Serum samples from41 periodontally healthy childrenaged 1 to 16 years were examined by ELISA for the presence of antibodies against a glass bead‐EDTA cell surface extract (GBE) and LPS of Porphyromonas gingivalisstrain ATCC 33277. P. gingivaliswas detected by immunofluorescence, using a species‐specific monoclonal antibody, in 41% (17/41) of the children, and isolated from a single subject (2.4%). IgM, IgG, and IgA against GBE were detected in respectively 39/41 (95%), 41/41 (100%), and 27/41 (66%) of the sera. In 22/39 sera, the IgG titer was below 50% that of a reference pool of adult sera (RP). In 13/41, the IgM titer was higher than that of the RP, mostly in the deciduous dentition group. Detectable IgA titers were always below 67% that of the RP. A polarized distribution of the children appeared, separating 21 non‐ and low IgA responders (IgA titer below 10% that of the RP) from the remaining 20 subjects. Anti‐LPS IgG, IgM, and IgA were detected in 41/41 (100%), 39/41 (95%), and 23/38 (61%) respectively of the children. In 32/41 sera, the anti‐LPS IgG titer was below 50% that of the RP, while in 20/39 sera, IgM titers were higher. A clearcut dichotomy in IgA response was observed, allowing us to distinguish non‐IgA responders (39%) and IgA responders to LPS (61%). Our results indicate that serum antibodies to P. gingivalisare highly prevalent in children, suggesting that an active primary immune response and a secondary immune response are well underway. Discordant results between detection rate, either by culture or by immunofluorescence, and prevalence of antibodies suggest that few but highly immunogenic cells of P. gingivalisare distributed at discrete sites in the oral cavity where they escape sampling for detection procedures. Thus our data support the hypothesis of P. gingivalisbeing a member of the endogenous oral microflora in humans. J Periodontol 1995; 66:369–376.