A questionnaire designed to screen Parkinson's disease (PD) in literate populations has been developed. It consists of nine questions, self-administered at medical facilities or by mail, and a scale of weights for ascribing scores to specific questions when the answer is positive. The questions were chosen to be symptom specific for PD and the weights were determined from answers provided by 37 PD patients in a neurological outpatient clinic. The questionnaire sensitivity was tested on a different PD population from the same outpatient clinic--50 individuals--and the specificity on a group of 100 ophthalmological patients. The sensitivity was 100% and the specificity was 100%. Three individuals who screened positive among the 100 ophthalmological patients were assessed and given a new diagnosis of PD. This questionnaire therefore constitutes an instrument that should prove valuable as the first stage of a door-to-door survey. It has high sensitivity and specificity.