The ‘insurance ova’ hypothesis (Anderson, 1990) views dizygotic twinning as a by-product of selection for multiple ovulation which sometimes – in error – results in the birth of twins. From this viewpoint, polyovulation is a mechanism which reduces the risks of fertilization failure or embryo defect/mortality. If DZ twin births are a ‘side-effect’ of a mechanism which compensates for defective embryos one might predict that embryo defect rates and twinning rates will covary. This prediction is tested using national-level data on twinning rates and rates of trisomy-21 (Down's syndrome), and a strong positive correlation is found, even when controlling for maternal age. One suggestion that follows from this finding is that intra- and interpopulation variation in both twinning rates and Down's syndrome rates may result, in part, from individual variation in pre-implantation rejection of embryos in the very early stages of pregnancy. In this paper it is proposed that the ‘insurance ova’ explanation for twinning in humans could be expanded to incorporate a model of rejection of anomalous embryos, be they anomalies of number or type. Variation in the efficiency of an embryo rejection mechanism, combined with variation in frequency of polyovulation, would have consequences for individual reproductive success.

In Tokyo Megalopolis, long-distance commuting between residences in the suburbs and work places in the centre of the city was examined. Using a questionnaire, heads of household in two suburbs were asked about the influences of long commuting hours on their daily time budgets. The results showed that on workdays, the workers who spent longer commuting rose and left home for work earlier, and came back home and retired later; accordingly, both time spent on in-home activities on workdays and hours slept on the day before a workday were shorter. Comparison of time budgets between the subjects who work 5 and 6 days per week revealed more vulnerable influences of long commuting hours in the former than the latter. The expected health consequences of these findings are discussed from a biosocial/human ecological viewpoint.

The strength of influence upon statural variation of: (1) the degree of urbanization of the locality of habitat, (2) family size, (3) paternal and (4) maternal educational status was analysed in three generations of 19-year-old Polish conscripts, examined in 1965, 1986 and 1995. Each of the above factors of an individual's social situation was described by a 4-level scale. Each factor was found to exert a highly significant residual effect on stature throughout the three decades considered, even after the effects of other correlated factors were partialled out by three-factor ANOVA. However, the stratifying force of each factor, as expressed by the dispersal of the level-specific main effects around the national mean, has been changing over time. For example, the growth-stunting effect of the condition of coming from a large sibship was dramatic in the 1965 cohort and considerably attenuated in 1986 but ceased to diminish thereafter. The growth-enhancing effect of the condition of being a large-city dweller, initially marked, has almost disappeared; but the growth-stunting effect of the condition of being a rural dweller has remained equally strong across all cohorts. These and other shifts in the relative importance of the social factors, as presumed determinants of family living standards, are described and some explanations attempted.

In 1993, the Navrongo Health Research Centre launched a new demographic research system for monitoring the impact of health service interventions in a rural district of northern Ghana. The Navrongo Demographic Surveillance System uses automated software generation procedures that greatly simplify the preparation of complex database management systems. This paper reviews the Navrongo model for data collection, as well as features of the Navrongo system that have led to its replication in other health research projects requiring individual-level longitudinal demographic data. Demographic research results for the first 2 years of system operation are indicative of a pretransitional rural society with high fertility, exceedingly high mortality risks, and pronounced seasonal out-migration.

In-depth interviews were conducted with married Asian women from Indian, Pakistani and Bangladeshi backgrounds, to investigate patterns of contraceptive use and influences on contraceptive decision making. The results show two distinctively different contraceptive ‘lifecycles’. Non-professional women typically have little knowledge about contraception until after their marriage or first birth. Their patterns of contraceptive behaviour show low levels of contraceptive use until after their first birth, when condom use is most prevalent. Non-professional women are influenced by their extended family, religion and cultural expectations on their fertility and family planning decisions. Professional women show an entirely different pattern of contraceptive behaviour. They are more likely to have knowledge about contraception before marriage, use some method of contraception throughout their childbearing years (typically the pill) and cite personal, practical or economic considerations in their fertility decisions rather than religious, cultural or extended family influences.

There is increasing evidence that age at menarche has decreased in Europe and the United States during the last century and in Japan over the last several decades. Data from a community-based survey conducted in two rural counties of Anhui Province in China indicate a similar, downward secular trend in age at menarche for Chinese women. The present study shows the mean age at menarche decreased by 2·8 years, from 16·5 to 13·7, over an approximate 40-year time interval. This rapid decrease in age at menarche may partly be due to better nutrition and living standards reflected by the improved socioeconomic standards experienced in China over the past few decades. To test this hypothesis, a number of determinants of age at menarche were assessed; year of birth, literacy status, county of residence, amount of physical labour, general health status, pesticide exposure before age at menarche, and drinking water source were all found to be associated with age at menarche.

The relationship between mood states, urinary stress hormone output (adrenaline, noradrenaline and cortisol) and adequacy of the menstrual cycle was examined in 120 recorded non-conception cycles from 34 women. It was hypothesized that women with higher stress levels would be more likely to experience abnormal cycles and that within women higher stress levels would positively relate to follicular phase length and inversely relate to luteal phase length. There was a non-significant trend for women to report higher stress levels during oligomenorrhoeic and unclear cycles compared with normal cycles. Analysis of covariance indicated that there was no consistent relationship between the measures of stress used here and follicular or luteal phase length within women. There was also no consistent pattern of relationship between reported mood states and stress hormone excretion within women. Further research is warranted to understand the role of stress and subtle menstrual cycle abnormalities in female fertility.

Qualitative research and cross-sectional survey methods were used in a study conducted in rural and urban areas of the Greater Accra Region, Ghana, to explore people's understanding of the cause of malaria and patterns of mosquito avoidance, in particular bed net ownership and use. The study indicated far higher bed net ownership and use in rural than urban areas, which was related partly to perceived affordability and partly to the different contexts of and reasons for avoiding mosquitoes. Knowledge of an association between mosquitoes and malaria, the most common cause of illness in both areas, was related to residence but not to literacy or formal education, and this knowledge did not predict bed net use. The paper points to the complexity of social and personal factors implicated in behavioural interventions for malaria control, and questions behavioural models that assume a linear relationship between knowledge and practice.

Epidemiological studies have revealed that stressful changes in social environment increase the risk of cardiovascular mortality. In this study, the influence of major negative and positive life changes on serum cholesterol was examined in middle-aged men to determine a possible biochemical link between life changes and cardiovascular mortality. The results showed no influence of negative life changes on serum cholesterols. However, positive life changes significantly predicted a reduction in total and LDL (low-density lipoprotein) cholesterol levels after adjustment for the baseline cardiovascular health status, baseline cholesterol level, diet, body mass index, waist-to-hip ratio and cardiorespiratory fitness. The odds ratio for lowering LDL cholesterol was 5·2 in the men reporting positive events compared with those reporting none. The findings suggest a predictive value of positive life changes for atherogenic lipid profile in middle-aged men.

Analysis of the 1992 Niger Demographic and Health Survey showed that although roughly two-thirds of both polygamous and monogamous women approve of birth control, polygamous wives are less likely than monogamous wives to discuss family size or birth control with their husband or to plan on using birth control. The study suggests that characteristics of polygamous couples have caused polygamous women to be more resistant to birth control use than monogamous women. The polygamous women tended to be married to older men who had not gone to primary school and who desired more children than monogamous husbands. The influence of marital structure is not significantly associated with intention to use birth control when the husband's age and the wife's ideal number of children were controlled for in the multivariate logistic regression model suggesting that background social factors may be more influential. In fact, educational level and age at first marriage were significantly associated with attitudes towards birth control and also with marital structure.

The purpose of this study was to validate the self-reported body height and weight of adult Japanese women. The subjects were women, aged 20–42 years, who participated in a survey on eating disorders in women in 1995. Physically measured height and weight data were obtained for 368 (89·8%) of the 469 women who self-reported their height and weight. The report-based heights and weights were compared with the measured values. The correlation coefficients for height and weight were 0·990 and 0·963 (p<0·0001), respectively. Mean reported height was 0·1 cm shorter and mean reported weight 0·2 kg lighter than the measured values. Shorter women tended to report a taller height than their actual height, and heavier women to report a lower weight than their actual weight. Despite these limitations, the self-reported heights and weights of adult Japanese women were precise and accurate, and their use in epidemiological surveys is considered acceptable.

Using data from a cross-sectional nutrition survey conducted in rural Ethiopia between March and April 1992, roughly a year after the end of one of the longest civil wars in modern human history, this study attempts to document the magnitude and correlates of childhood undernutrition in Ethiopia. Findings from the study reveal that, at the time of the survey, 59% of children in the country were exposed to long-term or chronic undernutrition (stunted); about 4% were suffering from acute problems (wasted); and about the same proportion were both stunted and wasted. Less than a third of the country's children had normal growth. The logistic–binomial regression results demonstrated the existence of significant clustering of risks of undernutrition within areas of residence and notable differentials by age of child, duration of breast-feeding, age at introduction of supplementary foods and number of under-five siblings. Significant variations were also noted by age of household head, per capita land holding, religious affiliation and prevalence of endemic diseases in an area.

Using a sub-sample of ever-married women from the 1993 Ghana Demographic and Health Survey (GDHS), this study examines differentials in contraceptive use in six cultural groups: Ga-Adangbe, Twi, Fante/other Akans, Ewe, Guan/others and Mole-Dagbani. Multivariate analysis is used to explore whether reported ethnic differentials in contraceptive use can be attributed to ethnicity or to other characteristics that distinguish the ethnic groups. Overall, the findings are generally more consistent with the 'characteristics' hypothesis, because contraceptive use differentials by ethnic group is accounted for by differences in socioeconomic and demographic characteristics of these women. However, for the Fante/other Akans, even after the necessary controls, ethnicity continued to emerge as a significant determinant of contraceptive use. Programmatic implications of these results are discussed.

The surnames of the 3443 males registered to vote in Kidlington in 1977 yield a Coefficient of Relationship by Isonymy of 0·000564 {Ri=Σ(n(n–1))/2 N(N–1), in which n=the number of men of each surname and N=Σn}. Those of the four wards separately average 0·000722. However, if one includes only one male of any one surname in each residence, the values are, respectively, 0·000534 and 0·000535. That is, the only structure seen between the two levels is in the influence of men of the same surname resident in the same house. An analysis of relationship by residence on the same street yields a value of Ri somewhat higher than that for the ward as a whole, however, suggesting that even within a ward there may be a tendency for the house of relatives occasionally to lie close together.

The incidence of obesity, defined as the fraction of persons with BMIs exceeding 30·0, was examined in two birth cohorts of 40–50-year-old occupationally active inhabitants of the city of Wrocław, Poland. In both cohorts and both genders obesity increased monotonically with decreasing position on a three-level educational scale. During the 1986–1996 decade obesity increased dramatically among males with trade school education only; concomitant shifts were much smaller or absent in the college-educated groups. The contrast in obesity between the opposite ends of the educational scale has widened markedly in both genders; and the between-gender gap has declined somewhat.

Fertility rates in Pakistan have remained consistently high over the past three decades. While numerous studies have examined sociodemographic determinants, the role of biological factors, and particularly consanguinity, has received little attention, even though marriage between close biological relatives continues to be the norm in Pakistan. Reproductive behaviour among women in consanguineous (first cousin) and non-consanguineous unions was compared, using data from a 1995 study of multi-ethnic communities in Karachi and the 1990–91 Pakistan Demographic & Health Survey (PDHS). The results show that, although female age at first marriage has been gradually rising in both study samples, women in consanguineous unions married at younger ages and were less likely to use modern contraceptive methods. In the Karachi sample, women in first cousin unions experienced a higher mean number of pregnancies and also reported a higher mean number of children ever born (CEB). However, their mean number of surviving children did not differ from those born to women in non-consanguineous unions, implying higher prenatal and/or postnatal losses in couples related as first cousins. On the other hand, the PDHS showed both lower CEB values for women in consanguineous marriages and a lower number of surviving children. Given the continuing popularity of consanguineous marriage, these findings have important implications for future fertility reduction in Pakistan.