In a case control study of the relationship between smoking habits and Parkinson's disease a negative association was demonstrated with a relative risk of 0 x 52. A history of smoking up to 20 years earlier was associated with a risk of developing Parkinson's disease equal to about half that in non-smokers. The type of disease, age of onset and rate of progression were associated with a similar reduction in risk implying that in respect of smoking history the disease is homogeneous. The positive correlation of degenerative vascular disease with smoking is further evidence that arteriosclerosis is not involved in the causation of Parkinson's disease. The negative association between Parkinson's disease and smoking is confirmed and is independent of other associated factors. The known and serious risks of smoking far outweigh the possible benefit of lowered risk of Parkinson's disease.