A proportional hazards regression model was applied to data on women aged 45–49 from the Egyptian Fertility Survey, to assess the effects of women's education, residence, and marital experience on their age at the birth of their last child. When age at marriage and parity were controlled, well-educated urban women tended to stop reproducing earlier than less educated rural women. Compared to intact first marriages, marital dissolution (divorce, widowhood, or separation) with remarriage tended to prolong the age at last birth, while failure to remarry tended to hasten it.

Data from birth records with information on previously born children from three maternity hospitals in Norway have been used to study the trend in breast-feeding practice from 1860 to 1984. During the whole period, the percentage of women breast-feeding for at least 1 week was remarkably high, above 90%. The results show a fairly stable duration of breast-feeding up until 1920, at least 3 months in approximately 80% of the women. After that year, the distribution of the duration of breast-feeding changed dramatically. The shortest durations were found in the late 1960s, when only about 30% lactated for 3 months or more. The duration then increased quickly, so that at the beginning of the 1980s about 80% of the women were again breast-feeding for at least 3 months.

By multiple regression methods, the relationships of several independent variables to the duration of breast-feeding were investigated. Married women lactated for 1·5–3 months longer than unmarried, the difference being largest before 1920. First-born children were breast-fed for a little shorter time than second and later-born children. Until World War II women of lowest social strata had the longest durations of breast-feeding, and then the situation was reversed, women of highest social strata continuing the longest.

Delivery records from Tanzanian hospitals reveal a marked seasonality of births in areas with holoendemic malaria. Accepted explanations of variations in conception rate are inadequate to account for these seasonal variations. The magnitude of the variation increases with higher parity but it has decreased over the past decade. The differences are related to different activity patterns of younger and older women, and to recent changes in the rural economy. Conception rate has a negative association with rainfall 4 months earlier; birth seasonality is therefore considered in relation to the agricultural cycle. While seasonal variations in sexual activity and pregnancy loss may be contributory factors, female stress due to the combination of malarial infection and physical exhaustion emerges as the major cause of seasonally depressed fecundity in areas with holoendemic malaria.

In a multi-centre survey in 1986, 400 married Filipino women aged 15–40 were interviewed about their use of contraceptive methods, and specifically about their perceptions of the effects on health of oral contraceptives and their attitudes to contraceptive methods. The sample was randomly selected in the urban and rural areas surveyed and cannot be considered representative for the country as a whole. The results showed that women hold definitive views on the health risks posed by oral contraceptives and point to the importance of family planning clinics as a source of contraceptive counselling. The overall rate of contraceptive use must be regarded as insufficient to meet contraceptive needs.

In summer 1987, 584 students at Oxford University were sent a questionnaire enquiring about their sexual behaviour and contraceptive practice, and about the effect that AIDS has had on these. Of the 374 students who replied, 65% of the women and 62% of the men had experienced sexual intercourse. Twenty-seven per cent of female undergraduates did not use adequate contraception on the first occasion of intercourse. During the 4-week period preceding the questionnaire, 15% of sexually active female undergraduates did not use adequate contraception. The use of contraception has changed over the last 5 years, with fewer students using oral contraceptives as their most frequent means of contraception and more using barrier methods. In general, knowledge about AIDS was high. Most students did not consider that they were at risk of becoming infected by the AIDS virus and, by their reported sexual behaviour, very few students appeared to be at any risk. However, 35% of women and 44% of men who were sexually active said that because of risk of AIDS they had been more likely to use the sheath, and 49% of women and 30% of men said that they had, or would have, fewer partners. These findings suggest that attitudes, and possibly behaviour, are changing towards a reduction in risks of sexually transmitted infection.

Results from a 1984 national survey indicate that the large majority of Thai women initiate contraception soon after giving birth. Among sterilized women, the vast majority undergo the operation during the immediate post-partum period and thus avoid the risks of subsequent unwanted pregnancy. Rates of initiation of other methods are unusually high at and immediately following the time at which menses return, and hence only modest proportions of Thai women remain at risk of unwanted pregnancy for very long after the birth of a child.

Socioeconomic differences and trends in infant and child mortality in Bangladesh are examined using data from the 1975 World Fertility Survey and 1979 Contraceptive Prevalence Survey. There is evidence of some recent decline in infant and child mortality. Logit analysis of infant and child mortality indicates that sociodemographic variables such as mother's education, recent period, or higher birth orders, had significant independent effects upon the reduction of infant and child mortality. Other variables such as fetal loss, father's education, or land ownership had no consistent significant effect of upon infant and child mortality. On the other hand, the effect of urban residence on infant and child mortality was positive after the control of the sociodemographic variables. Mere concentration on the supply of modern medical services may bring limited returns unless they are reinforced by appropriate social changes, in particular those affecting the socioeconomic status of women.

Longitudinal data were used to examine relationships between sibling variables and adolescents' aspirations at different social status, family environment and ability levels in a sample of 260 Anglo-Australian, 120 Greek, and 90 Southern Italian families. In this follow-up study of an earlier investigation each family had a 16-year-old child and the analyses related to those adolescents. The study suggests that there are different relationships between birth order and educational and occupational aspirations for adolescents from different ethnic groups, even after taking into account the effect on aspirations of measures of ability and family environment.

This paper presents time series data on spouse selection in Canada for the period 1921–78. Homogamous and endogamous trends are examined by age, sex and religion. There is considerable variation in the level of endogamy amongst religious groups. With respect to age, marriages were increasingly homogamous up to 1971 and thereafter less so. Younger males increasingly tended to select brides of their own age group until the 1960s, while older males have been increasingly heterogamous since World War II. Females under 20 years of age tend to select older grooms; other brides show no clear trend in their spouse selection by age.

Data from pregnancy histories collected by the 1976–77 Mexican Fertility Survey show wide variations in infant mortality among a sample of 125 communities. There are also large infant mortality differentials by population size, economic status, access variables, utilities, medical facilities and schools. These community-level factors, which have been little used in previous studies, are highly correlated and do not appear to affect infant survival independently of population size. As such, community size serves as a summary measure of a locality's overall level of development and comparative risk of early death for its young inhabitants.

Inbreeding coefficients were calculated from data on spouse relationships in a sample of 4833 marriages comprising 345 from three Nilotic tribes, 302 from two indigenous Negroid tribes, and 4186 from several Arab and other tribes. The average inbreeding coefficients were 0·0081 to 0·0145 in Nilotic tribes, 0·0251 to 0·0269 in indigenous Negroids, and 0·0286 to 0·0430 in Arab tribes. First cousin unions constituted about half of all the consanguineous marriages and the pattern of first cousin marriages was similar in all three groups. The inbreeding coefficient for the entire sample was 0·0317.

This paper presents information on the incidence of hysterectomy and tubal ligation in South Australia in the period 1980–82, and on the age, marital status and ethnicity of women undergoing these procedures in public hospitals during the period. A typical woman undergoing one of these procedures was married or previously married and in her mid-40s for hysterectomy or mid-30s for tubal ligation. South Australian women had a lifetime chance of one in six of undergoing hysterectomy and of one in five of undergoing tubal ligation. The findings are consonant with the suggestion that increasing numbers of women are choosing forms of sterilization as means of contraception.

The family registration records from a village population in a small island of Japan are used to assess the effect of demographic differentiation within a population on genetic measures. When the couples studied are classified by birth cohorts and origins, wives of the couples where one spouse came from elsewhere were older at marriage and had a shorter duration of marriage or registration than wives where both spouses were natives of the village. The mean number of offspring is statistically smaller in the former except for the latest cohort, due mainly to out-migration during the reproductive ages which also resulted in low rates of marriage among the offspring within the village. This leads to a small effective population size and an increased likelihood of genetic drift in the overall population.

The effect of migration on the maximum opportunity for selection is analysed using koseki data, the Japanese family registration records, of a village population in a small island. When selection intensity is divided into two components, migration greatly influences the variance of the mortality component of the original formula. The difference in selection intensity, with and without allowance for migration, is conspicuous in the recent cohort; the intensity, with a migration component, shows a consistent increase, whereas it declined in the original formula. The implications of this are discussed in the context of the genetic structure of a population of small size, changing rapidly over time.

Infant mortality in Sri Lanka between 1961 and 1980 is studied with reference to its trend and associated factors. Between the periods 1961–65 and 1976–80 deaths from exogenous and endogenous causes have declined considerably and nearly equally. The probability of survival has increased most in the first week of life.

National income or total food supply does not seem to be associated with infant mortality. Distribution of free supplementary food, increase of public health personnel, of immunization, and of institutional births appear to have initiated and sustained the transition of infant mortality during the period. Increasing levels of female education probably augmented these effects.

The duration of amenorrhoea among a group of chronically malnourished women in a rural area of Bangladesh is examined by application of multivariate hazard models with time-varying covariates, including the influence of maternal nutrition, seasonality and patterns of infant feeding. Both maternal weight at pregnancy termination and the pattern of infant feeding affected the length of post-partum amenorrhoea. Analyses focused on season of birth showed the importance of differences by education in infant feeding.

A survey of consanguineous marriages in Japan in couples and their parents was conducted in 1983 through questionnaires. The total number of couples studied was 9225; they were chosen from six widely different areas of the country. The kinship coefficient between parents did not decrease with the year of birth. Thus, the recent decline in the frequency of consanguineous marriages in Japan seems to be largely a post-World War II phenomenon. The kinship decreases with marital distance in the parental generation, and socioeconomic class (level of education and occupation) effects are small by comparison with those of the present generation. The rate of consanguinity is significantly higher in the older generation for almost all areas. The kinship between spouses is less than a half of that in parents for the whole of Japan. Geographical variation in the rate of consanguinity is more remarkable in spouses than parental couples.

Multivariate analysis of the effects of maternal age at birth, birth order and the preceding birth interval on mortality risks in early childhood, using data from the Bangladesh Fertility Survey, 1975–76, confirms that the length of the preceding birth interval is the most influential single factor. But the lower mortality risks among infants and children of educated mothers are due neither to the age at which childbearing was initiated nor to the spacing between births.

Using data from a 1982–83 survey, infant and early childhood mortality are investigated in a large rural region of Senegal. Because of evidence of under-reporting of early mortality an adjustment factor was applied to the data. Infant mortality was estimated to be about 113 deaths per 1000 live births. Childhood mortality was very high relative to infant mortality. Some noteworthy differentials in mortality between geographic, ethnic, and socioeconomic groups were recorded. It was estimated that diarrhoeal and respiratory diseases accounted for about one-half of reported deaths. Tetanus was an important cause of death during infancy. Measles and malaria were important causes only after the first birthday.

From analysis of the 1974 Korean National Fertility Survey data, the changing patterns of demographic and socioeconomic determinants of infant and child mortality are generalized. The pattern for infant mortality is: (1) in a traditional society demographic factors affect infant mortality more than socioeconomic factors; (2) at the early stage of development, demographic factors are replaced by socioeconomic factors as the main determinants; (3) when the difference in living standards between social classes narrows, the socioeconomic differentials in mortality also diminish; and (4) at the stage of high development the effects of demographic factors remain although the absolute differences are very small. But the pattern of changing determinants of child mortality is nearly the reverse of that of infant mortality.