Six adult male research volunteers, in two groups of 3 subjects each, lived in a residential laboratory for 15 days. All contact with the experimenters was through a networked computer system, and subjects' behavior was monitored continuously and recorded. During the first part of each day, they were allowed to socialize. Two cigarettes containing active marijuana (2.7% delta 9-THC) or placebo were smoked during the private work period and the period of access to social activities. Three-day contingency conditions requiring subjects to engage in a low-probability work activity (instrumental activity) in order to earn time that could be spent engaging in a high-probability work activity (contingent activity) were programmed during periods of placebo and active-marijuana smoking. During placebo administration, the contingency requirement reliably increased the amount of time that subjects spent engaged in the low-probability instrumental activity and decreased the time spent engaged in the high-probability activity. During active-marijuana administration, however, the increases in instrumental activity were consistently larger than observed under placebo conditions. The decreases in contingent activity were similar to those seen under placebo conditions. Smoking active marijuana was thus observed to produce increments in instrumental activity under motivational conditions involving contingencies for "work activities."