A programme was undertaken to study the attitudes of visitors to mental hospitals. Initially a series of individual non-directive tape-recorded interviews were conducted on a small sample of visitors, who freely discussed various topics, including staff attitude, visiting problems and difficulties in having a mentally ill relative. A provisional questionnaire was constructed from the significant points of the recorded interviews and was used on another sample of visitors in a pilot study, on the basis of which a final ‘multiple choice’ questionnaire was designed incorporating 28 items. Copies of the final questionnaire were then distributed from the wards of two mental hospitals to consecutive visitors, with a stamped envelope addressed to Nottingham University, informing them that the study was in connection with an independent survey. The distribution continued until few new visitors attended. Of the two hospitals, the first, The Pastures (Hospital A), serves a wide county catchment area and has an ‘unrestricted visiting hours' policy; the second, Kingsway (Hospital B), on the other hand, has a ‘limited visiting hours' policy and serves a compact urban catchment area. Altogether 510 questionnaires (75 per cent) were returned completed, 306 (8o per cent) for Hospital A and 204 (70 per cent) for Hospital B at the end of the study. The rest of the paper is confined to the analysis of these 510 questionnaires in relation to the question on the effect of unrestricted visiting, which has already been studied by Barton et al., from the nurses' point of view. The particular question asked by us was, 'some hospitals allow visitors to visit at any convenient time. What effect do you think this would have on (1) the patients, (2) the visitors, (3) the nursing staff, (4) the doctors?’ The replies were to be given on a five point scale for each question. The scale was ‘Very good-Good-No effect-Poor-Very poor.’ the graded adjectives being determined from the earlier tape-recordings.