Parental age and birth order in homosexual patients: a replication of Slater's study

Parental age (at patient's birth) and birth order were studied in 623 male and 89 female patients diagnosed as homosexual, who were born in England and Wales and first attended the Maudsley Hospital between the years 1961-1975. For male patients, mean parental ages were significantly higher than expectation, and mean birth order was significantly later. Regression analysis indicated that the raised fathers' age was probably of more aetiological importance than the raised mothers' age. The results confirm those of previous studies. The female patients did not show a raised parental age or late birth order.

A selective model for electrophoretic profiles in protein polymorphisms

In a previous paper the theory of a model of electrophoretic profiles due to Ohta & Kimura was considered. This model assumes a finite population with a linear series of possible alleles with mutation between nearest types but no selection. In the present paper a model with both mutation and selection is constructed which results in a stable population distribution closely fitting empirically observed features of the Ohta–Kimura model. The problem of discriminating between selective and non-selective models for electrophoretic models is considered.

Are there two maternal age groups in Down's syndrome?

Penrose and Smith (1966) have reviewed the literature on Down's syndrome in great detail, and this has been followed by an important recent review by Richards (1973). In Chapters 10 and 11 of Penrose and Smith's book they discuss the remarkable frequency distribution of the ages of mothers of patients, compared with that of the general population at the corresponding place and time, and they summarize the large number of studies made on this subject. The mean age of the mothers is shifted upwards by amounts which vary in different countries from about 6 to 8 years. The remarkable feature, however, is that there appear to be two bumps in the curve. These are usually (but not always) not large enough to make the curve bimodal, and J. B. S. Haldane therefore coined the term ‘bitangentiality’ for this phenomenon, which appears in most published studies and in the group of all sample cases (9,441) given by Penrose and Smith, Fig. 76. Collmann and Stoller (1962) make a complete survey of all mongol births in Victoria, Australia, from 1942 to 1957 and here there is a distinct bimodality.

The estimation of psychological genetic differences between classes and races

For genetic characteristics controlled by a single locus which are such that the degree of dominance is not dependent on the environment, statistical methods of determining the mode of inheritance and the gene frequencies are now well understood in principle. If, however, there is incomplete penetrance, depending on the environment (taken in the wide sense), the problem becomes much more difficult, especially in human genetics. The situation is even worse if the character is dependent on several or many loci and such characteristics are often very dependent on environmental factors. This seems to occur in man in some psychiatric diseases and is certainly the case for intelligence. The genetical analysis of such phenomena is very difficult, and it is essential that the mathematical and statistical bases of such investigations be carefully understood by anyone who undertakes such work.

Familial effects in schizophrenia and homosexuality

This paper begins by reviewing the problem of determining whether there is a genetic factor in schizophrenia and homosexuality by considering the problem explaining the resulting apparently stationary polymorphisms. By analysing unpublished data on homosexuals it is concluded that the latter can not be explained by a genetic factor. The implications of the observed increases in parental age and parental loss in these diseases are then discussed, together with their implications for the genetic hypotheses.

The schizophrenia-like psychoses of epilepsy: relation between ages of onset

In a series of papers, Slater, Beard and Glithero (1963) reported on a group of 69 patients who had suffered from a schizophrenia-like psychosis at some time after the onset of an earlier but continuing epilepsy. They gave a number of arguments, which need not be considered here, to support the view that the epileptic and the schizophrenic states were not independent of one another and merely coincidental, but that the pre-existing epileptic state was the direct cause of the schizophrenia-like illness, and that this could in fact be regarded as a special form of an epileptic psychosis, which took on for a time the appearance of a (symptomatic) schizophrenia. One of the arguments on which reliance was placed was that there was a non-random relationship between the ages of onset of the epilepsy and of the schizophrenia, such that the interval between the two varied about a mean of 14 years. The three clinical groups into which the 69 patients were classified had significantly different mean ages of onset of epilepsy of 20·5, 16·3 and 9·3 years respectively; but there was no significant difference between the three clinical groups in respect of the mean interval between onset of epilepsy and onset of psychosis, 15·6, 14·0 and 12·8 years respectively. In fact, in the case of the duration of epilepsy, with an F value of 0·24, between-group variance was only a fraction of the within-group variance.

Parental loss in homosexuals

In a previous paper (Abe and Moran, 1969) we analysed observed increases in paternal and maternal ages in 291 male homosexuals admitted to the out-patient or in-patient departments of the Bethlem Royal and Maudsley Hospitals during 1949–1960. It was shown that the mean mothers’ age of these patients was 2.31 years greater than would be expected in the general population after taking account of the years of birth of the patients. The similar shift in mean paternal age, using an indirect method of calculating the expectation, was between 3.15 and 3.55 years. On carrying out an approximate regression analysis of each of these shifts on the other it was shown that the shift in maternal age could be completely explained as an effect due to a shift in paternal age, but not vice versa.

Parental age of homosexuals

During the past two decades an increasing number of investigations have been made into the relationship between maternal age and incidence of disease. The following may be mentioned by way of illustration.

Maternal age and parental loss

Maternal age at parturition, birth order, and parental loss have all been extensively investigated as possible causal or concomitant factors in a wide variety of mental disorders. Thus, to take an extreme example, the influence of maternal age on one of the two main kinds of mongolism is well known and can hardly have other than a biological explanation, as is also true of a number of other causes of mental defect whose incidence is known to increase with maternal age (Penrose, 1963); and similarly much is now known of the biological effects of maternal age in animals (for a survey see Parsons, 1964).

The establishment of a psychiatric syndrome

Recently several papers have appeared which apply the techniques of multivariate analysis in attempts to show that patients suffering from depression can be divided into two groups, often called “endogenous” and “neurotic”, which tend to have different groups of symptoms and different outcomes with E.C.T. or other treatment. The purpose of the present paper is to look at this problem from the point of view of a statistician in an attempt to clarify the logic involved in the various procedures.