Opportunity for natural selection in human populations has so far mainly been studied on anthropological data for tribal populations or on census data for nations. The present study is mainly based on data on individual lifehistories but also, for part of the longitudinal study, on census data. Six of the populations, Nedertorneå, Tuna, Svinnegarn, Trosa, Locknevi and Fleninge are parishes. These sets of data covers the period 1800-1850 as defined by the birthyears of the women. The data for the longitudinal study are derived from two sour­ces, a biography over all clergymen in the diocese of Linköping, cove­ring the period 1600-1845, and material published by the National Swe­dish Central Bureau of Statistics (SCB) that covers the period 1750-1980. For each subpopulation data on childhood mortality and female fertility has been collected and from these data Crow's index of opportunity for natural selection has been calculated. The original index has also been modified in order to estimate the importance of childlessness in relation to the total index. The study shows that for the periods and the populations studied, there is a considerable opportunity for natural selection both through mortality and through differential fertility and that, during our cen­tury, differential fertility has become the main asset for natural se­lection, as mortality has been reduced to very low levels. It is also obvious that childlessness is an important factor as regards natural selection in human populations. The cross-sectional study shows signi­ficant differences between the populations for all components of the index. The longitudinal study covers when, the two sets of data are combined, a period of over 350 years, 1600-1980. Over this period changes in index of opportunity for natural selection have occured but these changes are not very drastic as compared to other longitudinal studies. However, within a separate region there can be drastic chang­es in index between decades and there are large differences between regions. Mortality and fertility patterns have been studied from different angles. With the exception for the census data, each woman in the stu­dy has be followed from 16 to 40 years of age and each of her children (if any) has be followed from birth to 16 years of age or death, if prior. Therefore it was possible to obtain distributions for age at first childbirth, sibship size, succesful sibship size, childhood mor­tality by age at death, female mortality, and childlessness, total and marital. In some cases a study of sex ratio at birth and at 16 years of age, and birth intervals, have been made. Statistical analysis of the results shows significant differences between populations for all tests that have been applied. The Linköping data was analysed for dif­ferences between periods. Significant differences were found for all of the parameters with the exception of female mortality. / <p>Härtill 5 uppsatser</p>