The possibility that early nutrition has long term consequences in man has been much debated. There have been limited opportunities to perform formal randomised studies on the effect of early nutrition in man and many studies have been flawed by problems with study design. Infants born preterm are a special group. At the start of this study in 1982 evidence on which to base choice of diet was inconsistent and related only to short term outcome, and diets available for such babies differed greatly in nutrient content. In this group it was both ethical and practical to conduct a formal, randomised trial of early diet and outcome and the results were clearly needed for management decisions.

A long term prospective outcome study was undertaken on 926 preterm infants randomly assigned to the diet received in the neonatal period. Surviving children have been followed at 9 months, 18 months and now 7½–8 years of age. The findings suggest that children fed a nutrient supplemented preterm formula perform better than those fed a standard formula milk, and also that human milk may contain factors which promote brain growth or development. Outcome data from the randomised trials show that a very brief period of dietary manipulation (on average for the first 4 weeks of life) influences later development.

Using a cross-sectional sample of 1147 urban adolescents, aged between 14 and 20 years, the variability of some menstrual cycle indicators was related to weight loss. Nearly 40% of the adolescents tried to lose weight and the results showed that attempting to lose weight is significantly associated with increased prevalence of menstrual irregularity and menstrual pain. This finding is independent of body mass index.

This paper presents an ecological analysis of the relationship between infant mortality and economic status by race in metropolitan Ohio, using census data on mother's residence, and economic status determined by the percentage of low income families living in each area. White–non-white comparisons for total infant mortality are examined for the US censuses of 1960, 1970, 1980 and 1990; and more detailed period- and broad cause-specific rates are presented for 1990. A pronounced inverse association is found between income status and infant mortality for whites, but not for non-whites. Non-white post-neonatal death rates were higher for the lowest income area, but for neonatal mortality, total infant deaths, and exogenous and endogenous cause-specific death rates, there was no discernible socioeconomic differential. It is concluded that low income whites and non-whites at all income levels have infant mortality rates that are substantially higher than the overall rate for the population. Policy implications are discussed.

The study examines the impact of body build on differential fertility patterns in 93 !Kung San and 85 Kavango females from northern Namibia. In both groups shorter females had more offspring, but also more dead offspring, than taller females. The interaction between body weight and fertility was different in both groups and may be due to different nursing behaviour. There was a positive relationship between the typical pedomorphy and fertility in the! Kung San, and directional selection processes may be important.

The tertiary sex ratio, as measured by the masculinity index, was examined in four regions of the Province of Jujuy in north-west Argentina over a period from 1869 to 1991. The results showed that inter-regional, as well as age group differences in MI existed which could be related to changing social and economic conditions in this area over the last two centuries.

South Korea and Cuba are dissimilar in religion, economy, culture and attitudes toward premarital sexual relations. In 1960, Korea instituted a national family planning programme to combat rapid population growth. Cuba explicitly rejected Malthusian policies, but made family planning universally available in 1974 in response to health needs. Both countries have undergone rapid fertility declines and today have less than replacement level fertility. Both countries have also used a similar mixture of methods, including a high prevalence of female sterilisation. Abortion has played a major role in the fertility decline of both countries, rising in the first half of the fertility transition and then falling, although remaining a significant variable in the second half. It is concluded that access to contraception, voluntary sterilisation, and safe abortion has a direct impact on fertility and has been associated with a rapid fall in family size in two very different countries.

An obvious requirement before embarking on molecular genetic investigation of a trait is prior evidence from ‘classic’ genetic studies that there is indeed a genetic component. Many behavioural traits are familial and these range from comparatively uncommon single gene disorders such as Huntington's disease which has a typical mendelian dominant pattern of transmission, to much commoner characteristics such as career choice or religious denomination which, it might be assumed, are heavily influenced by cultural factors. In between, there is a wide range of attributes including personality type, cognitive ability and liability to common disorders such as depression, that show a tendency to run in families, and which could conceivably be explained by shared genes, shared environment or a combination of the two.

This study uses Bongaarts' model to examine the relative contributions of three proximate determinants (non-marriage, contraceptive use and postpartum infecundability) to fertility change using data from the 1984 and 1988 Botswana Family and Health Surveys. Breast-feeding is shown to be the most important proximate determinant of fertility, followed by contraceptive use, and finally non-marriage, both in 1984 and 1988. However, contraceptive use increased between 1984 and 1988, leading to fertility decline over this period. Marriage is the least important proximate determinant of fertility, probably due to the high prevalence of premarital childbearing. Other factors such as induced abortion could have played a major role in the fertility decline but their effect could not be estimated due to lack of accurate data.

This paper examines the effects of exposure to mass media messages promoting family planning on the reproductive behaviour of married women in Nigeria using cross-sectional data. Longitudinal data are also used to ensure that exposure to media messages pre-dates the indicators of reproductive behaviour. Cross-sectional analysis suggests that: (1) contraceptive use and intention are positively associated with exposure to mass media messages, and (2) women who are exposed to media messages are more likely to desire fewer children than those who are not exposed to such messages. Similarly, analysis of the longitudinal data shows that exposure to mass media messages is a significant predictor of contraceptive use. Thus, exposure to mass media messages about family planning may be a powerful tool for influencing reproductive behaviour in Nigeria.

A demographic survey among a probability sample of 980 married migrant women was carried out in Sydney in 1988. The sample included 507 Lebanese, 250 Turkish and 223 Vietnamese women. The study revealed differences in family formation patterns within and between the three groups and between them and the general population. Family size had declined among all three groups compared with their family of origin, and it was clear that the younger women would not achieve the same family sizes as the older women. Migrant women tended to marry earlier than the general population and to start their families earlier. While they showed a strong preference for their children to marry within their own ethnic and religious group, nearly one-third said it was up to the choice of the individual. Overall, the future family size of younger migrant women is expected to converge towards the Australian norm. Migrant families are in a state of transition between two cultures which needs to be recognised by health and family planning service providers.

This case-control study was undertaken to determine socio-demographic risk factors for low birth weight in Al-Ain (United Arab Emirates) over a 12-month period in 1992–93. A total of 3485 live births occurred of which 293 (8·4%) were low birth weight. The risk factors considered were mother's occupation, house conditions, place of residence (urban or rural), maternal smoking habits, antenatal care, availability of help in the home, maternal BMI and educational status. Multiple logistic regression analysis showed that mother's occupation, maternal smoking, antenatal care, and lack of help in the home were associated with increased risk of low birth weight.

It has long been known that mild mental retardation (meaning that associated with an IQ in the 50–69 range) shows a strong tendency to run in families and that there is a much increased recurrence risk if either a parent or a sib has mental retardation. For example, Reed & Reed (1965) found that, having had one retarded child, the chance of a further retarded child was 6% if both parents and their sibs were ‘normal’, 13% if both parents were normal but one had a retarded sib, 20% if one parent was retarded, and 42% if both were retarded. A somewhat similar gradient was found in the more recent study by Bundey, Thake & Todd (1989). The appreciation that mild mental retardation was so strongly familial had led Lewis (1933) some years earlier to refer to it as ’subcultural‘ (meaning that it was a normal variation), in order to differentiate it from severe mental retardation, which was viewed as ‘pathological’. Penrose (1938, 1963) made the same classificatory distinction and the utility of a two-group approach became generally accepted (Burack, 1990).

This paper reviews the historical development of the IUD, describing the challenges and successes, and attempts to offer a balanced perspective for family planning service workers today. Modern IUDs are an important component of family planning services and an excellent contraceptive choice for properly screened women, providing contraception that is safe, effective, long lasting and cost effective. Potential research strategies for the future are also discussed.

Some social and economic factors related to breast-feeding durations in Tanzania are analysed using current status data taken from the 1991–92 Tanzanian Demographic and Health Survey. Proportional hazards and proportional odds models are estimated. The results show that breast-feeding durations vary according to the region of residence of the mother and child (and whether they are living in a rural or an urban area), the age of the mother at the time of the birth, the order of the birth, and the mother's religion.

When asked whether he would discuss man in the Origins of the Species, Darwin replied, ‘I think I shall avoid the subject, as so surrounded with prejudices, though I fully admit it is the highest and most interesting problem for the naturalist’. Galton on the other hand replied to the same question, ‘I shall treat man and see what the theory of heredity of variations and the principles of natural selection mean when applied to man’ (Pearson, 1914–30, Vol. II, p. 86).

This study is based on a cross-sectional sample of 459 Sikh adults, male and female, from three castes representing three levels of socioeconomic affluence. The sedentes group stayed all their lives in the native state of Punjab, India; the migrant group completed their active physical growth in the Punjab and left for the UK at around 20 years of age. The pattern of stature, weight, BMI and skinfolds in both groups reflects their level of socioeconomic affluence in their native Punjab. In the migrants, the caste differences in the mass measures largely disappear, but the differences in the skeletal dimensions remain. Both male and female migrants are heavier than their sedente counterparts and the male migrants have substantially higher amounts of subcutaneous fat. The change in weight, BMI and skinfolds in the migrants is inversely proportional to the original values of these variables in the native settings. Male sedentes and migrants are not statistically significantly different in stature, but the female migrants are taller than their sedente peers in all three castes. In a multivariate analysis, caste remains the most significant factor in the skeletal variables; in the mass characters, both caste and migration are statistically significant factors; age explains very little of the variation.

Of all the controversial issues surrounding IQ tests, few have generated more heat and less light than the question whether different groups in our society differ in average IQ. Those who study, and find evidence of, such differences believe that they are addressing a legitimate scientific issue and one whose implications, however unpalatable they may be, society ignores at its peril. Their critics insist that the question has no intrinsic scientific interest whatsoever, and conclude that those who study it are either naive, deceiving themselves or only too happy to find justification for their own comfortable position within an unjust and unequal society. In the case of sex differences in IQ scores, at least, one could argue that both sides have got it largely wrong. With one exception, there is relatively little social or political implication in such differences as there are. More importantly, however, arguments will be presented against the critics' view that such research is of no intrinsic scientific interest. On the contrary, it has already helped to answer some important scientific questions about the nature of IQ, and has the potential to answer more. Unfortunately, those who engage in such research have usually failed to see this.

Son preference is strong in Vietnam, according to attitudinal surveys and studies of contraceptive prevalence and birth hazards. These techniques assume a single model is valid for all families, but it is more plausible that son preference is found for some, but not all, families. Heterogeneous preferences may be addressed with a mixture model. This paper specifies and estimates a two-Weibull regression model, applied to the interval between the second and third births. The data come from the Vietnam Living Standards Survey of 1992–93. Applying information criteria, graphs, and martingale-based residuals, the two-Weibull model is found to fit better than a one-Weibull model. Roughly half of parents have son preference and, curiously, a propensity for fewer children. The other group has more children, no son preference, and is colourless in the sense that the birth interval is difficult to predict on the basis of the regressors used.

Studies of parent–child correlations in stature require data which can be viewed as random samples of some general population and which are large enough to allow partition of the variable and evaluation of non-genetic and genetic influences. In a sample of 4336 individuals drawn from a cohort of all persons born in England, Scotland and Wales in 1 week in 1958, the correlation of statures of the males with their fathers, the females with their fathers, the males with their mothers and the females with their mothers were 0·36, 0·43 and 0·41 and 0·47 respectively at age 16 of the offspring and 0·41, 0·41, 0·47 and 0·46 respectively at age 23. Allowance for the occupational social class of the fathers lowers the correlations, but in no case by more than 5%. Allowance for the occupational class achieved by the offspring by age 23 has little effect on the correlations.