Subgingival plaque samples were taken from active and inactive lesions in 33 subjects exhibiting active destructive periodontal diseases. Active diseased sites were those which showed a significant loss of attachment within a 2-month interval as computed by the "tolerance method". The predominant cultivable species from 100 active sites were compared with those found in 150 inactive sites of comparable pocket depth and attachment level loss. Among the 33 subjects, W. recta, B. intermedius, F. nucleatum, B. gingivalis and B. forsythus were elevated more often in active sites; whereas, S. mitis, C. ochracea, S. sanguis II, V. parvula and an unnamed Actinomyces sp. were elevated in inactive sites. The likelihood of a site being active was increased if B. forsythus, B. gingivalis, P. micros, A. actinomycetemcomitans, W. recta, or B. intermedius were detected in that site, and decreased if S. sanguis II, the Actinomyces sp., or C. ochracea were detected.