A one-year birth cohort was studied in Jimma town, South West Ethiopia, in 1992-93. We report here on the design and on the methods used in the study and describe the principal health outcomes. Infants were visited bimonthly until their first birthday. Background data on the physical, cultural and economic environment of the home were collected at the first visit, and data on nursing and weaning, on traditional surgical and other practices, and on vaccination at the first visit and at each subsequent visit. Length, weight and mid upper arm circumference were measured, and details of the mother's handling of illness episodes recorded. Of 1563 children born, 86% were successfully followed to the end of their first year or to an earlier death. There were 141 deaths, indicating an infant mortality of 115/1000 (estimated probability of surviving to 1 year 0.8851, with s.e. 0.0101). The mean length and weight of the singleton infants at the end of their first year was -1.41 and -1.52SD below from the median of the NCHS/WHO reference population. Weights throughout the first year were analysed in more detail using a Reed model, fitted as a random coefficient regression model in ML3-E. There were clear differences in growth across the different ethnic groups, with the best growing group weighing on average about 1 kg more at the end of the first year than the groups growing least well.