A 10 year study of whooping cough in a discrete general practice community was performed to assess longitudinally the efficacy of pertussis vaccine from one to seven years after immunisation. Of the 436 cases of whooping cough over 10 years, 326 occurred in children aged 1-7 years. The rate of immunisation was known for each cohort of children born during each year, and the attack rate of whooping cough was thus calculated for those immunised and unimmunized. The attack rates were highest in those cohorts exposed to the epidemics of 1977-9, 1981-3, and 1985-7. The efficacy of the vaccine was calculated as a percentage as (attack rate in unimmunized group--attack rate in immunised group) x 100/attack rate in unimmunized group. It fell from 100% in the first year to 46% in the seventh, being 84% in the fourth and only 52% in the fifth. Thus the pertussis vaccine or its schedule of use does not seem to provide sufficient herd immunity to prevent outbreaks of whooping cough. Matters might be improved if vaccination against pertussis were included in the preschool immunisation programme.