-June, when 
Pacific ocean perch are closer to the embryo-release 
period. 
Nevertheless, it appears that after 20 years in a 
depleted state, the stocks of Pacific ocean perch off 
Washington have partially compensated for a loss in 
reproductive potential by reducing their age at ma- 
turity from 10 years to 8. Comprehensive studies 
have shown long-term declines in age at maturity 
from 10.5 years (1923) to 8 years (1976) in a heavily 
fished stock of Northeast Arctic cod (Jprgensen, 
1990), and from 5-7 years (early 1900’s) to 4-5 years 
in North Sea plaice (Rijnsdorp, 1989). Although a 
genetic basis for such changes has been documented 
in some species (Policansky, 1993), disentangling 
genetic changes and phenotypic plasticity is often 
difficult. Given the long generation time of Pacific ocean 
perch and the relatively short time span involved, these 
changes probably reflect reaction norms of phenotypic 
plasticity rather than changes in genotype. 
In contrast, growth, as reflected in size attained 
at age 15, showed no substantial changes between 
1968-70 and 1992. Although monitoring length-age 
relationships at fixed bathymetric locations allows 
the depth effect to be controlled for, it is difficult to 
maintain the sampling depth within a range of less 
than about 18 m with trawls on the continental slope. 
Aggregations of Pacific ocean perch often have dif- 
ferent growth characteristics and vary interannually 
in their availability (Gunderson, 1974), and further 
sources of bias and variability are inherent in using 
different age-determination techniques. All of these 
factors make it difficult to detect changes in growth 
rates unless they are substantial. Nevertheless it 
should be kept in mind that interannual variations 
in food availability can often be more influential than 
changes in population density in determining growth 
rates (Rijnsdorp et al., 1991; Rijnsdorp and van 
Leeuwen, 1992). 
Most adult rockfish in the Oregon- Vancouver Is- 
land region probably migrate to a very limited ex- 
tent, and stocks within these regions represent a 
mosaic of small, highly localized stocks. Neverthe- 
less, practical considerations in terms of data collec- 
tion, data assessment, and management enforcement 
often force the geographic scale of fishery manage- 
ment to be relatively broad. For example, although 
the Pacific Fishery Management Council has at- 
