Optimality models for life-history strategies predict distinct patterns of reproductive investment in environments differing in food availability. However, attaining the predicted optimal phenotypes may be prevented by genetic constraints, such as lack of additive genetic variation or genetic correlations. We tested predictions of optimality models in the laboratory using the eastern mosquitofish, Gambusia holbrooki, reared in high- and low-food environments. We quantified juvenile and adult growth rates, age and size at maturity, fecundity, egg size, reproductive allotment, and storage lipids in somatic and ovarian tissues. To test for genetic constraints on optimal phenotypic plasticity, we used a split-family, half-sib/full-sib quantitative genetic breeding design to quantify genetic correlations and additive genetic variation in reaction norms for these nine life-history traits. Fish in the high-food treatment had higher juvenile growth, decreased age at maturity, increased fecundity, increased reproductive allotment, and reduced storage lipids in both somatic and ovarian tissues relative to fish grown on a low- food diet. Patterns of age at maturity, reproductive allotment and energetic investment per egg were generally consistent with optimality models. Differences in somatic lipid content were inconsistent with previous dietary studies, and may have been due to the differences in time since maturity for fish on either diet. Genetic correlations between traits expressed in the two feeding environments were generally low, indicating that the potential for independent evolution of these traits were high. The levels of additive variation for reaction norms were significant for three of the nine traits measured (adult growth, age at maturity, and reproductive allotment), but were low for the other six traits. These genetic results suggest that the potential for adaptive phenotypic plasticity should not be constrained in this population, which corresponds with the observation that the fish responded to the diet treatments in directions predicted by optimality models.