I t was a decade and a half ago when our laboratory initiated its investigation on the chemical and immunological nature of the colicins.These potent antibacterial substances, which are elaborated by a variety of enteric microorganisms, had never been fully characterized since their discovery by Gratia (1) in 1925. To be sure, over the years there have appeared assertions that these are proteins, and perhaps for good reason, for their biological activity was known to be destroyed by proteolytic enzymes (2). Yet a search of the literature, until but a few years ago, yielded no other convincing evidence that this was indeed their true nature. More recently this situation has changed. There are now several investigators who have obtained highly purified colicins, the syntheses of which are induced with the antibiotic mitomycin C, a reagent which kills the growing colicinogenic bacillus and radically alters its metabolic processes, but which permits, indeed even enhances, the synthesis of the bacteriocin (3-5). From such cultures there have been obtained bacteriocins which are unquestionably protein in nature. One must not lose sight of the fact, however, that the conditions under which these bacteriocins are elaborated are considerably different from those of microorganisms growing in a medium devoid of mitomycin.Our earlier work with Escherichia coli K235 (6) and with a colicinogenic strain of Shigella sonnei phase I I (7), as well as with a colon bacillus which elaborates colicin V (8), has all pointed toward the fact that the type-specific O antigen of these bacilli is in each instance bactericidal. It has been our premise that the colicins are an integral part of the O antigenic complexes and that they cannot be separated from them by electrophoretic or chromatographic means (9). It must be emphasized, however, that we have always been alert to the possibility that colicin protein and O antigen are in some way intimately bound, and that their separation might not have been achieved by the procedures employed.Early in our work we showed that the colicin K derived from E. coli K235 elaborated two types of antibodies, one which precipitated the substance, and the other which neutralized its bactericidal activity, apparently without