Upright stance and its reflex control were studied in parkinsonian patients and in age-matched normal subjects. They stood unperturbed on a force-measuring surface (static conditions), or were displaced by movement of a supporting platform (dynamic conditions). During quiet stance the following variables were analysed, with eyes open or closed: position of the centre of foot pressure (CFP), average sway area, length of sway path, amplitude and distribution of tonic leg muscle EMG activity. Perturbations of stance were induced by toe-up or toe-down rotations, and by backward or forward translations of the platform. Amplitude of short, medium and long-latency EMG responses to displacement were measured in the tibialis anterior (TA) and in the three heads of the triceps surae (TS) muscle. The perturbations were produced during both free and supported stance (holding onto a stable structure), under which condition normal subjects suppress medium and long-latency responses. Under static conditions, the only significant finding in parkinsonians was represented by a shift in the position of the CFP. This was correlated with the severity of the disease (Webster scale), the less affected patients being shifted backwards and the more affected patients forwards, with respect to normals. Under dynamic conditions, the reflex responses to perturbations of free stance were similar in both groups. Only the medium-latency burst of gastrocnemius lateralis and the long-latency burst TA evoked by TS stretch were larger in parkinsonians. The amplitude of these responses, as well as of all the others, was not related to the Webster score. Within the patients' group, a relationship between position of CFP and area of EMG burst was found for both medium and long-latency TA responses evoked by forward translation and toe-up rotation, respectively. Under supported conditions, the capability to suppress all medium and long-latency muscle responses to any perturbation was lost or impaired in the parkinsonians. The degree of impairment was unrelated to the position of the CFP, but was significantly related to the severity of the disease. The suppression to 40% (supported/nonsupported), of TA response to toe-down rotation is proposed as the point of separation between normals and parkinsonians. The forward projection of the CFP, occurring in the severe stages of the disease, and the increase in amplitude of some responses to perturbations of free stance might be a compensatory adaptation to the anomalous upright posture.(ABSTRACT TRUNCATED AT 400 WORDS)