Low serum 25-OHD in female Arab subjects, which may predispose their infants to hypocalcaemia, has been suggested to be due to inadequate sunshine exposure, but may include other sociobiological factors. The effects of duration of sunshine exposure - weighted against the magnitude of clothing (UV exposure) and other sociobiological variables such as age, education and living accommodation - on serum 25-OHD and mineral status of 33 UAE national women of childbearing age were compared with those of 25 non-Gulf Arabs and seventeen Europeans. Serum concentrations of calcium, phosphorus, alkaline phosphatase and intact parathyroid hormone among the groups were not significantly different. The serum concentration of 25-OHD in UAE nationals was 8·6 ng/ml (4·5–17·4), mean±1 SD, and in non-Gulf Arabs 12·6 ng/ml (6·0-26·4); both these values were significantly lower (p=<50·0001) than the 64·3 ng/ml (49-84·3) found in Europeans. Compared with Europeans, the UAE and non-Gulf Arabs in this study were younger, had fewer years of education and had significantly lower clothing and UV scores (p<0·0001). Furthermore, there was a positive correlation (r=0·59425) between serum 25-OHD and UV score, but not with length of exposure. After adjusting for other confounding variables, nationality, clothing and UV scores remained major determinants of serum 25-OHD (p<0·0001). Therefore, limited skin exposure to sunlight appears to be an important determinant of vitamin D status in our subjects. Strategies to increase vitamin D stores should include vitamin D supplementation or advice on effective sunlight exposure.

Information on reproductive attitudes is now routinely collected in fertility surveys in developing countries, and has become very important for understanding fertility behaviour. The quality of this information, however, is rarely assessed, partly due to lack of necessary data. In this paper, the recently completed panel survey in Morocco by the Demographic and Health Surveys (DHS) was used to investigate the consistency of reporting of ideal number of children, reproductive intentions and the planning status of the last birth. The validity of reproductive intentions for subsequent fertility behaviour was also examined. The findings indicate that the three measures of reproductive attitudes are subject to different degrees of measurement error. The measure of reproductive intentions is the most consistent of the three, followed by ideal number of children. The latter is much more consistent at the aggregate level than at the individual level. The reliability of the planning status of the last birth is marred, mainly by an unwanted child being reclassified as 'wanted' with time since the birth of the child. The reproductive intentions of women were also found to be positively related to their subsequent fertility.

The existence of very low rates of fertility among non-contracepting human populations has intrigued researchers in demography and reproductive ecology. Long inter-birth intervals, driven primarily by the lactational amenorrhoea associated with breast-feeding, have been shown to be important determinants of low natural fertility in several populations. Other reports have suggested that sterility brought about by sexually transmitted diseases (STDs) might also explain low fertility in some populations, especially in Africa. This report presents an analysis of 1981 Indian census data that documents low natural fertility in Ladakh, a high-altitude region of the Himalaya in north India. In Ladakh, there is evidence for low rates of marriage among women that may be attributable to the practice of polyandry, but also high rates of primary and secondary sterility within marriage, resulting in low completed parities for post-reproductive age married women. Age-specific fertility rates derived from the number of current births are also unexpectedly low. Hypotheses to explain very low fertility in Ladakh are considered from among the likely proximate determinants and evaluated using two supplementary sources of information derived from fieldwork in Ladakh in the early 1990s. The most likely explanations for low marital fertility include sterility from STDs, high rates of fetal loss, and possibly nutritional contraints on ovarian hormone status.

Breast-feeding and weaning practices were observed for a cohort of 81 children under 3 years of age in the context of a longitudinal study of social, ecological and nutritional factors affecting growth in a semi-nomadic population of pastoralists in northern Tanzania (WaDatoga of Mbulu District). The adequacy of indigenous infant and young child feeding practices was assessed in relation to current international recommendations. Objectives were to provide baseline data for future investigations of any changes in young child feeding practices which accompany population shifts towards settlement and non-pastoral modes of subsistence, and to improve understanding of the strengths and limitations of indigenous feeding practices in this type of population. It was found that while breast-feeding was universally initiated, other aspects of young child feeding practices do not meet current international recommendations. Prelacteal feeds are commonly used, supplementary feeding with non-human milks usually occurs before 4 months of age, use of solid foods normally begins later than 6 months, and breast-feeding does not continue until 2 years of age for the majority of children. The data have implications for the design of breast-feeding promotion and improved weaning food interventions among African pastoralists.

The purpose of this study was to assess factors associated with the use of family planning at first sexual intercourse among young adults aged 15 to 24 in urban Ecuador. The study population consisted of 1443 young adults (494 females and 949 males) in the cities of Quito and Guayaquil, interviewed by the 1988 Ecuador Young Adult Reproductive Health Survey, who reported having experienced consensual sexual intercourse. Approximately 11% of females and 15% of males reported using contraception at first intercourse. Binary logistic regression was performed to assess jointly the effect of multiple factors on contraceptive use at first intercourse. The regression model was first run on the entire study population and then separately for males and females. In the overall population, the following variables were significantly related to using family planning at first sex: being male; being from Guayaquil; older age; father's completion of secondary school. Having lost one's virginity to a prostitute was significantly associated with non-use of family planning. Males were 3·6 times more likely than females to use family planning during their first sexual intercourse. For each year older a young adult was at first sex, his or her odds of using family planning was multiplied by a factor of 1·3. Twenty-eight per cent of males in this study experienced their first sexual intercourse with a prostitute, and these young men were highly unlikely to use family planning. A male who experienced first intercourse with his girlfriend was more than five times as likely to use contraception than a male who lost his virginity to a prostitute.

Homosexual men have a higher mean birth order than do heterosexual men, primarily because they have a greater number of older brothers. The purpose of this study was to determine whether the same difference occurs in homosexual vs heterosexual women. The probands were 964 homosexual and heterosexual, male and female volunteers, from whom birth order data were collected with self-administered questionnaires. The homosexual men had more older brothers than the heterosexual men, but they did not have more older sisters, younger brothers, or younger sisters. The homosexual women did not differ from the heterosexual women with regard to any class of sibling. These results are consistent with the hypothesis that the high birth order of homosexual men reflects the progressive immunization of certain mothers to H-Y antigen by succeeding male fetuses, and the increasing effects of H-Y antibodies on sexual differentiation of the brain in succeeding male fetuses.

Data were obtained on surnames of the parents and places of birth of the parents and grandparents of children in Siena, Italy. Isonymy and total inbreeding coefficient, and their random and non-random components, are 0·005, 0·00125, 0·00019 and 0·00106, respectively. Isonymy and inbreeding figures are similar to those of other medium-sized Italian towns, while higher values have been reported for Italian villages and Italian ethnic minorities. City endogamy, and endogamy of Contrada for grandparents have the same values (44·1 and 44·8%, respectively), but for parents, endogamy of Contrada is lower than city endogamy (15·2 and 33·4%, respectively). The difference between the extent of Contrada endogamy expected at random and observed in the parents' generation does not seem to affect the genetic structure of the present population. However, the bulk of marriage migration (more than 70%) is short range, with people coming from Tuscany. There is no statistical difference in marital migration between males and females.

The author's theory that among adults men have a higher mean IQ than women has been critized by Mackintosh. His criticisms are examined and found wanting. New evidence is adduced that men obtain higher means than women on Raven's Progressive Matrices.

Do males and females differ in general intelligence? Lynn (1994, 1998) and I (Mackintosh, 1996) agree that the answer to this question must depend on the answer to two prior questions: what is the definition of general intelligence, and what tests best measure it? Our disagreement arises partly from different answers to these prior questions, but also from differences in our reading of some not wholly consistent evidence. What answers might one give to these prior questions? They are not independent, for a possible (severely operational) definition of general intelligence is that it is a person's score on one particular IQ test battery. Let us start with that simple possibility.

From 1971 to 1984, the Israeli Defence Forces administered two tests unaltered that qualify as IQ tests. These were the Matrices and the Instructions tests. The former is an adaptation of Raven's Progressive Matrices, similar to the original in that there are clusters with progressive difficulty and an escalation of difficulty throughout the test. The raw score range is from 4 to 35. The Instructions test is adapted from the old US Army Alpha Instructions Test and is described as an Otis-type verbal test. It includes 21 open-ended instructions. Typical items: write the last letter of the word which is the opposite of black; write the third letter of the largest word among the following four words; find the word beginning and ending with the same letter and write that letter. There are similar questions using numbers rather than letters. Both tests are timed, and most subjects do not complete either. There is no penalty for wrong answers (Col. Moshe Even-Chen, personal communication, 14 March 1990; Gal, 1986, pp. 79 and 95; Judy Goldenberg, personal communications, 4 March 1991 and 25 April 1993).

Students of the problem of sex differences in intelligence will be indebted to Mackintosh and Flynn for providing useful further clarification of the issue of sex differences in intelligence. I have little to say about Flynn's paper except to note that it confirms my thesis that among adults males have a small advantage over females in abstract reasoning ability. Mackintosh and I have some common ground but there remain substantial disagreements.