into acetic-acid precipitates has been studied with a preparation from cells ground with alumina. The time-course of mannose incorporation was biphasic (rapid reaction for 2-5min followed by a slower, linear reaction for 5-20min) and the incorporation was stimulated by uridine diphosphate glucose (UDP-Glc) during the early part of the reaction. The reactions (mannose incorporation after 2 min and 15 min, & UDP-Glc) were nevertheless proportional to protein concentration. [14C]Glucose was incorporated into acetic-acid precipitates from UDP-[14C]GIc and this reaction was slightly stimulated by GDP-Man. Radioactivity from GDP-[14C]Man was incorporated into butanol extracts and the early part of the reaction (2-5 min) was markedly stimulated by UDP-Glc. Bacitracin did not inhibit the incorporation of mannose into acetic-acid precipitates in the early part of the reaction, but did inhibit incorporation about 2001, between 5 and 30 min.Larger-scale incubation mixtures were extracted with 4501, aqueous phenol. The product was of high molecular weight as indicated by Sephadex G-100 chromatography and yielded a single radioactive peak corresponding to mannose after total hydrolysis and paper chromatography. This high-molecular-weight product precipitated with anti-09 antiserum in an antigen-binding assay, indicating that it was serologically active. Thus, it is likely that the 0 9 antigen of E . wli 0 9 : K30(A) : H12 is synthesized from GDP-Man and UDP-Glc through a butanol-extractable (lipid-linked) intermediate. Reactions appear to be analogous to those described for lipopolysaccharide biosynthesis in Xalmonella.I n Escherichia coli, 150 different 0 antigens are known. These are lipopolysaccharides and have been typed according to their sugar composition [l]. I n addition to the O-antigenic lipopolysaccharides about 100 different K antigens are known [2]. Combinations of these surface antigens on the bacterial cell give rise to an enormous variety of E . coli serotypes. These combinations, however, are subject to a restriction not yet understood. All K antigens of the A-type [Z], i.e. such K antigens which are true serologically-distinct capsules of unique acidic polysaccharides and which are readily lost by mutation, occur only with E.coli cells bearing either antigen 08 or 09. We became interested in studying the biosynthesis of O-antigenic lipopolysaccharides and Kantigenic acidic polysaccharides of E . coli strains 0 8 and 0 9 to investigate the relationship of the K and 0 antigens.