Total population surveys were carried out in 1980—84 (LEP-1) and 1986—89 (LEP-2) in Karonga District, northern Malawi, as part of a major cohort study of leprosy and tuberculosis. Retrospective information on fertility was collected from adult women in selected villages. Mortality estimates for the adult population were based on reported inter-survey deaths. This paper describes the demographic dynamics of the population in this time period, which coincided with the start of the HIV epidemic. A total of 112 886 individuals were interviewed in LEP-1, of whom 88 544 were also examined in LEP-2. Of this population, 46 per cent were under 15 years of age and about 20 per cent could not give a precise year of birth. Procedures for smoothing the age distribution and correcting the under-ascertainment of infants are explained. Total fertility in the interval between the two surveys was 6.4 children per woman and the crude birth rate was 48 per thousand. The crude death rate for the same interval was about 18 per thousand, and the growth rate was 4 per cent per year. Infant mortality was around 120 per thousand live births, under-5 mortality was around 190, the index of adult mortality (proportion of those who survived to age 15 dying before age 60) was 37 per cent, and life expectancy was around 50 years. Analyses adjusted for age, sex and socio-economic factors show higher mortality in north compared to south Karonga (rate ratio of 1.29, 95 per cent CI: 1.19, 1.38); and in those with estimated rather than precise years of birth (rate ratio of 1.14, 95 per cent CI: 1.03, 1.25). The similarities of the estimates of the age—sex distributions, and summary measures of mortality and fertility to the 1987 Malawi census and 1992 Malawi Demographic and Health Survey results mutually support the reliability of Karonga epidemiological survey data, and add details not available through the standard demographic survey and census methods, particularly with respect to adult mortality.