We studied the relative benefits of carbidopa/levodopa (25/100 mg and 50/200 mg) and propoxyphene (100 mg and 200 mg) in six subjects with periodic limb movements (PLM) using a randomized, double-blind, placebo-controlled, cross-over design with a 4-day placebo wash-out period between test medications. Each subject received, for successive 2-week periods, 4 days of placebo followed by 5 days of low-dose and then 5 days of high-dose medication. Outcome was measured with polysomnograms and sleep latency tests on the last night of the high dose, and leg activity was monitored for each night of the study. Carbidopa/levodopa normalized PLM and improved sleep, particularly in the first 3 hours in most subjects. Propoxyphene had marginal benefit and significantly decreased arousals and mildly decreased leg movements for leg activity monitoring, but not for PLM on the polysomnograms. Subjects reported better sleep and alertness on carbidopa/levodopa than on propoxyphene or placebo. Our results support the premise of a central dopaminergic abnormality in PLM.