The treatment of neurotic emotional responses and behaviour patterns has probably generated more controversies and fewer verifiable observations than any other aspect of psychiatric practice. Tentative hypotheses have been all too quick to take root, expanding into comprehensive theoretical systems, and often obscuring the growth of objective information. Lately, however, a more cautious and empirical attitude appears to be gaining ground. For example, Marks (1971) has pointed out that workers in this field increasingly recognize the presence of factors influencing outcome which are not explained or even contradicted by those concepts and strategies which underlie the techniques employed. Psychotherapists of both behaviourist and psychodynamic persuasions are beginning to express an awareness of the limitations and dangers of a too rigid theoretical approach. Such a movement towards uncommitted empiricism seems a healthy trend. This may be furthered by the more careful examination of specific treatment regimes applied to a variety of diagnostically homogeneous groups of patients, an approach which is clearly preferable to sterile arguments about their relative overall effectiveness in the usual heterogeneous collections of neurotic patients with which the literature abounds. In the recently published Handbook of Psychotherapy and Behavior Change (1971), the editors comment on the frequency with which their contributors criticize the lack of replicated studies and the common failure to describe in adequate detail the specific features of patient, method and therapist. From whatever aspect it is viewed, psychotherapy emerges as a complex personal interaction containing many components, all difficult to quantify and unlikely to combine their effects in an easily predictable way.