The professional relationships between family physicians (general practitioners or GPs) and specialist doctors are crucial in a referral-based national health service such as that in the United Kingdom. They may, for example, have a fundamental influence on what in the past have been assumed to be logical or proximity-based utilisation patterns of secondary an even tertiary hospital facilities if GPs refer to specialists with whom they have greater rapport. Until recently, research has identified a generally poor mutual understanding between the two key clinical branches of the National Health Service (NHS), and this research was conducted to determine the state of the current relationship. A range of qualitative methods was employed - including key informant interviews, focus groups, and in-depth interviews - with family doctors and specialists in the south and west region of the United Kingdom. The participants described a much better mutual relationship than that found in previous studies. Four major issues were identified relating to professional interaction. These included the reasons for choosing one of the two main branches of medicine in which to practice, the ways in which doctors interact over clinical issues, the effects of teaching and learning together on the relationship, and possible future models of professional interaction. In process terms, the theme of cooperation was more evident than that of conflict and there was minimal evidence of a struggle for power or status. If anything, the balance of influence within the profession seemed to be shifting in favour of the primary practitioners rather than the specialists. This research therefore indicates key variables for inclusion in future behavioural studies of the use of health services in physician referral-based systems. We illustrate tlie strengths and potential of qualitative methods in health care research and health care geography for investigating professional relationships which are likely to impact on patient care and service quality as well as being factors that might influence spatial patterns of patient behaviour.