Diet selectivity by the planktivorous bluegill sunfish (Lepomis macrochirus) is correlated with the fishes' visual resolution. Poor visual capabilities are thought to impose a constraint on the ability of small sunfish to choose prey. Li et al. (1985) suggested that sunfish subjected to conditions of poor visual resolution select prey in accordance with a model in which prey choice is determined by apparent prey size. Further, enhanced visual resolution permitted large sunfish to assess absolute prey size more accurately than could smaller conspecifics and, consequently, large-sized sunfish consumed relatively more large prey than was predicted by the apparent size model. Here, we examine the visual capabilities and prey selection of very small sunfish, extend the analyses of Li et al. by determining the optimal diet sets for the fish used in their study, and compare prey selection by all sunfish size categories to that predicted by the apparent size and optimal foraging models. Contrary to the conclusions of Li et al., small sunfish (< 20 mm standard length [SL]) were more selective than was predicted by the apparent size model and preferentially attacked small-sized Daphnia. The optimal foraging "prey" model predicted correctly the size-related changes in mean size of prey attacked by sunfish, whereas the apparent size model predicted a preference for larger Daphnia for all sizes of fish. Yet, the observed diet breadths were always larger than those predicted by the optimal foraging model. The visual system and energy considerations constrain prey selection in small sunfish but operate in opposite fashions.