Over a 12-year period, 16 human strains of a gram-negative, catalase-positive, halophilic, aerobic, nonmotile, small coccoid bacterium were received for identification. On the bases of biochemical characteristics and cellular fatty acid profiles, 14 of these strains were similar to the "Philomiragia" bacterium (Yersinia philomiragia, species incertae sedis). Additional characteristics were growth on Thayer-Martin agar but no growth or sparse, delayed growth on MacConkey agar; oxidase positive; acid production, often weak and delayed, from D-glucose, sucrose, and maltose; urease negative; no reduction of nitrates; and H2S produced but often delayed in triple sugar iron agar. Both the human isolates and the "Philomiragia" bacterium contained C10:0, C14:0, C16:0, C18:1 omega 9c, C18:0, 3-OH C18:0, C22:0, and C24:1 as major cellular fatty acids and ubiquinone eight (Q8) as the major isoprenoid quinone. These cellular acids in these relative amounts have been found previously only in Francisella tularensis and Francisella novicida, suggesting a relationship between the "Philomiragia" bacterium and Francisella species. Of the 14 human "Philomiragia"-like isolates, 9 were from blood, 3 were from lung biopsies or pleural fluid, and one each was from peritoneal fluid and cerebrospinal fluid. DNA relatedness studies (hydroxyapatite method, 50 and 65 degrees C) showed that these 14 strains were a single group that was the same species as the "Philomiragia" bacterium. Two other human strains were oxidase negative and H2S negative. They formed a single DNA relatedness group that was indistinguishable from the type strains of both F. tularensis and F. novicida. DNA relatedness of "Philomiragia" bacterium type and other strains to strains of F. novicida and F. tularensis, including the type strains, was 35 to 46%. One of the two F. novicida- and F. tularensis-like strains was isolated from blood, and the other was isolated from a cervical lymph node. On the basis of these findings, we propose transferring Y. philomiragia from the genus Yersinia to the genus Francisella as Francisella philomiragia comb. nov. Having confirmed that F novicida and F. tularensis are the same species and having shown that F. novicida is pathogenic for humans, we further propose eliminating the species F. novicida and demoting it to a biogroup of F. tularensis.