During the past decade, an occupational disease surveillance scheme has been created in the UK, based on systematic reporting of newly diagnosed cases by six groups of clinical consultants and by specialist occupational physicians. Labour Force Survey statistics have proved a reasonably satisfactory denominator for the former, but not for the occupational physicians, who provide services for only a selected subsection of the employed population. To remedy this deficiency, approximately 700 occupational physicians who were recorded as having been a reporter at some time were invited to provide their best estimate of the number of employees for whom they were responsible. After various exclusions--mainly physicians who were not, or were no longer, responsible for any defined workforce, and others who had not reported for at least 3 years--the number of active participants for whom data, by industry, occupation and sex, were obtained or estimated was 503. The resulting total number of employees served was estimated at 3.2 million, comprising 12% of the general working population. The proportion with access to an occupational physician varied enormously, from 43% in the health and social services to 1% in agriculture, forestry and fishing, and 6% in the rest of industry. Numbers estimated for each industrial sector were fairly reliable, but by occupation less so, especially in the health and social services.