When naloxone-precipitated abstinence occurs in morphine-dependent mice, they exhibit stereotyped jumping. The number of jumps has been used as an abstinence measure and the motivation for them described as 'escape'. Studies were designed to assess stimulus control and motivation here. Groups of ten dependent and ten saline mice were singly tested in both light and dark conditions in each of five covered cylinders (2-23 in high). Jumps were recorded automatically for 10 min after 10 mg/kg naloxone. At 2 in, which permitted tactile contact with the cover, no jumps occurred in the light or the dark; at 5-11 in, a high incidence of jumping occurred in both the light and the dark; at 23 in, too high for the animal to reach, relatively few jumps occurred in the light but many occurred in the dark. In another study, similar groups were singly tested in the light after naloxone on each of five circular platforms (18 in high, 3-48 in diameter). As the diameter increased, the number of mice jumping to the floor showed a linear decrease concomitant with an increase in latency. This suggests that jumping is controlled by visual and tactile cues; mice do not jump when escape is prevented by a cover which can be touched or a height which cannot be reached; but do jump when such cues are eliminated by intermediate heights or darkness. Latency varies with size of immediate explorable environment.