Temporal and spatial variability in fecundity was investigated for three species of north-west Atlantic flatfishes: American plaice Hippoglossoides platessoides, yellowtail flounder Limanda ferruginea and witch flounder Glyptocephalus cynoglossus. Significant log10 linear relationships existed between fecundity and total length (LT) for all three species and for all combinations of area (NAFO regulatory areas) and year. Fecundity differed between areas for yellowtail flounder and witch flounder and changed significantly for all three species since the last reported fecundity studies in the region 30–40 years ago. Three populations experienced decreases in fecundity (3LNO American plaice and yellowtail flounder, 3Ps witch flounder), one experienced an increase (3NO witch flounder) and one did not change (3L witch flounder). Significant yearly variability in fecundity at length and relative fecundity existed for all three species within the 6 year sampling period (1993–1998). Estimates of population egg production based on new and historic fecundity data revealed that using the old, invariant fecundity and LT relationship resulted in overestimation of total annual egg production by up to 24% for 3LNO American plaice and as high as 41% for 3LNO yellowtail flounder. These results clearly demonstrate the variable nature of fecundity for flatfishes in Newfoundland waters and suggest that the use of an invariant fecundity and LT relationship may result in inaccurate estimates of stock reproductive potential.