I review the role of environmental variability in the survival of juvenile fish and shellfish by examining the success of previously published environment–recruitment correlations when tested with new data. The proportion of published correlations that have been verified upon retest is low. There is one generalization that stands out: correlations for populations at the limit of a species' geographical range have often remained statistically significant when re-examined. An examination of environment–recruitment correlations that were reviewed 13 years ago by Shepherd and co-workers shows that only 1 out of 47 reviewed studies is currently used in the estimation of recruitment in routine assessments. The results suggest that future progress will require testing general hypotheses using data from many populations.