Factors associated with not exclusively breastfeeding at discharge from the obstetric hospital and with duration of breastfeeding were examined in 700 randomly sampled infants. Obstetric records were examined in 97.7% of the subjects and 73.0% of subject families were interviewed. There were 444 (66.5%) infants exclusively breastfed at discharge from the obstetric hospital. Factors associated with not exclusively breastfeeding at discharge from the obstetric hospital after adjusting for potential confounders were: mother leaving school aged less than 18 years, mother not attending antenatal classes and the use of a dummy in the 2 weeks before the interview. Bed sharing practice in the 2 weeks before the interview was associated with exclusively breastfeeding at discharge from the obstetric hospital. Factors associated with a shorter duration of breastfeeding were: mother leaving school before 18 years of age, smoking 20 or more cigarettes per day and use of a dummy. Dummy use may causally reduce breastfeeding or might be a marker for breastfeeding difficulties. There was a dose-response relationship with smoking, with heavier smokers breastfeeding for the shortest time periods. Bed sharing was associated with a longer duration of breastfeeding. This may not necessarily be a causal relationship because breastfeeding may promote bed sharing. The effect of dummy use and bed sharing on breastfeeding warrants further study.