Hot and cold aqueous extracts were prepared from 22 commonly ingested fruits, vegetables, and seeds. When tested by agar diffusion, extracts from 13 and 10 of the foods formed precipitin bands with samples of normal rabbit serum and human saliva, respectively; extracts from four of the foods also reacted with antigen extracts of strains of Streptococcus mutans. When added to rabbit antiserum, extracts from 18 of 21 foods tested inhibited reactivity with antigen extracts derived from S. mutans MT3. Extracts from 16 foods agglutinated whole S. mutans cells, whereas those from 10 foods agglutinated human erythrocytes of blood types A and B. The lectin-like activities of extracts which reacted with human saliva were studied further. Pretreatment of saliva-coated hydroxyapatite (S-HA) beads with extracts of bananas, coconuts, carrots, alfalfa, and sunflower seeds markedly reduced the subsequent adsorption of S. mutans MT3. Pretreatment of S-HA with banana extract also strongly inhibited adsorption of S. mutans H12 and S. sanguis C1, but it had little effect on attachment of Actinomyces naeslundii L13 or A. viscosus LY7. Absorption experiments indicated that the component(s) in banana extract responsible for inhibiting streptococcal adsorption to S-HA was identical to that which bound to human erythrocytes. The banana hemagglutinin exhibited highest activity between pH 7 and 8, and it was inhibited by high concentrations of glucosamine, galactosamine, and, to a lesser extent, mannosamine. Other sugars tested had no effect. The selective bacterial adsorption-inhibiting effect noted for banana extract was also observed in studies with purified lectins. Thus, pretreating S-HA with wheat germ agglutinin and concanavalin A inhibited adsorption of S. mutans MT3 cells, whereas peanut agglutinin, Ulex agglutinin, Dolichos agglutinin, and soybean agglutinin had little effect; none of these lectins affected attachment of A. viscosus LY7. Collectively, the observations suggest that many foods contain lectins which can interact with components of human saliva and S. mutans cells. Because of their potential to influence host-parasite interactions in the mouth and elsewhere in the gastrointestinal canal, these reactions warrant further study.