Diet selection among several prey types present in a dense aggregation, permitting a predator to become satiated without changing patches, may be important for predators that can eat many small prey items in a single bout. Choice in this scenario differs from that in optimal foraging models for sequential diet choice model and simultaneous choice models when travel time between patches is needed. Furthermore, satiation and depletion effects may be important in dense prey aggregations. We predicted that in dense prey aggregations, predators should eat the most profitable prey first, switching to smaller prey as larger ones become depleted and predators become satiated, and that prey below some minimum profitability should be rejected. When large numbers of prey of varying sizes were presented simultaneously, broad-headed skinks (Eumeces laticeps) preferentially consumed large crickets, ate some medium-sized crickets late in ingestion sequences, but ate no small crickets. Prey depletion, with selection of the currently most profitable prey type, appears to account for much of observed prey switching, and satiation may contribute. When four crickets of each of four sizes were presented, lizards ate largest first, then medium-sized. Some then ate small crickets, but none ate very small crickets. These observations and exclusion of small crickets from the diet by many lizards when larger ones were unavailable support the predictions. In tests with three sizes of juvenile mice presented singly, the smallest were attacked at shortest latency and eaten, medium-sized mice were attacked at greater latency but could not be subdued, and large mice were not attacked. These data suggest that as prey become too large to subdue and eat readily, profitability declines until they are excluded from the diet. Unsuccessful attacks on medium-sized mice suggest that lizards had to learn their own capabilities with respect to a novel prey type.