In a double-blind trial two-, three- and five-component acellular vaccines were compared to a British whole-cell vaccine: in areas using three doses at three, five and 12 months of age (3-5-12 schedule), 72,698 children and in areas using a two, four and six months schedule (2-4-6 schedule), 10,194 children were evenly randomized to the four groups. The background incidence of pertussis was higher in the 3-5-12 schedule areas than in the 2-4-6 schedule areas; in spite of this, the point estimates of the relative risks for the 3-5-12 schedule versus the 2-4-6 schedule were close to or below one for the multicomponent acellular and the whole-cell vaccine groups, indicating a lower overall risk of pertussis when the third dose was delayed. The risk of whooping cough according to parents was lowest for the five-component and whole-cell vaccine groups in both schedules. The delayed third dose elicited booster responses for filamentous haemagglutinin but not for the other pertussis antigens. For highly efficacious pertussis vaccines two doses in infancy followed by a third dose in the second year of life may be recommended.