The variation in the range of services provided by general practitioners (GPs) is not only related to personal characteristics and features of the country's health care system but also to the geographical circumstances of the practice location. In conurbations health services are more widely available than in the countryside, where GPs often are the only providers. With highly mobile populations and a plentiful supply of doctors, in cities the prevailing regulations for access and use of services are more difficult to maintain. It is also more difficult to control access and thus opportunities for inappropriate use are greater. Against this background an international study was conducted on variation in task profiles of GPs, especially focusing on differences between urban and rural practices. In 1993 standardised questionnaires in the national languages were sent to samples of GPs in 30 countries. Various aspects of service provision were measured as well as practice organisation, location of the practice and personal backgrounds of the GP. Completed questionnaires were received from 7,233 respondents, an overall response rate of 47%. Sources of variation have been analysed by using a two-level model. Rural practices provided more comprehensive services regardless of the health care system. Approximately half of the variation was explained by features of a country's health care system. The GP's position at the point of access to health care was strongly associated with the gatekeeper function controlling access to secondary care. In western countries where the GPs were self employed they had greater involvement in technical procedures and chronic disease management. There was a considerable gap between the task profiles of GPs in eastern and western Europe. We found evidence of a reduced gatekeeper role in inner cities in those countries where GPs held this position. GPs with an estimated overrepresentation of socially deprived people and elderly in the practice population reported a wider range of services. Differences also appeared to be related to factors which are largely controlled by the individual doctor, such as level of training and education, availability of equipment and practice staff. The results have important implications for education, policy development and health care planning both in eastern and western Europe.