The first five otolith growth zones of 744 Pacific halibut from 26 year-classes (1953–1978) were analyzed to identify patterns of annual growth and the sources of temporal variation. Sexual differences in growth appeared at an early age, and size selectivity of the gear types was shown to influence the perception of past growth. We adjusted the growth record for these two effects by removing variation in annual otolith growth that was linearly related to the size of the fish at capture. Correlation patterns of the otolith zones at the youngest ages suggested a general uncoupling of individual growth from one year to the next. Temporal variation was examined by developing a linear model to partition growth at different ages into both year and year-class effects. Randomization tests indicated that both effects are significant when all five ages are included, but when subsets of the data are examined, the year effect is strongest for youngest juveniles, whereas the year-class effect is significant for older juveniles. The year effect is probably attributable to interannual temperature changes, as indicated by a strong linear relationship between sea surface temperature (SST) and otolith growth from ages 0 to 2 years. However, analysis of residuals indicated a remaining year effect, suggesting that SST is an imperfect measure of the actual environment that regulates juvenile growth. The year-class effect observed in older juveniles was not present in the data adjusted for size selectivity, suggesting that this effect could be influenced by sampling bias, though intrinsic influences cannot be ruled out. Density-dependent growth did not appear to be a factor accounting for growth variation; instead, there appeared to be a slight positive correlation between otolith growth at ages 1 and 2 years and estimated year-class abundance (at age 8 years), suggesting that early growth is a factor in determining year-class strength.