Kainic acid, a dicarboxylic acid containing pyrrolidine, is a potent neuronal depolarizing agent in the mammalian central nervous system, being some 50 times more effective than the putative neurotransmitter L-glutamate1–5. As kainic acid is a structural analogue of glutamic acid, it has been proposed that the potent excitatory effects of kainate reflect its restricted conformation, which is optimal for interaction with excitatory glutamate receptors2,4. More recent studies, however, do not support this hypothesis. In invertebrates kainic acid has been found to be a rather weak agonist at excitatory glutamate receptor sites6,7, whereas at the crustacean neuromuscular junction8, on certain neurones of Helix aspersa9 and in the mammalian cerebral cortex3, kainic acid seems to act at receptors distinct from those which mediate the neuroexcitatory action of glutamate. In addition, the ionic requirements for kainate-induced neuronal depolarization differ from those for glutamate-induced depolarization in the mammalian brain10, and glutamate-induced excitation of rat cortical neurones can be antagonized preferentially by drugs which are very weak in blocking kainate-induced depolarization11. Finally, the characteristics of binding sites for 3H-kainic acid in rat brain are consistent with the existence of specific kainate receptors on neuronal membranes12. Thus, the kainate sites may represent receptors for an unidentified endogenous substance probably containing the glutamate structure. Here we report that the pteroylmonoglutamate compound methyltetrahydrofolate (MTHF) is a potent competitor for 3H-kainic acid binding sites in rat cerebellar membranes. This action is specific in that MTHF is very weak or inactive at sites for 3H-glutamate or 3H-γ-aminobutyric acid (GABA) respectively. Electrophysio-logical studies in frog spinal cord suggest that MTHF is an agonist at the kainate receptors. The implications of these results are discussed.