Maximizing the average rate of energy intake (profitability) may not always be the optimal foraging strategy for ectotherms with relatively low energy requirements. To test this hypothesis, we studied the feeding behaviour of captive insectivorous lizards Psammodromus algirus, and we obtained experimental estimates of prey mass, handling time, profitability, and attack distance for several types of prey. Handling time increased linearly with prey mass and differed significantly among prey types when prey size differences were controlled for, and mean profitabilities differed among prey taxa, but profitability was independent of prey size. The attack distance increased with prey length and with the mobility of prey, but it was unrelated to profitability. Thus, lizards did not seem to take account of the rate of energy intake per second as a proximate cue eliciting predatory behavior. This information was combined with pitfall-trap censuses of prey (in late April, mid-June and late July) that allowed us to compare the mass of the prey captured in the environment with that of the arthropods found in the stomachs of sacrificed free-living lizards. In April, when food abundance was low and lizards were reproducing, profitability had a pronounced effect on size selection and lizards selected prey larger than average from all taxa except the least profitable ones. As the active season progressed, and with a higher availability of food, the number of prey per stomach decreased and their mean ize increased. The effect of profitability on size selection decreased (June) and eventually vanished (July-August). This variation is probably related to seasonal changes in the ecology of lizards, e.g. time minimization in the breeding season as a means of saving time for nonforaging activities versus movement minimization by selecting fewer (but larger) prey in the postbreeding season. Thus, the hypothesis that maximizing profitability could be just an optional strategy for a terrestrial ectothermic vertebrate was supported by our data.