Cocaine (5.6-30 mg/kg, i.p.) was administered to nondeprived male rats trained to eat a palatable sweetened mash. Over a 60-min period, their behaviour was observed and recorded for a microstructural anaylsis. Cocaine suppressed feeding in a dose-dependent manner (significantly at 10 mg/kg and greater), and this was due in the main to a reduction in the frequency of eating bouts. In contrast, the mean duration of eating bouts was unaffected, except at the highest dose, 30 mg/kg. In addition, the rate of eating was not significantly affected by cocaine at any dose. Time-course data revealed that cocaine, at anorectic doses (10-30 mg/kg), initially suppressed feeding completely, and the duration of this suppression was proportional to the dose. In effect, cocaine delayed the initiation of feeding, thus bringing about the reduction in the number of eating bouts. Cocaine caused some stimulation of locomotor activity and rearing to the side of the observation tank, but did not affect rearing away from the centre, or immobility. Grooming proved to be very sensitive to cocaine's suppressant effect, with substantial inhibition occurring at 5.6 mg/kg (a sub-anorectic dose). These data are compared with previously published work with D-amphetamine and are contrasted with results for selective D1 and D2 dopamine receptor agonists.