Mice descending from lines previously selected for high and low levels of interfemale aggression and from a nonselected control line were exposed to live crickets on two consecutive test days. Latencies both to attack and to eat the cricket were recorded. No line or sex differences in attacking were found, although latency to attack decreased over test days. Control line mice were more likely to engage in the eating of the cricket than either the high or low selected lines, suggesting that the selection procedure may have dissociated attacking and eating components of predatory behavior.