Serum samples were obtained from 44 infants vaccinated against diphtheria at the ages of 3, 5 and 12 months with an aluminium-adsorbed diphtheria-tetanus toxoid vaccine containing 15 Lf units of diphtheria toxoid. Toxin-neutralizing antibodies (antitoxin) were measured by the Vero cell assay and IgG, IgM and IgA antibodies against diphtheria toxoid by enzyme-linked immunosorbent assay. A neutralizing antibody titer of 10 corresponded to 0.01 IU/ml, the level considered necessary for short-term protection. Geometric mean neutralizing antibody titers at 3, 5, 6, 12, 13 and 30 months were 28, 21, 173, 61, 1076 and 61. All children had titers of greater than or equal to 10 (greater than or equal to 0.01 IU/ml) between 6 and 30 months of age. At 30 months only 48% had titers of greater than or equal to 100 (greater than or equal to 0.01 IU/ml), the level considered necessary for long-term protection. Geometric mean IgG antibody levels were 13, 36, 216, 64, 649 and 57. IgG antibodies significantly correlated with neutralizing titers and predicted neutralizing antibodies above or below 10 and 100 with an accuracy of 96 and 82%, respectively. IgG antibodies could not, however, be used to predict individual neutralizing antibody titers with great accuracy. IgM antibodies were only detected after the third vaccination. IgA antibodies were not detected in any serum sample from ten infants tested. In conclusion, the Swedish vaccination schedule results in protective antibody levels in infants until at least 30 months of age. The decline of the antibody titers indicates a need for further studies to establish the duration of protection.