SUMMARY Factor I of the anthrax toxin was isolated and showed one major component in the ultracentrifuge and on paper electrophoresis; it contained less 5.5% of extraneous antigens detectable by serological precipitation in gels. The final preparation contained all the usual amino acids (N = 10.1%) and some carbohydrate (6%, calculated as glucose) and phosphorus (0.7%). The most striking aspects of its analysis were a high ash (10-13%) and a light absorption at 260 mμ. The high ash was not due to one element but to a highly variable metal content (mainly Ca, Mg, Ni, Cu) indicating a powerful and indiscriminate chelating action of factor I. This chelating action might have been due to the chemical entity which absorbed light at 260 mμ and which was not RNA or DNA. The final preparation of factor I was not toxic when injected alone but when mixed with purified factor II it evoked oedema in the skin of a rabbit and killed mice. However, the concentrationof this mixture which killed mice formed a much larger skin reaction in rabbits than a comparable dose (based on mouse LD 50) of either curde toxin or a mixture of crude factors I and II. An investigationof this fact led toteh demonstraion and partial purification of a third factor (III) of the anthrax toxin which: (1) was different serologically from factors I and II; (2) was present in anthrax toxin produced in vivo; (3) was non-toxic when injected alone; (4) was lethal for mice when mixed with factor II but not with factor I; (5) increased the lethality of mixtures of factors I and II for mice and decreased their capacity to produce oedema in the skin of rabbits. A mixture of factors I, II and III showed synergic action in toxicity tests in mice; the mixture killed guinea pigs which showed signs of ligaemic secondary shock (as did guinea pigs killed by anthrax infection).