The immunogenicity and adverse effects of an acellular pertussis vaccine consisting of a purified pertussis toxin inactivated with hydrogen peroxide (PTxd) was evaluated. Children aged 15 to 30 months were injected with 10 (n = 33) or 50 micrograms (n = 34) of PTxd or with diphtheria and tetanus toxoids and whole cell pertussis vaccine (DTP) (n = 34). All children had previously received three doses of DTP during infancy. Both dosages of PTxd induced higher IgG antibody (p less than 0.05 for 10 micrograms dose and p less than 0.01 for 50 micrograms dose) and pertussis antitoxin responses (p less than 0.01 for 50 micrograms dose) than DTP. The 50 micrograms dose gave slightly higher (though not significantly) antibody responses than the 10 micrograms dose of PTxd. None of the vaccines induced detectable IgM or IgA antibody responses to pertussis toxin. At 24 h, local reactions occurred in none of the children injected with 10 micrograms PTxd, 12% with 50 micrograms PTxd and 78% with DTP. Fever at 24 h occurred in 13% after 10 micrograms PTxd, in none after 50 micrograms PTxd and in 53% after DTP. Recipients of DTP, but not of PTxd, had significant increases in neutrophils and decreases in lymphocytes and haematocrit at 24 h (all p less than 0.05). None of the groups showed changes in blood glucose at 24 h. PTxd induced pertussis toxin antibody levels similar to those observed in patients convalescing from natural pertussis. This acellular pertussis vaccine deserves further evaluation for safety and immunogenicity in infants and for efficacy in preventing pertussis.