Previous studies have shown that long-term 5-hydroxytryptamine (5-HT) depletion induced by combined dorsal and median raphe injections of 5,7-dihydroxytryptamine (5,7-DHT) leads to impairments in the acquisition and performance of behaviour maintained under a differential-reinforcement-of-low-rate (DRL) schedule of reinforcement. The present studies examined the relative importance of dorsal versus median raphe 5-HT projections in mediating these effects by investigating the impact of separate versus combined infusions of 5,7-DHT into these sites on DRL responding. Adult male rats received injections of 3 micrograms 5,7-DHT into both dorsal raphe and median raphe, dorsal raphe only, or median raphe only. Sham-operated controls received vehicle injections. Animals were then trained to respond on a DRL 20s schedule for 36 days. Combined dorsal raphe and median raphe 5,7-DHT infusions depleted striatal and hippocampal 5-HT by > 90%, increased responding, decreased the number of reinforcers earned, lowered the mean inter-response time (IRT), and shifted the frequency distribution of IRTs to the left. Median raphe 5,7-DHT infusions induced similar behavioural effects and reduced hippocampal 5-HT by 64%. Dorsal raphe 5,7-DHT lesions, that depleted striatal 5-HT by 56%, and hippocampal 5-HT by 30%, had no effect on behaviour. In a final experiment it was shown that providing an external cue light to signal food availability, and that removed the need to rely on internal timing processes, resulted in rapid acquisition of responding, and prevented the behavioural deficits in 5,7-DHT-lesioned rats. Following removal of this cue, responding deteriorated, and the 5,7-DHT treated animals performed worse than controls. These results suggest that destruction of 5-HT neurons arising from the median raphe is sufficient to disrupt behaviour maintained by a DRL 20s schedule, and that the behavioural changes observed may involve a deficit in timing.