In a randomised trial, infants living in a large village in The Gambia were immunised either at 4 months of age with 40,000 plaque forming units (PFU) of the Edmonston-Zagreb (EZ) measles vaccine or at the usual age of 9 months with 6000 TCID50 of a conventional Schwarz measles vaccine. Measles developed in 2 of 119 children who received the EZ vaccine, in 1 before and in the other after 9 months of age. In the Schwarz group measles developed in 7 of 120 children--in 5 before and in 2 after 9 months of age. Serological responses measured at 5 months after vaccination and at 18 months of age were satisfactory in both groups although in the Schwarz group levels were on average 2-fold higher than in the EZ group. The frequencies of fever, cough, vomiting, and diarrhoea were no higher in the EZ vaccinees in the 3 weeks following vaccination than in age-matched non-immunised controls. Long-term morbidity as assessed by clinic attendances and weight at 18 months of age was much the same in the two groups. The EZ measles vaccine is thus safe and clinically and serologically effective when used in a high dose to immunise young Gambian infants.