Psychiatry and other medical disciplines have been paying increasing attention in recent years to the relationship between patient and doctor, and psychoanalytic concepts have been applied in order to formulate various aspects of this relationship. One of those most commonly applied is that of transference, which is now often loosely used as a synonym for relationship in general. The concept of transference will be discussed in detail in a further paper in this series, but a distinction has always been made within clinical psychonanalysis between transference ‘proper’ and another aspect of the patient's relation to the doctor which has been referred to in recent years as the ‘therapeutic alliance’, ‘working alliance’ or ‘treatment alliance’ (Zetzel, 1956; Loewald, 1960; Stone, 1961, 1967; Gitelson, 1962; Tarachow, 1963; Greenson, 1965, 1967; Friedman, 1969). The concept has been used in connection with the psychoanalytic treatment situation to describe aspects of what is familiar to all as the ‘therapeutic contract’ between the patient and his doctor (Menninger, 1958), and has been defined as ‘the non-neurotic, rational, reasonable rapport which the patient has with his analyst and which enables him to work purposefully in the analytic situation …’ (Greenson and Wexler, 1969). As far as the psychoanalytic situation is concerned, it has been maintained that the recognition of the difference between treatment alliance and other aspects of the patient-analyst interaction (such as transference) leads to an increased understanding of the processes which occur in that situation and, in particular, of those which enter into therapeutic failure. It is the purpose of this paper to examine the treatment alliance concept as it has developed within psychoanalysis and to consider its applicability outside the psychoanalytic treatment situation (Sandier, Dare and Holder, 1970).