Rats were trained to self-administer intravenously-delivered cocaine. Four lever-press responses resulted in a cocaine infusion (0.2 mg/kg) during daily 24-h sessions. The rats were also trained to obtain water from tongue-operated solenoid-driven drinking spouts. Ground food and water from a standard drinking bottle were also available. When cocaine injections reached stable levels, L-tryptophan was mixed with the rats' food for 5 days. Three concentrations of L-tryptophan (2, 4, and 8%) were tested in different groups of five rats each. Three other groups of five rats each received the same L-tryptophan treatments; however, in these rats saline was substituted for cocaine and a sweet drinking solution consisting of glucose and saccharin (G + S) replaced water in the automatic drinking device. Two other groups consisting of five rats each self-administered a higher (0.4 mg/kg) or lower (0.1 mg/kg) unit dose of cocaine and food adulterated with 4% tryptophan. At the two higher concentrations L-tryptophan reduced cocaine infusions by at least 50% during the 5 days of treatment, and cocaine infusions returned to baseline levels within 48 h after the regular diet was restored. Responding reinforced by the G + S solution was not altered by any of the L-tryptophan concentrations. Food intake was substantially lowered by the 8% L-tryptophan concentration; however, water intake, responding on an inactive lever, and the number of saline infusions were not affected by addition of L-tryptophan to the food. L-Tryptophan had the same magnitude of effect on self-administration of the 0.1 and 0.2 mg/kg unit doses of cocaine, but behavior maintained by the highest cocaine dose (0.4 mg/kg) was resistant to the effect of L-tryptophan. The results of this experiment indicate that L-tryptophan reduces behavior reinforced by IV cocaine infusions.