This study examined the pharmacokinetics of levodopa and carbidopa in the rat after different modes of administration. The drugs were given simultaneously by the intravenous, intraarterial, oral, duodenal, and intraperitoneal routes, as single doses. The ratio of levodopa to carbidopa given was always 4:1. Two iv doses (5 and 15 mg/kg of levodopa) were given to test for nonlinearity. Three ip doses of levodopa were given (5, 7.5, and 15 mg/kg), and the 15 mg/kg dose was given in three volumes (2, 4, and 20 mL/kg). One oral dose and two intraduodenal doses of 15 mg/kg were given. The drugs were dissolved in saline in one of the intraduodenal doses and suspended in 1.8% methylcellulose in the other. The elimination of levodopa was nonlinear. There was a comparatively high degree of interindividual variability in absorption with the oral route, but this was substantially reduced when levodopa was given intraduodenally. There was also much less variability with the intraperitoneal route compared to the oral, and the degree of absorption was generally high. There was a significantly higher extent and slower rate of absorption when levodopa was administered ip in a large volume of vehicle. These results suggest that the oral route may not be the optimal method of delivering levodopa to patients who have a fluctuating response and that a continuous delivery system via the intraperitoneal or intraduodenal routes might be a better alternative.