A study is presented which fails to replicate a recent report that peak years of birth of patients later developing Parkinson's disease are related to the influenza pandemics of the period 1890-1930. The years of birth of a whole population cohort of 243 patients suffering from Parkinson's disease examined in Aberdeen in 1983 and reexamined in 1986/7 were compared with deaths due to influenza in the City of Aberdeen in the years 1900-1930. Although a significant peak of Parkinson births (compared with the age profile of the Aberdeen population in 1983) occurred in 1902, there appeared to be no systematic relationship between Parkinson births and influenza deaths. In addition, no season of birth effect could be detected in a comparison with 232 matched controls. The presence of peaks of birth years, for whatever aetiological reason, is of significance to epidemiological studies in that prevalence estimates may be influenced by the year of study relative to these mini-cohorts.