Some of the literature on the interrelationships between suicidal acts, accidents, surgical operations and other forms of violent experience is reviewed.It was postulated that persons making suicidal attempts would, more commonly than non-suicidal controls, have encountered violent experiences during their life-time. To test this hypothesis, 50 persons attempting suicide were compared with 50 non-suicidal psychiatric patients and with 50 healthy persons attending a chest clinic. The groups were matched for age, sex and social class.Using a questionnaire, all relevant data were recorded. Classes of violent experience were graded numerically on a basis of severity and on the degree of responsibility of the person involved. It was found that the suicidal patients had significantly higher violence scores than either control group, a finding which remained significant when previous suicidal attempts were excluded from the score. The possible implications of this finding are discussed.

Reliability of screening for selection for postgraduate training in clinical psychology was assessed by correlating ratings by four readers of the application papers. Correlations for the 128 applicants were reasonably high. Reliability of the interview procedure used with 29 of the shortlisted applicants was assessed by correlating ratings from four interviewers who worked in pairs. Those who interviewed together correlated significantly but other correlations were disappointingly low. It was shown also that the selection of preferred candidates from the total interviewed did not differ from chance. The results are discussed in terms of the quality of the candidates applying for training and of the reliability and validity of the selection procedure.

SUMMARY The performance of samples of English children at five and eleven years on the Draw‐a‐Man test is assessed by the scoring methods of Goodenough and Harris. Proposals are made for improving definitions of some scoring instructions. Inters‐corer reliabilities are reported. Average scores of the samples are compared to American norms. Relationships between scores on this test and other cognitive and educational tests arc interpreted.

Studies have shown that subjects are not random in their responses, even when attempting t o achieve randomness. In this study, normal and schizophrenic subjects were asked to give numbers from one to nine at random and paced. They were given an interpolated task, and then retested. Some statistical tests of degree of randomness were devised and it was shown that no subject was random, but that stereotypy decreased within and over test-retest. The results are discussed in terms of learning and extinction of stereotypy. Differonces between introverts and extraverts were rtlmost significant in the opposite direction from that predicted. Schizophrenics were significantly more stereotyped than normals. Finally, it was shown that for a normal subject, his test could be identified from other subjects' tests on the basis of retest protocol, the test protocol providing a kind of psychological 'fingerprint'.A number of writers have discussed the inability of human subjects to generate random sequences of choices from a finite set of responses. Tune (1964b) gives a review of the literature on evidence of response preferences, and puts forward the hypothesis that the non-randomness is due to limitations of short-term memory (Tune, 1964a). Weiss (1964) also gives a review and criticizes Tune's hypothesis of a limitation of short-term memory on the grounds that the problems of encoding and storing the necessary frequency information, even in short-term memory, are overwhelming (Weiss, 1965). He states that the use of memory devices may lead to even less success than is usual for most subjects. Weiss further points out that, according to the memory hypothesis, it would be most difficult for subjects' responses to become less stereotyped with time, as more and more information has to be stored. He suggests that in any alphabet of symbols some of the combinations of symbols will be dominant over others, and to randomize successfully a sequence of symbols the subject must inhibit attention to each response after it occurs.The present experiment set out firstly to test whether subjects become more or less random over time, and in this way to test the short-term memory hypothesis proposed by Tune. A second purpose of the experiment was to investigate possible differences in degree of randomness due to personality differences. Weiss (1964) remarked on large individual differences and suggested the need for investigation of the effects of organismic variables. Eysenck (1960) postulated that extraverts build up inhibitory potentials (both reactive inhibition, I,, and conditioned inhibition, faster than do introverts. A paced task such as giving random numbers at a certain speed would result in a build-up of inhibition, and, if there are patterns within the protocols,

Conventionalsaline pad electrodes were held to the scalp by a standard rubber head-cap. Bipolar recordings were made from electrode pairs C@-O@, 02-C4, 01-02 ‘¿ 99