Variables measured before or during psychological treatment may be used in several related ways: as an empirical guide to prognosis with any given treatment, as a means of studying possible treatment mechanisms, or as a method of identifying which stage or process in treatment is associated with therapeutic change. Using data from the same patients already described in a recently published study (Gelder et al., 1973), questions of relevance to each of these areas will be examined in turn.

The Royal College of Psychiatrists is at present concerned with establishing criteria whereby psychiatric hospitals and posts are to be accredited for training purposes. With regard to the role of the library, it is stated that ‘there should be a medical library adequately housed and suitably fitted. The library should contain sufficient books and subscribe to an adequate number of journals to meet the needs of the trainees. For psychiatric units in general hospitals and for those psychiatric hospitals where a library at a postgraduate centre is within very easy access, the psychiatric sections of the library should contain sufficient books to meet the trainees’ needs’. It is by no means clear, however, what constitutes ‘sufficient’ books and journals, nor what sort of library is ‘adequate’ for the needs of psychiatrists in training.

Nearly all the authors who have concerned themselves with puerperal psychoses have discussed the occurrence of mental illness in the families of the patients. However, as Jansson (1964) has pointed out, there exists no systematic study on the importance of hereditary factors in this disorder.

Sleep disorders are important diagnostic and therapeutic aspects of many common psychiatric syndromes. Clinicians have to be content with the patients' own assessment of their sleep patterns, or at the best, those of a night nurse; both methods are known to be unreliable, particularly where insomnia is present (Kuper et al., 1970). In a previous paper, Johnson and Kitching (1972) described a simple device for assessing the duration of sleep in ward situations. It depended upon the patient's ability to signal wakefulness by pressing a mechanical switch in response to an intermittent light stimulus throughout the night. This was considered to indicate a level of cerebral arousal compatible with wakefulness: failure to respond to the light stimulus indicated that the patient's level of cerebral arousal was lowered to the level of ‘sleep‘. A number of objections were anticipated, however, which might invalidate these assumptions. Most important of these were:

1. (i) The intermittent occurrence of the light stimulus, every 15–20 minutes, limits the device to an approximation of the duration of sleep. The system of scoring was to assume a period of 10 minutes of sleep on either side of a missed stimulus.

2. (ii) The possibility that the patient could press the mechanical switch in response to the light stimulus whilst in Stage I or even Stage II of sleep (Rechtschaffen et al., 1968) without returning to wakefulness.

3. (iii) The intermittent light stimulus might itself act as a disruptive stimulus to sleep, producing arousal and thus altering the sleep level.

Professor Penrose was a human geneticist but above all a human biologist interested in all aspects of man's adaptation and behaviours. However, I need not remind you of the fact that his interests were wider. They were, among others, with the field of biology itself. This interest was one reason why he was particularly delighted with some of the discoveries made on human chromosomes and their anomalies, and he shared very abundantly in the work and in the excitement of these discoveries.

A large part of the data gathered in psychiatric and clinical psychological research consists of ratings carried out by various members of the clinical team. A set of clinical ratings consists essentially of a questionnaire that is completed by someone who is familiar with the patient. Data of this kind provide information about the way in which members of the clinical team think about their patients; but not necessarily about the ways in which the patients themselves actually behave, feel and think. One would of course expect to find a relationship between the two kinds of data, and both are of psychological interest.

The systematic search of aetiological agents from a variety of slowly progressing or subacute neurological diseases has revealed causative viruses or virus-like agents from kuru, Creutzfeldt-Jacob disease and other forms of presenile dementias, subacute sclerosing panencephalitis, progressive multifocal encephalopathy, and from many similar neurological diseases in animals (Gajdusek and Gibbs, 1973; Gajdusek, 1973). The first two diseases called subacute spongiform virus encephalopathies (Gajdusek and Gibbs, 1971) have many interesting features including heredo-familial occurrence and totally non-inflammatory neuropathology. Thus the epidemiology of these diseases is not typical for diseases with infectious aetiology, and the histopathological studies do not suggest the presence of extremely high-titred infectious material in brain cells.

Beginning with the hypothesis that ‘there exists a set of correlated behaviour variables indicative of predisposition to psychotic breakdown, demonstrable as a continuous variable in the normal propulation, and independent of E and N’ (H. Eysenck and S. Eysenck, 1968), the Eysencks have been concerned to develop a questionnaire scale to measure such a variable. From an initial pool of potential items, hypothesized to be related to psychoticism, several factor analyses of adult responses have been conducted to obtain a small group of items measuring a factor orthogonal to the factors of neuroticism and extraversion, and these items have been called the Psychoticism or P Scale (H. Eysenck and S. Eysenck, 1968; S. Eysenck and H. Eysenck, 1968; S. Eysenck and H. Eysenck, 1972). The same arguments have been applied to children, and a similar set of items have been assembled to measure this variable (H. Eysenck and S. Eysenck, 1971; H. Eysenck, G. Easting and S. Eysenck, 1971; S. Eysenck and H. Eysenck, 1969).

This paper reports a study of maternal psychiatric illness in school-phobic adolescents admitted to a psychiatric in-patient unit. These cases have been previously described and studied in various ways (Berg, Nichols and Pritchard, 1969; Berg, 1970; Berg and McGuire, 1971; Berg, Butler and McGuire, 1972; Berg and McGuire, 1974).

For some years we have monitored a steady increase in the rates for parasuicide (attempted suicide) in Edinburgh (Aitken et al., 1969; Kreitman, 1972). Some changes in the character of the group have been demonstrated; in particular, the proportion of young people admitted has increased. The proportion readmitted within 12 months has, however, remained constant at around 15 per cent. Identification of potential repeaters is dependent on the group possessing characteristics which remain consistent over time. In this paper the characteristics of both repeaters and non-repeaters are compared over a three-year period, and stable predictors of repetition are identified. An earlier study (Buglass and McCulloch, 1970) suggested that prediction of repetition was best considered separately for each sex. This question is reconsidered in the light of new data. Thirdly, the interaction between the type of disposal recommended after the parasuicidal episode and subsequent repetition is examined.