A collection of 120 strains of saccharolytic intestinal Bacteroides was examined for plasmid deoxyribonucleic acid (DNA). These strains were previously characterized for standard phenotypic properties (variety of carbohydrates fermented, metabolic end products, and complex carbohydrate hydrolysis). Additional phenotypic characterizations included bacteriocin production, cell wall analysis, and antibiotic susceptibilities. The interrelatedness of the strains had been determined by DNA homology experiments. The strains were distributed among 10 major DNA homology groups or genospecies. Plasmid DNA was found in 52 strains (43%) distributed among all ten major DNA homology groups. The plasmids ranged in size from 2 to 65 megadaltons (Mdal) as estimated from their migration distance in agarose gels during electrophoresis. A small plasmid of 2 to 6 Mdal was found in 42 of the plasmid bearing strains. Seventeen of these strains also contained larger plasmids (20 to 65 Mdal). Ten strains contained only a large plasmid. There was no apparent correlation between the presence of any particular plasmid size with the DNA homology grouping of the host strains. Size estimations calculated from plasmid DNA reassociation curves (C₀t curves) agreed with the agarose gel electrophoresis size determinations. Four plasmids were selected for reference use in DNA homology studies. They were isolated from B. fragilis strains 2553 and 4076, B. thetaiotaomicron strain 5482, and B. uniformis strain T1-1 respectively. These plasmid preparations were radioactively labeled in vitro using the 4 deoxyribonucleic triphosphates (³H labeled thymidine 5'- triphosphate) and Escherichia coli B polymerase I. The percent guanine plus cytosine content of two duplexed plasmid DNA preparations was determined by comparing their thermal stabilities with those of duplexed DNA from three bacterial strains whose percent guanine plus cytosine contents were known. Plasmid DNA p4076 and p5482 were found to have a percent guanine plus cytosine content of 39 and 43 respectively. Three of the reference plasmids (p2553, p4076, and p5482) were large species (22 to 23 Mdal) and did not have any significant base sequence homology with one another. The small reference plasmid (pTl-1, 3 Mdal) had 30% DNA homology with the p4076 reference plasmid. Plasmid DNA preparations from all the plasmid carrying strains were allowed to renature with each of the labeled reference plasmids. Many of these plasmid DNA preparations contained multiple plasmid species. The pT1-1 reference plasmid had moderate (30 to 60%) or higher (61 to 100%) DNA homology with 35 of the plasmid DNA preparations. The reference plasmid DNA, p2553, had moderate or high homology with 10 unlabeled plasmid preparations while the p5482 reference plasmid had similar homology with 9. The p4076 reference plasmid did not have significant base sequence homology with any of the plasmid DNA preparations. Several strains contained plasmid DNA that had homology with both pTl—1 reference plasmid and one (but not both) of the larger reference plasmids (p2553 and p5482). The homology between the unlabeled plasmid DNA preparations and the two reference plasmids may be due to either separate or hybrid plasmids. This was not determined. A computer analysis of 81 phenotypic traits was made, comparing each with the plasmid DNA homology groups. No significant correlation was noted. Plasmid—mediation of bacteriocin production was considered, but there was no correlation between bacteriocinogenesis and the occurrence of any plasmid DNA homology group. / Ph. D.