Abstract The therapeutic effects on parkinsonism of a combination either of L-dopa (levodopa) and L-tryptophan, or of levodopa and a placebo, were compared. The trial was carried out on forty parkinsonian patients. Psychological assessments before the trial revealed little conspicuous psychiatric morbidity, and, in particular, little depressive affect. " Validity ", a dimension of personality measured by the Marke-Nyman temperament scale, which is related to energy and drive, was lower than normal. Levodopa produced a striking improvement in the ratings of tremor, rigidity, akinesia, gait, and posture equally in patients treated with this compound alone or in combination with L-tryptophan. Tests of functional ability to do certain tasks tended to improve in all patients but showed a significant improvement only in patients treated by the combination of levodopa and L-tryptophan. There was no deterioration in the mental state in either treatment group, and the patients treated by the combination (but not by levodopa alone) showed a small but significant improvement in mood and drive. Cerebrospinal-fluid concentrations of homovanillic acid and 5-hydroxyindoleacetic acid were estimated in fourteen patients before treatment. The levels of both compounds were low before starting treatment, but there was no correlation between the pretreatment levels of these compounds and the baseline neurological and psychological scores, nor with the patients' subsequent response to treatment.