1. Normally during walking, the heel strikes the ground before the forefoot. Abnormalities of foot strike in idiopathic Parkinson's disease may be amenable to therapy: objective measurements may reveal response which is not clinically apparent. Occult changes in foot strike leading to instability may parallel the normal, age-related loss of striatal dopamine. 2. The nature of foot strike was studied using pedobarography in 160 healthy volunteers, aged 15 to 91 years. Although 16% of strikes were made simultaneously by heel and forefoot, there were no instances of the forefoot preceding the heel. No significant effect of age on an index of normality of foot strikes was detected (P greater than 0.3). 3. The effect on foot strike of substituting placebo for a morning dose of a levodopa/carbidopa combination was studied in a double-blind, cross-over trial in 14 patients, aged 64 to 88 years, with no overt fluctuations in control of their idiopathic Parkinson's disease in relation to dosing. On placebo treatment there was a highly significant (P = 0.004) reduction in the number of more normal strikes, i.e. heel strikes plus simultaneous heel and forefoot strikes. The effect appeared unrelated to the corresponding difference between active and placebo treatments in plasma concentration of levodopa or a metabolite of long half-time, 3-O-methyldopa (3OMD). However, it correlated negatively (P less than 0.05) with the mean of the 3OMD concentrations. 4. It appears that some abnormalities of foot strike due to Parkinson's disease are reversible. Employing test conditions, designed to provoke abnormalities of foot strike, might be useful in screening for pre-clinical Parkinson's disease.