The aim of the present study was to examine more closely than has previously been done some widely accepted generalizations about the ritualistic behaviour of patients suffering from obsessional neurosis. Observation of such patients had suggested that their behaviour is less consistent and is governed by a much greater complexity of factors than is normally assumed. In addition it seemed that such over-simplification has led to the construction of inadequate theoretical models to explain ritualistic behaviour. Moreover it appeared that this over-simplification was linked to the fact that there have been very few systematic studies of the obsessional ritual itself: most descriptions of obsessionals' behaviour seem to be based on the retrospective and rather general accounts, given by the patients themselves in the relative calm and detachment of the consulting-room situation. Observation of patients actually performing rituals suggested that such accounts might well be misleading. Therefore, it seemed likely that it would be profitable to study the rituals of individual obsessional patients systematically and in some detail. In the investigation reported here, three such studies were carried out. The first two, which are reported briefly, were relatively uncontrolled and had as their aim the derivation of more specific hypotheses to be tested in a later investigation. The third study, which is reported more fully, was concerned with testing out these hypotheses by the more rigorous investigation and manipulation of the rituals of a further obsessional patient.