The recent availability of selective ligands for NMDA and AMPA receptors has enabled neuroscientists to test the hypothesis that Parkinson's disease is a glutamate hyperactivity disorder and hence treatable with glutamate antagonists. This review takes a critical look at the motor characteristics of this new class of drugs in rodent and primate models of parkinsonism and assesses the clinical potential and pitfalls of this radical new approach. Monotherapy of Parkinson's disease with glutamate antagonists appears impractical at the present time, due to their low efficacy and unacceptable side effects, but polypharmacy with L-DOPA and a glutamate antagonist as adjuvant is a more realistic prospect. This review will focus on the ways in which glutamate receptor blockade facilitates motor recovery with L-DOPA and will examine whether the basis for this beneficial effect can be traced to a specific interaction with dopamine at D1 or D2 receptors, and therefore to discrete motor pathways within the basal ganglia.