Incidental to a vaccine study involving 783 immunized children conducted at two study sites, inner city children had significantly higher geometric mean pertussis agglutinin titers compared with suburban children just before the fourth dose of diphtheria-tetanus-whole cell pertussis vaccine (47 vs. 25; P less than 0.001). Higher titers in the inner city were correlated with residence in census tracts where cases of pertussis were reported. Three hundred thirty-two children in a placebo arm of the study who were clinically observed and had paired serum samples taken during a 2- to 4-month period were analyzed for evidence of natural Bordetella infection. Twelve (11%) inner city children and three (1.3%) suburban children had spontaneous 4-fold or greater rises in at least two different pertussis antibodies measured (agglutinin, antitoxin or enzyme-linked immunosorbent assay for IgG to pertussis toxin, IgG and IgA to filamentous hemagglutinin). Eighty percent of these children had IgA to filamentous hemagglutinin. Nine of 12 inner city children with serologic evidence of pertussis lived within 6 blocks of a case of pertussis reported within 1 month of the observed antibody rise in study subjects; none had a household member with pertussis and none had symptomatic disease.