Body sway was studied at various body inclinations, voluntarily maintained for about 1 min, in young and elderly normals and in idiopathic parkinsonians. They stood on a dynamometric platform, whose output gave the instantaneous centre of foot pressure (CFP), its mean value and body sway area, with eyes open (EO) or closed (EC). Subjects held the normal upright stance, or the maximum possible inclined posture (body straight, rotated at the ankle joints) in forward or backward direction, or intermediate postures. EMG was recorded from tibialis anterior (TA), soleus (Sol), extensor digitorum brevis (EDB) and flexor digitorum brevis (FDB). The cross-correlation function between the profile of the EMG envelope and the profile of the shift of CFP along the sagittal plane was calculated. In young subjects standing with EO, the maximum extent of antero-posterior (A-P) displacement of CFP was about 60% of foot length. EC reduced this value to about 50%. In the elderly normals, the maximum A-P displacement was about 40% (EO) and 30% (EC). In both groups, sway area was minimal during normal stance with EO and increased progressively when the subjects leant forward or backward. With EC, sway area further increased during normal stance and the rate of increase in relation to inclination augmented markedly. Sol was tonically active during normal stance. Forward leaning increased Sol EMG and induced activity in FDB. TA and EDB were active during backward leaning. The peak of the cross-correlation function between Sol EMG and instantaneous CFP was higher during normal stance than forward inclination, while the reverse was true for FDB. This suggests a role of FDB in the fine-tuning of postural adjustment during forward leaning, and a weight-supporting role of Sol. During backward inclination, TA but not EDB was cross-correlated with CFP. In the parkinsonians, maximum A-P displacement of CFP was just about 30% of foot length (EO; about 20% with EC); its extent was inversely correlated with the severity of the disease. The relationship between sway area and A-P displacement was similar to the elderly, both with EO and EC, within the common range of inclination. In the patients affected by the long-term syndrome, A-P displacement was further reduced while sway area increase at the critical postures was often absent. In all patients, the relationship between muscle activity and body inclination was comparable to normal.(ABSTRACT TRUNCATED AT 400 WORDS)