SYNOPSIS Using only 12 antisera it has been possible to group 56-6 of all strains. Forty-four per cent of all strains were typable in the first eight of the serological groups used. Of the primary isolates from symptomatic infections, 65X7 % were typable; 48-6 % of all strains of the serological groups 01, 02, 04, and 06 occurred in these primary symptomatic isolates.There was a good correlation in the serological typing of five colonies isolated from the initial isolation plates. There was complete agreement in all the typable strains, 325 of the 350 colonies. Twenty-five strains were not typable due to roughness. It is considered that this reproducibility confirms the validity of using one colony for serotyping of urinary isolates.Statistically there was no significant difference in the number of typable strains from specimens with high bacterial counts and specimens with low bacterial counts. Nor was there any difference in frequency of the typability of strains in patients with primary symptomatic infection compared with those with associated urinary tract abnormalities. There was a possibly significant increased frequency of typable strains in the primary infection compared with the recurrent infection. There was no significant difference in the frequency of specific serological groups in the high and low bacterial count groups. There was no statistical difference in the frequency of individual serological groups isolated from the two sexes.Where Escherichia coli has been found epidemiologically associatedwith disease, as in infantile gastroenteritis, the pathogenic significance of certain serological types has been noted. Kauffmann (1947) suggested that certain serological groups of Escherichia coli were isolated more frequently from the urine than from the faeces. Vahlne (1945) and Sjostedt (1946) considered that although certain serological groups were found more commonly in human infections these were not limited to the urinary tract infections. More recently suggested that certain serological groups were more invasive for the urinary tract than other types and that these caused clinically evident pyelonephritis more frequently than other serological groups. Turck and Petersdorf (1962) confirmed that certain groups were associated with non-enteric infections but suggested that this may be due to an increased prevalence in the environment. The majority of patients studied in this paper were not exposed to the risk of nosocomial infection; thus there is little possibility of the incidence of serological groups being biased by the inclusion of environmentally prevalent strains.
MATERIALS AND METHODS
Five hundred and thirty-four strains of Escherichia coliReceived for publication 5 February 1965. were investigated; 316 were from urines with bacterial counts more than 100,000 per ml., 86 from the intermediate count group, and 132 from counts below 10,000 organisms per ml. They were all identified biochemically as Escherichia coli by routine methods (Edwards and Ewing, 1962 These strains were grown on...