Based on 1) neuroanatomical and neuroimaging studies indicating aberrations in brain regions that are rich in glutamate neurons and 2) similarities between symptoms produced by N-methyl-D-aspartate (NMDA) antagonists in healthy subjects and those seen in autism, it is proposed in the present paper that infantile autism is a hypoglutamatergic disorder. Possible future pharmacological interventions in autism are discussed in the light of the intimate interplay between central glutamate and serotonin, notably the serotonin (5-HT) 2A receptor. The possible benefit of treatment with glutamate agonists [e.g. agents acting on the modulatory glycine site of the NMDA receptor, or so-called ampakines acting on the alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptor] is discussed, as well as the potential usefulness of a selective 5-HT2A receptor antagonist.