Ten systemically healthy subjects (ages 28 to 60 years) with untreated moderate to severe periodontal disease and evidence of presence of A. actinomycetemcomitans underwent standard mechanical periodontal treatment consisting of oral hygiene instruction and systematic deep scaling and root planing. Before, and 4 to 5 weeks after treatment, clinical measurements and separate subgingival microbiological samples were taken from the mesial and distal aspect of every tooth, with the exception of the third molars. A. actinomycetemcomitans could still be detected in all patients after treatment. In 9 of the 10 patients, all tested isolates from both examinations were of a single type. Two patients carried serotype a; 2 serotype b; 2 serotype c; and 1 serotype e. Two individuals showed only non-typeable isolates lacking serotype a, b, c, d, or e specific antigens. Another subject was colonized by serotype c and, in addition, yielded a non-typeable isolate. Persistence of A. actinomycetemcomitans after treatment was significantly correlated with the frequency of A. actinomycetemcomitans before treatment (P < 0.001) and the mean probing depth before treatment (P < 0.05). No serotype-specific patterns of treatment outcome could be recognized. The analysis of the site specific effect of treatment showed a significant relationship between post treatment levels of A. actinomycetemcomitans and both probing depth reduction as well as attachment gain. Individuals showing evidence of A. actinomycetemcomitans in a multitude of sites appeared to be more difficult to treat than patients with few positive sites only. Within such individuals, the deeper pockets showed the greater resistance to eradication of A. actinomycetemcomitans.