Vol. 30, no. 3, p. 583, column 2: The following acknowledgment was inadvertently omitted: "We thank John G. Holt for his helpful discussions about Latin and Greek names and for his assistance in deciding on a specific epithet." 109

In 1972 there were only 11 genera and 26 species in the family Enterobacteriaceae. Today there are 22 genera, 69 species, and 29 biogroups or Enteric Groups. This paper is a review of all of the new organisms. It has a series of differential charts to assist in identification and a large chart with the reactions of 98 different organisms for 47 tests often used in identification. A simplified version of this chart gives the most common species and tests most often used for identification. The sources of the new organisms are listed, and their role in human disease is discussed. Fourteen new groups ofEnterobacteriaceae are described for the first time. These new groups are biochemically distinct from previously described species, biogroups, and Enteric Groups of Enterobacteriaceae. The new groups are Citrobacter amalonaticus biogroup 1, Klebsiella group 47 (indole positive, ornithine positive),

Enterobacter sakazakii is the name proposed for the organism previously known as "yellow-pigmented Enterobacter cloacae." The type strain (holotype) of this species is ATCC 29544. The proposed change in the classification of this organism is based on differences between E. cloacae and E. sakazakii in deoxyribonucleic acid (DNA) -DNA hybridization, biochemical reactions, pigment production, and antibiotic susceptibility. By DNA hybridization, E. sakazakii was about 50% related to E. cloacae, Citrobacter diversus ("Citrobacter intermedius" biotype b), and "Citro bacter amalonaticus" ("Citro bacter intermedius" biotype a). The new species was placed in Enterobacter rather than Citrobacter because of its closer phenotypic and DNA similarity to E. cloacae, the type species of the genus Enterobacter, and because it was only 41% related by DNA hybridization to Citrobacter freundii, the type species of Citrobacter. E. sakazakii had biochemical reactions very similar to those of E. cloacae but was D-sorbitol negative and positive for extracellular deoxyribonuclease at 2 to 7 days and produced yellow-pigmented colonies. E. sakazakii had larger zones of inhibition around ampicillin and cephalothin antibiotic disks, which also helps to differentiate it from E. cloacae. E. sakazakii grew on the nonselective (but differential) plating media commonly used in enteric bacteriology, but its plating efficiency was reduced on more inhibitory enteric plating media. It has been isolated from human clinical specimens such as sputum, feces, and wounds, where it is probably only a colonizer and not clinically sigruficant. However, it is also a documented, although rare, cause of neonatal meningitis. Other sources have included food, a physician's stethoscope, and an uninoculated blood culture bottle.As new knowledge accumulates, changes in the classification of members of the family "Entwo bacteriaceae" (scientific names in quotation marks were not included in the Approved Lists "241) become necessary (4,8,9, 12, 19,21,23). A few of these changes have involved consolidation of organisms previously thought to be distinct species. The four Shigella species and Escherichia coli although phenotypically distinct, are closely related by other criteria and could be considered as belonging to the m e species (7). Similarly, Klebsiella ozaenae and Klebsiella rhinoscleromatis are comprised of host-adapted strains p patho-bio-sero-geogroups") which belong in the same species with KlebsielZapneumoniae (11). Other changes have involved "splitting" rather than "lumping." New species previously thought to be biogroups (biotypes or biovars) within a recognized species have been proposed recently. Examples include Klebsiella oxytoca, previously thought to be an indole-positive biogroup of K. pneumoniae (19); Proteus tThe participants and their affiliations are listed in the Acknowledgments at the end of this paper.

Escherichiafergusonii (formerly known as Enteric Group 10) and Enterobacter taylorae (formerly known as Enteric Group 19) are proposed as new species in the family Enterobacteriaceae. By DNA hybridization (32p, 60°C, hydroxyapatite), strains of E. fergusonii were 90 to 97% related to the type strain (holotype) ATCC 35469. They were most closely related to Escherichia coli and more distantly related to species in other genera. E. fergusonii strains are positive for indole production, methyl red, lysine decarboxylase, ornithine decarboxylase, and motility. They ferment D-glucose with gas production and also ferment adonitol, L-arabinose, L-rhamnose, maltose, D-xylose, trehalose, cellobiose, and D-arabitol. They are negative for Voges-Proskauer, citrate utilization (17% positive), urea hydrolysis, phenylalanine deamination, arginine dihydrolase, growth in KCN, and fermentation of lactose, sucrose, myo-inositol, D-sorbitol, raffinose, and a-methyl-D-glucoside. By DNA hybridization (32p, 60°C, hydroxyapatite), strains of E. taylorae were 84 to 95% related to the type strain

of the non-Cedecea reference strains used in the DNA relatedness study have been reported previously (3,4,11). Methods. A suspension of strain 005 in distilled water (about 0.02 ml) was deposited on a carboncoated specimen grid, negatively stained with 2% (Wt/ Asbury, Abst. Annu. Meet. Am. SOC. Microbial. VOl) phosphotungstic acid, and exambed with a Sicby the method of Clowes and Hayes (5). Growth on various compounds as carbon and energy sources was determined by multiple inoculation (9) onto M70 minimal medium (14) supplemented with 0.01 pg of thiamine per ml, 100 pg of cysteine per ml, 0.5 pg ofpara-317 318 GRIMONT E T AL. INT. J. SYST. BACTERIOL. TABLE 1. Origins of the Cedecea strains studied Strain designations Origin" Source Location This paper Centers for Disease Species Control C