Laboratory feeding trials were conducted with the predaceous stonefly Hesperoperla pacifica and a number of mayfly and dipteran prey species to investigate the effects of predator size, and prey size and morphology, on the predator's success. Observations under dim red light permitted estimation of encounter rate (E/min), attack propensity (A/E), capture success (C/A) and handling time (HT). For prey of a particular species and size, HT decreased log-linearly with increasing predator size. Across all prey categories, HT increased log-linearly with increasing values of the ratio prey dry wt/predator dry wt, and differences among species appeared to be small. Overall, capture success was low, but C/A was higher for dipterans than for mayflies, especially with large H. pacifica. Predator size affected C/A when prey fell within a certain size range, but was not a detectable influence with very small or very large prey. Values of A/E of near 10% typified many predatorprey combinations; however, ephemerellid mayflies suffered markedly fewer attacks, and values of A/E up to 30% were obtained with some species-size combinations. We estimated benefit to the predator first as prey wt ingested per unit time (dry wt/HT), and second by mutliplying the former term by capture success. Values increased with increasing size of the predator, and inclusion of the C/A term indicated that predators would obtain greater reward from small relative to large prey, and from dipterans relative to mayflies. Howerver, there was little evidence that attacks were biased toward more profitable prey. We compare the relative contributions of E/min, A/E and C/A to prey choice, and discuss their applicability to predation events in nature.