Stretch reflexes were elicited in flexor carpi radialis (FCR) of healthy subjects and patients with Parkinson's disease by forcible ramp and hold extensions of the wrist joint. Individual patients were studied off treatment when rigidity was detected clinically at the joint and throughout the clinical response to anti-parkinsonian medication that abolished or reduced their rigidity. In this way the possible effects of inter-subject variability upon the relationship between reflex behaviour and rigidity were eliminated. The long-latency (M2) stretch reflexes of the patient group were increased on average compared to those of healthy subjects. However, in the large majority of individual patients there were no significant correlations between the amplitudes of their M2 or total (short-latency (M1) + M2) reflex activities, recorded off and on treatment, and the accompanying changes in clinically assessed rigidity. These results suggest that parkinsonian rigidity cannot be uniquely attributed to the increased reflex responsiveness measured by the present laboratory techniques. However, the techniques used to test reflex function in our study differed in several respects (e.g., background activity, stretching wave form) from those employed during clinical assessment of rigidity so that the balance of reflex mechanisms may have varied in the two situations. Therefore, these results cannot be taken as definitive evidence against a reflex origin of rigidity.