Of patients with ‘conspicuous psychiatric morbidity’ in British general practice, 5 per cent annually are seen by psychiatrists, roughly equal numbers are referred to non-psychiatric specialists (Shepherd, Cooper, Brown and Kalton, 1966), and cross-referral rates are low (Mezey and Kellett, 1971). The clinical, as opposed to the social, reasons why so few patients are referred to psychiatrists must be inferred from the indirect evidence of epidemiological surveys (Shepherd et al., 1966), studies of patients already referred to an out-patients clinic (Kreitman, Sainsbury, Pearce and Costain, 1965), doctors' referral letters (Mowbray, Blair, Jubb and Clarke, 1961), and surveys of doctors' attitudes (Shepherd et al., 1966; Mezey and Kellett, 1971). Few studies of referral have been prospective (Kaeser and Cooper, 1971) and none has been by interview of patients before referral has taken place. Meanwhile, detailed clinical studies of psychiatric illness continue to focus almost exclusively on patients already in care of psychiatrists. It follows that more knowledge is needed of the clinical factors which lead to specialist referral; otherwise the significance of hospital-based studies for the wider context of psychiatry in general medical practice cannot be properly appreciated. This need is particularly acute with respect to the so-called minor affective disorders which are common in general practice but of which only a minority come to the notice of psychiatrists (Watts, 1966; Taylor and Chave, 1964).