It is generally recognized that bacteria in dental plaque at sites of periodontal diseases are not commonly found at sites of periodontal health. One hypothesis to explain the etiology of periodontitis is that pathogenic bacteria from diseased sites infect healthy sites. It has been suggested that gingival inflammation may predispose sites to colonization by bacteria associated with periodontal diseases. The purpose of this cross-sectional study was to determine whether the detection frequency of selected bacteria at sites of periodontal health or gingivitis differed between subjects who were in good periodontal health, subjects who had gingivitis, or subjects with periodontitis. The clinical status of every tooth (except third molars) from 106 subjects was characterized by means of clinical attachment level, probing depth and by signs of inflammation. Subgingival plaque was collected from mesio-facial and disto-lingual surfaces. Specific monoclonal antibodies were used in an immunocytochemical assay to identify Campylobacter rectus, Eikenella corrodens, Porphyromonas gingivalis, pathogen-related oral spirochetes (PROS, using Treponema pallidum subspecies pallidum monoclonal antibodies), T. denticola (serotypes A-D), T. socranskii subspecies buccale and T. socranskii subspecies socranskii. Differences in detection of bacteria between groups of subjects were measured using odds ratios (OR). Results of this study indicate that PROS was the only identified bacterium at sites of both health and gingivitis that demonstrated a significant positive relationship with the presence of periodontitis. These findings do not prove that bacteria spread from periodontitis sites, nor do they imply that disease necessarily results from infection. However, these data do suggest that some bacteria associated with periodontitis are more likely than others to tolerate conditions at healthy sites and that the presence of periodontitis increases risk of infection at healthy sites.