Samples of apical plaque were taken by means of an anaerobic gas-flushed syringe from 21 sites in eight patients. The samples were anaerobically dispersed, diluted and plated and incubated in an atmosphere of 80% N2, 10% H2 and 10% CO2 for 7-21 days. All colonies on plates containing 20-50 isolates were picked, repeatedly restreaked, characterized and identified where possible by a probabilistic computer identification program. The sites were divided into four groups on the basis of clinical features. The significance of differences between bacterial populations in the groups was determined by the Kruskal Wallis and Mann-Whitney U tests, while the Spearman rank correlation coefficient was used to determine the rank correlation of clinical features of diseases and microbial species. The subgingival microbiota in advanced destructive sites was predominated by Gram-negative rods. The microbiota of two young adult patients with generalized extensive bone loss, extensive clinical inflammation and suppuration was dominated by Bacteroides asaccharolyticus and an organism with characteristics consistent with Actinobacillus actinomycetemcomitans. The predominant cultivable microbiota in two patients with extensive bone loss but minimal clinical inflammation was predominated by Bacteroides melaninogenicus ss intermedius and Eikenella corrodens in one patient and E. corrodens and a slow growing fusiform-shaped Bacteroides in a second patient. A third group of four patients demonstrated moderate levels of clinical inflammation and evidence of continued bone loss in the last year. Predominant organisms in this group were more heterogeneous and included B. asaccharolyticus, Fusobacterium nucleatum, the "fusiform" Bacteroides and anaerobic vibrios. Sites with minimal disease in the patients revealed higher proportions of Gram-positive organisms including Rothia dentocariosa, Actinomyces naeslundii and Actinomyces viscosus. A positive rank correlation could be detected between clinical inflammation including suppuration and B. asaccharolyticus and a negative rank correlation between inflammation and E. corrodens.