When invited to deliver this year's Caradog Jones Lecture, I was familiar with neither the lecture as an institution nor the work of the man in whose honour it is held. However, in view of the interest of Caradog Jones in the quality of British society it became clear that the subject of the lecture should be related, amongst other things, to the improvement of social conditions in general and to the improvement of the quality of the population of Great Britain in particular.

It is hardly necessary to say that the present state of British society leaves room for improvement, beset as it is with severe social and economic problems. Now the key to the solution of these problems must be a sound understanding of the way social and economic systems work. What is apparent from the briefest scrutiny of the economic and social reforms attempted in the recent past is that they have not succeeded and this, I strongly suggest, is because the theoretical premises on which they were founded are unsound.

For 926 persons aged 5 years and over, a large majority of the population of three western Newfoundland communities, associations of ocular refraction with inbreeding (Wright's F) have been evaluated. Some evidence was found among children of a tendency towards myopia with increasing homozygosity and a strong such tendency in the elderly (over 60) was found. No relationship of refraction to F was apparent for persons aged 15–60 years. When the strong relationship of refraction value to nearwork activity was taken into account the relationship to F was enhanced.

This investigation was prompted by Shuster's (1978, 1980) suggestion that acne should be regarded as a normal feature of adolescence. Acne was recorded and graded in white school children (722 boys and 788 girls) seen at approximately half-yearly intervals from age 9 to 17. Losses over the 8 years of follow-up were 16·8% for boys and 11·4% for girls. Facial acne appeared at an earlier age in girls, the prevalences being estimated to rise to a peak of 75–80% in girls aged 15–16 and 90–95% in boys aged about 17–18 years. However, the prevalence of acne at a given age depended much more upon the level of sexual maturity attained (assessed from a variety of somatic indicators) than upon age per se. On average, acne became manifest at a relatively late stage of puberty, at or shortly after peak height velocity in boys and menarche in girls. Prevalences were not affected by relative obesity nor by socioeconomic status, apparent correlations among girls disappearing when maturity status was taken into account. The proportions of subjects with moderate or severe acne increased with length of time from onset, and only about 10% showed remission during the first 3 years after onset. Acne on the body was observed in 20% of boys and 15% of girls.

Analysis of the data from the tenth Italian census shows that women born to older fathers tend to marry older men. The correlations between the ages of the father and the husband in various situations were always found to be positive and significant when the age of the woman and her educational level were held constant. Correlation coefficients are higher in rural and economically less developed areas. Behavioural ‘imprinting’ is suggested as a possible explanation.

It is suggested that dizygotic twinning rates and fecundability are closely related and that the recent declines in dizygotic twinning rates may have been accompanied by declines in fecundability in developed societies. Records on twinning rates for many countries are readily accessible in national vital statistics; demographers may therefore find that changes in dizygotic twinning rates may be a useful monitor of changes in fecundability. The question arises whether this decline in fecundability has been responsible, to any appreciable extent, for the decline in birth rates experienced in recent years by developed countries.

The hypothesis that women who experience accidental pregnancy are at risk for all types of accident has been examined to test indirectly the validity of a commonly postulated abortion personality profile. Identified women undergoing abortion in Odense, Denmark, were compared with non-abortion community women as to their registered incidence of accident-related trauma over a 2-year period. Age- and cause-specific accident rates were found not to vary significantly between the abortion and non-abortion groups of women.

Fourteen variables including socioeconomic factors, demographic variables, reproductive ideals and contact with family planning programmes are tested as correlates of contraceptive use among a low-income population of Cali, Colombia. Data from surveys in 1976 and 1978 show two factors to be key determinants: number of children and contact with the local contraceptive distribution post. The availability of family planning services is discussed as an important concern for policy makers.

Data from a survey of 1699 currently (once) married males are used in examining ideal family size in Northern Ireland, an area where fertility remains quite high by west European standards. Although there are several problems concerning the measurement and use of ideal family size, an analysis of the Northern Irish data indicates that the responses appear to be valid. For example, most respondents can formulate an ideal family size in numerical terms, there is a considerable degree of variation in the responses and these are not necessarily a function of actual family size. Northern Irish family size ideals (mean 3·6) are considerably greater than those of many west European countries and these ideals also vary within Northern Ireland. Religious grouping proves to be the major factor influencing these variations, with Roman Catholics generally expressing an ideal (mean 4·4) for large families. However, other factors are also associated with variations in ideal family size and the paper considers the roles played by education, age, occupation, region of residence and religiosity. Also, ideal family size does not always correspond with actual family size, indicating that while some married men possess relatively large family size ideals but have smaller families, others have large families but favour more moderately sized ideals. It is evident that attitudes to fertility in Northern Ireland are in a state of change. The value of ideal family size data in understanding fertility patterns is emphasized.

This paper extends the analysis of sibling groups from Ghana (Goody et al., 1981) by a comparison with seven sets of data from various parts of the world. Sex preference is analysed from the actual holdings of children, taking into account the numbers and sequence of sexes by which the family is built up. The hypothesis examined is that sex preferences are related to general socioeconomic factors and in particular to differences between extensive and intensive systems of primary production. The results suggest that such preferences are shown in the holdings of completed families, though their effects are of small magnitude and they vary from society to society.

Using data collected through follow-up home visits to a sample of mothers, whose singleton babies were born at the Dufferin Hospital, Lucknow, September 1976—August 1977, estimates are made of infant mortality in Lucknow city, and its differentials by socioeconomic, demographic and health care characteristics of the infants and their families. Infant mortality was 74·5 per 1000 births during the year 1976–77 and there is evidence of some recent decline in it.

Among socioeconomic factors considered, infant mortality is found to be inversely related with caste status among Hindus, education of mother, father's monthly income, and father's occupation. Infant mortality exhibits a J-shaped distribution in relation to maternal age and parity and a reversed J-shaped distribution in relation to maternal age at marriage. Among health and health care factors, antenatal care of the mother appears to exercise a negative influence on infant mortality, which also shows a marked negative relationship with birth weight and feeding index of the child. The findings suggest important policy implications, which, if followed, would help to reduce infant mortality, and would thus remove a serious obstacle in the way of acceptance of the small family norm and its achievement through the adoption of family planning.

The results of a demographic study of two occupationally distinguishable caste communities inhabiting the neighbouring settlements of Balaramchak and Bamanchak in the Geonkhali region, coastal Midnapore district, West Bengal, are reported. Although there is no significant difference in mortality between the two settlements, fertility is generally higher in Balaramchak than in Bamanchak. Further, fertility seems to have declined over time in both the settlements, with the decline commencing earlier in Bamanchak. Possible explanations for the inter-settlement and temporal differences in terms of occupation and family planning practices, as well as the possible genetic consequences of these differences, are discussed.

This paper examines urban-rural differences in the fertility of Pakistani women, as shown by the 1975 Pakistan Fertility Survey, part of the World Fertility Survey. The urban fertility rates were slightly higher than the rural rates. Urban marital fertility exceeded rural marital fertility at the younger reproductive ages, mainly on account of shorter duration of breast-feeding, which reduces birth intervals, but fell below it at the older reproductive ages. The beginnings of parity-specific birth control are discernible at the older reproductive ages in urban areas.