We examined the effects of amount of food on vitellogenic oocyte size and number (fecundity) between two experimental groups of Atlantic herring (Clupea harengus) kept in separate but identical compartments of a large seawater tank. Growth and maturation were monitored for 1.5 years. Each group of 200 fish was fed exclusively on krill. At the end of the experiment, a high-ration group, fed 12 mg ·g fish-1 ·day-1, had similar weights-at-length as wild specimens of the same stock. A low-ration group fed about half the amount of the high-ration group had significantly lower weights-at-length. Ration had no effect on oocyte size and relative potential fecundity (oocytes per gram), but mean potential fecundity (oocytes per fish) of the low-ration fish was 26% lower than that of the high-ration fish. The relative intensity of early stages of atresia was quantified histologically. Atresia is a common phenomenon in Atlantic herring and increases markedly with suboptimal feeding. Mean relative atretic intensities were estimated at about 3 and 6% for the high- and low-ration groups, respectively, but with large intragroup variation. Modeled realized fecundity (spawned eggs per fish) indicated a 9 and 40% reduction compared with the measured potential fecundities, respectively.