This paper reports the results of a sample survey of ZPG members undertaken in April 1971 to ascertain if any changes had occurred in membership attributes since the first survey in September 1970 and to investigate additional characteristics. Age, education, religion and student/ non-student status are considered in relation to length of membership of ZPG. Income, political activities, intended family size and opinions about ideal population size and population control are also examined, including attitudes towards vasectomy and abortion.

It is concluded that, among members of voluntary organizations, demographic factors such as age, education, religion and sex do not possess a causal relationship with social behaviour.

It is well established that, in general, pregnancies at advanced maternal ages or of advanced birth order are at greater risk of aborting spontaneously. It does not follow that each individual woman, as she proceeds through her reproductive life, becomes steadily more prone to abort. The present paper attempts to analyse evidence on how individual women's abortion probabilities vary during their reproductive lives. Statistical examination of the data suggests that:

1. A spontaneous abortion in a given birth rank is more likely than a live birth in the same rank to be followed by a further pregnancy. Abortion-prone women have more pregnancies than other women.

2. There is no significant overall tendency for spontaneous abortions to appear early or late in sibships even when account is taken of the reproductive compensation reported above.

3. Pathological examination of spontaneously aborted fetuses suggests that more (perhaps many more) than 10% of spontaneous abortions (e.g. those associated with trisomy and maternal isoimmunization) are causally associated with advanced maternal age or birth order.

4. The mean maternal age of spontaneously aborted fetuses which fail to grow in culture is higher than that of fetuses which are successfully cultured.

One would infer:

A. (from suggestions 1 and 2 above) that reproductive compensation (rather than a real birth order effect) is a major reason why abortion rates are higher at higher maternal ages and higher birth ranks;

B. (from suggestions 2 and 3 above) that there must be another group of abortions, roughly equal in number, which are causally associated with low maternal age or early birth order.

Little is known about this second group of reproductive casualties (those associated with low birth order or maternal age), but it is suggested that some have CNS malformations. I have made similar suggestions in regard to stillbirth and neonatal death.

An analysis of 96% of Israeli live births, 1968–70, gives a mean birth weight of 3299 ± 543 g with 5·7% weighing less than 2500 g. The mean values increased from 3193 ± 510 g for first born to 3469 ± 604 g for tenth and subsequent births.

There were significant differences between the mean birth weights, adjusted for birth order, of the different ethnic-origin groups. Infants of Jewish immigrants from North African countries weighed 3356 g followed by Christian Arabs (3337 g), Israeli-born Jews (3310 g), immigrants from Western countries (3303 g), Moslem Arabs (3251 g), Druze (3244 g) and immigrants from countries of Asia (3223 g).

For the Jewish groups, the differences were diminished in the second generation of Israeli-born but those descended from Asian immigrants continued to be the lighter group.

In order to determine the extent to which these differences are truly ‘ethnic’ further analyses of more homogeneous groups will be necessary taking into account a number of other demographic and socio-economic variables. (Summary in Hebrew, p. 58.)

The method of decomposition is applied to evaluate the role played by fertility and age composition of women at risk in annual fluctuations in the number of legitimate live births in England and Wales during the period 1956–70. The analysis shows that the upward trend in the numbers of births during the period 1956–64 was mainly due to a rise in the level of period fertility and that the downward trend thereafter has been due to a declining trend in fertility alone. The analysis also shows that the size and age distribution of women at risk have strongly affected the size of birth cohorts born since 1964. Such effects have been in the direction of increasing the numbers of births to women under 25 years of age in a period of falling fertility.

A recent paper published in this journal (Poole & Kuhn, 1973) has reported that ‘a relatively large family appeared to be no impediment to the educational achievements of middle-class children’. This finding, however, is based upon an analysis which fell into a relatively unknown statistical trap. Poole & Kuhn are only the latest in a long line of victims of this trap and some of their predecessors are reviewed below.

In accordance with JBS policy, Mrs Poole was offered the opportunity of replying to Mr Bytheway, for simultaneous publication with his paper. Her comment, which has been seen by Mr Bytheway, appears below. So far as JBS is concerned, the discussion is now closed.

The object of the study was to investigate the genetic nature of personal preferences in such spheres as food, friendship, art, music, ambition, marriage and so on. The subjects chosen for the research were forty-eight pairs of identical and forty-eight pairs of non-identical twins, divided equally into males and females, with ages ranging between 17 and 40 years. Each subject, individually, was given an oral questionnaire (‘Fads and Fancies’), the answers to which were analysed statistically. The results indicate very little difference between the concordance rates of identical and non-identical twins, suggesting that the issues discussed might not be genetically influenced. One possible reason for such findings could be the unrepresentative nature of the sample.

Data are presented on a representative sample of legitimate live-born first infants in England, Scotland and Wales. When they are categorized simultaneously by paternal age, maternal age and social class of the father, it is found that:

(1) Mean parental age difference (father's age — mother's age) is higher in upper class births than in other births.

(2) If a man is younger than his wife, he is probably more likely than otherwise to come from the lower social classes, though this conclusion becomes less secure as the age difference increases in this direction.

(3) Mean parental age difference increases with paternal age and decreases with maternal age.

(4) For a given paternal age, parental age difference increases with a decline in social class, and for a given maternal age, parental age difference increases with a rise in social class.

(5) In general, young parents have smaller mean parental age differences than old parents.

If such mating patterns are characteristic of white North Americans one would infer:

(1) that the risks of neonatal death and stillbirth previously reported in association with aged paternity are the result of biological rather than sociological phenomena, and

(2) that the risks of neonatal death and stillbirth previously reported in association with youthful paternity are the result of sociological rather than biological phenomena.

In general, it seems that mean parental age difference is unlikely to be a fruitful clue to the aetiologies of pathological conditions unless there are simultaneous controls on the age of at least one parent and on social class.

The question is raised whether age disparity between spouses is associated with childlessness as well as with perinatal death and congenital malformations.

Twenty-eight South African Negro albinos have been tested for intellectual maturity and body image boundary, characteristics and a comparison made with carefully matched normally pigmented controls. The albinos were found to be more intellectually mature than the controls but the control group showed slightly more diffuse body image boundary differentiation than the albino group.