In order to produce choreic movements in monkeys (Macaca species), we have used the combination of unilateral kainic acid injections into the striatum and the systemic administration of L-DOPA. Two control monkeys received kainic acid injections alone and 7 were also given L-DOPA (150-300 mg/day) orally. No spontaneous involuntary movements were seen after kainic acid injection alone (over a period of 4-14 days). In the 7 monkeys that received L-DOPA 4-7 days after the kainic acid injection, in 5, whose striatal lesions were located in the rostral dorsolateral striatum, choreic movements appeared in the contralateral limbs following each dose of L-DOPA, whereas in 2 monkeys in which the lesion was large enough to occupy more than 60% of the rostral striatum, no choreic movements occurred. Preservation of some part of the striatum is therefore necessary for the generation of choreic movements. A considerable number of neurons in the rostral ventromedial putamen, which escaped the kainic acid lesion showed burst discharges when L-DOPA was administered. These were not seen in the caudate nucleus. The burst discharges recorded from some neurons were time-locked to the choreic movements. Since deafferentiation by dorsal root section had no effect, these burst discharges appeared to be generated in the unlesioned striatum during the choreic movements. Brains of 6 monkeys were analysed biochemically for gamma-aminobutyric acid concentration, choline acetyltransferase and tyrosine hydroxylase activity, and spiroperidol binding. Concentrations of gamma-aminobutyric acid, choline acetyltransferase activity and spiroperidol binding in the basal ganglia did not correlate with the generation of choreic movements. A finding which appeared to have a clear correlation was an increase in excess of 50% in tyrosine hydroxylase activity in the rostral ventromedial striatum on the lesioned side which escaped the kainic acid lesion in choreic monkeys as compared with the corresponding area on the intact side. It is, therefore, assumed that certain populations of 'activated' nigrostriatal dopaminergic neurons which innervate the unlesioned area of the striatum may be more likely to attain a level of activity, induced by L-DOPA, sufficient to produce choreic movements in the contralateral limbs. Methamphetamine, but not apomorphine, given to 2 monkeys, produced an effect that was essentially the same as that of L-DOPA. These results support the hypothesis that the 'activated' dopaminergic components in choreic monkeys are the presynaptic dopaminergic nerve terminals rather than the postsynaptic dopaminergic receptors.