The growth rate of larval fish is a predictor of the annual supply of young fish, but fisheries science has not exploited this remarkable finding, particularly for Australia’s sparse and diverse fisheries. Using a rare 28 year archive of herring otoliths from the Bay of Fundy off eastern Canada, with robust year class strengths based on virtual population analysis, we tested the hypothesis that strong year classes were produced from larger otoliths, and weak year classes were produced from smaller otoliths. Two weak year-classes were selected with ~5*10^8 individuals at 1 year old, and two strong year-classes were selected with abundances ten-fold greater. We examined otolith size (diameter and area) of 2, 3, 4, and 5 year-old fish at approximately 18 months of age, and measured the otolith radius of 3 year-old fish at 6 months, 2 months and 1 month post-hatch. In 2 year-old fish, the area of the otolith at ~18 months old was correlated with year class strength, but effects were quickly removed as the cohort moved onto 3, 4, and 5 years old. Measures of daily increment growth in 3 year-old fish did not yield significant differences among year-classes due to high variability of larval growth rates in certain year-classes.