Ancient and Renaissance physicians implicated the soul's imaginative faculty in the genesis and remission of disease. The theory stated that images of objects of desire or aversion aroused emotions, which in turn set up humoral imbalances, disturbing digestion and other vital functions, culminating in various forms of pathology. This concept explained the therapeutic efficacy of ‘placebos’ in forms such as spells, charms and talismans. The theory merits close attention by contemporary physicians and researchers.

The life and the ideas of a mentally aberrant figure in mid-seventeenth century England are presented in relation to the sociocultural background of the time and the available clinical information. The material illustrates the need to draw on both sources when applying psychiatric concepts to particular historical epochs.

Kanner's descriptions of the children showing the abnormal behaviour pattern he observed and named 'early infantile autism’ are summarized. E. A. and N. Tinbergens' ethological theory of the aetiology of this syndrome is outlined and criticized. The major problem in evaluating this theory, apart from the absence of any evidence in its favour, is the lack of precision with which the Tinbergens use the terms ‘autism’ and ‘Kanner's syndrome’.

Schizophrenia may be associated with increased prostaglandin synthesis in certain parts of the brain. This hypothesis is based on the following findings: (1) Catalepsy, which is the nearest equivalent in animals to human catatonia, develops in cats when prostaglandin E1 is injected into the cerebral ventricles and when during endotoxin or lipid A fever the prostaglandin E2 level in cisternal c.s.f. rises to high levels; however, when fever and prostaglandin level are brought down by non-steroid anti-pyretics which inhibit prostaglandin synthesis, catalepsy disappears as well. (2) Febrile episodes are a genuine syndrome of schizophrenia.

The Editor has kindly offered us space for replying to the paper by Wing and Ricks, in which they reject, and advise their readers to discard, our views on early childhood autism. Rather than using this space for a ‘counter refutation’ we prefer to call attention in this journal to what, after careful study of Wing's and Ricks' criticism, we still consider a potentially fruitful approach to the problems posed by the autistic deviation. We take this course because we feel that our critics have missed the point of our publications. As we see it, the disagreement is not so much a matter of facts per se as one of how to look at the problems. The primary aim of what we have written (as was indicated, for instance, in the subtitle of our joint paper of 1972) has been to sketch a method of approach, to illustrate its potential, and to advocate the application of this approach in child psychiatry. The method has been developed in studies of animal behaviour, where it has led lo valuable insights. But we have not simply extrapolated the results of animal studies. We have looked at certain details in the behaviour of the children themselves. When we applied this ‘ethological’ approach to autistic behaviour, we arrived at certain, naturally tentative conclusions which seemed, and still seem to us, to have the merit of being plausible interpretations of quite a number of hitherto puzzling parts of the syndrome.

The planning of health services consists of a process that generally involves the following steps: (a) situational analysis, or the description, definition and statement of the problem, its characteristics and dimensions in relation to population and time; (b) the formulation of alternative tactical approaches to the handling and solution of the problem; (c) decision analysis or the selection of a plan; (d) discussion and implementation of the plan selected; (e) evaluation of the results achieved in relation to the problem, situations or populations concerned. This paper discusses the content of a programme of mental health services research for collecting and analysing the information needed to apply these processes to the planning of mental health services, monitoring the manner in which the plan is being implemented, and assessing its effectiveness in achieving short-term and intermediate objectives and long-term goals.

Statistical and epidemiological information play an important role in these processes, particularly in the situational analysis and the evaluation processes. Illustrations have been provided of types of data that are produced in the national mental health statistics programme in the United States. Difficulties in using such data to answer questions concerning the needs for mental health services, and manpower requirements for delivering services to meet these needs are discussed. In many instances, currently available data are quite inadequate for answering key issues such as these, plus others related to living arrangements of the population, the effect of services on the persons who receive them, their families and the communities in which they live. Thus, much still remains to be done to develop systematic, comparative morbidity statistics on the incidence, duration and prevalence of mental disorders in the general population, on the needs for mental health services, and on the effectiveness of our efforts to prevent disorders that can be prevented, and to reduce the amount of disability and distress caused by those that cannot be prevented or terminated.

A series of problems have been identified, the solutions to which would assist materially in providing data that would narrow gaps between available knowledge and that which is needed. The importance of establishing well-staffed research units at the catchment area level with stable funding to accomplish this is underscored. It is a matter of the greatest urgency that adequate resources - financial, manpower, scientific and administrative - be made available to solve these problems. If this is not done, then efforts to document quantitatively the effectiveness of programmes to prevent and control mental disorders will continue to suffer from many of the same shortcomings that have impeded our past and continue to impede current efforts to accomplish this.

The increasing interest in the treatment of phobias underlines the need for clarifying their definition and classification. A brief historical note highlights some of the difficulties which have been, and still are, encountered.

Serum luteinizing hormone(LH) responses to a course of clomiphene citrate were studied in eleven patients with anorexia nervosa at different stages of the illness. In malnourished patients basal levels of LH were invariably low. With the resumption of a normal weight a small but definite rise in LH levels was observed but this spontaneous response to weight gain was variable in that many patients continued to exhibit abnormally low LH levels.

The response to clomiphene in terms of a rise in basal LH levels after administration of the drug, followed by a second peak of LH and subsequent menstrual bleeding, was clearly dependent in part on the patient's nutritional state. In the malnourished state the response to clomiphene was usually either absent or incomplete. After the resumption of a more normal weight, the patients invariably showed an initial rise in LH after the clomiphene, but the second LH peak and subsequent menstruation were frequently not demonstrated. Six patients maintained a normal body weight for at least six months after a course of clomiphene, but only three of them resumed cyclical menstrual bleeding. It was concluded that factors additional to the nutritional state contribute to the prolonged amenorrhoea in anorexia nervosa and that clomiphene appears to have only a limited role in the treatment and management of patients with the disorder. Some aspects of current knowledge of the endocrine mechanisms that regulate normal menstruation and of the mode of action of clomiphene are outlined.

The results of the present study are discussed against this background in an attempt to elucidate further the hypothalamic disorder underlying the amenorrhoea in anorexia nervosa.

Histories of depression-related symptoms were obtained from 3845 randomly selected adult residents of Kansas City, Missouri, and Washington County, Maryland. Depressed persons were slightly more common in Kansas City than in Washington County but within the latter area no urban–rural differences were observed. More depressed persons were found among blacks than among whites. Slightly more white females than males were depressed; no significant differences were found between black females and males. After adjustment for the effects of other independent variables, the probability of having symptoms of depression was highest among persons who were young adults, unmarried, not employed outside the home, poorly paid, and not well educated.

Fifty patients attending a neurological outpatient clinic for Parkinson's disease were assessed by standardized methods for both physical and psychiatric symptoms. The patients then received treatment with L-dopa, L-dopa with carbidopa or anticholinergic drugs and/or amantadine. During the following six-month period the subjects were assessed at intervals, both physically and psychiatrically. Forty patients were followed up for the full six-month period. The severity of physical signs and affective symptoms was shown to be significantly related at several stages of the investigation. Initially, the patients showed a high psychiatric morbidity. During treatment, 22 patients developed a depressive disorder, 12 of which had a history of previous depressive episodes. By contrast, of the 11 patients who showed very few affective symptoms during follow-up, none had a history of depression. Of the 22 patients with a depressive disorder, only two were in the anticholinergic/amantadine group, compared with nine and 11 in the other groups. L-dopa was not an effective antidepressant agent. The probable relevance of the findings of the study to the management of patients with Parkinson's disease is outlined.

Statements about prevalence, of the type ‘x% of the general population are mentally ill’, imply a concept of prevalence that is difficult to define operationally. It requires the specification of a cut-off point, a level of severity from which the presence of an illness can be inferred, and this is largely arbitrary. A more generally useful type of statement compares the frequency distribution of severity for declared cases with that for the rest of the population. This avoids the need for a defined cut-off point. We are still left with the problem of defining a dimension of severity and a declared case, but these should follow easily from a clear statement of the purpose of any specific inquiry. The argument is illustrated from the results of a study of self-referrals to general practitioners.