Plaque samples from caries-active subjects showed a higher incidence of S. mutans than plaque samples from caries-free subjects. This was especially evident in approximal incisor plaque. S. mutans serotype d was almost exclusively present in approximal plaque obtained from caries-active subjects. Tooth surfaces infected with S. mutans still harbored this micro-organism 10 months later, while uninfected tooth surfaces remained free of S. mutans. Caries development predominantly occurs on those tooth surfaces which harbor relatively high percentages of S. mutans (> 5%). It is unlikely that serum or saliva antibodies against S. mutans play a major role in the protection against dental caries in these caries-free subjects since subjects with the greatest number of decayed surfaces showed the highest antibody titre as measured by haemagglutination or by the enzyme-linked immuno sorbent assay (ELISA).