The paper is based on the author's epidemiological studies of sensorineural hearing loss in Ghana, Jamaica, Nigeria and the United Kingdom. Basically, the method has been one of direct examination of random or total samples. Such a study is more difficult than a study of conductive hearing loss because there is no qualitative difference between sensorineural hearing loss or of normal hearing. These epidemiological, studies therefore, resolve themselves into measuring the hearing levels of ears which have been otologically screened to exclude conductive hearing losses, calculating values for 'average' hearing levels and their dispersion and determining factors which appear to influence these levels.Whereas in European and North American populations, noise and 'ageing' appear to be the only factors which influence these average values, in other populations, other factors emerge. In particular, there are dietary factors and the influence of systemic disorders becomes apparent.The evidence from epidemiology i...