An investigation of the chemical nature of bacterial toxins must await the isolation of these substances in a state of relative purity. It is unnecessary to emphasize the importance of such an investigation for the purpose of elucidating the nature of toxins as antigens, as substances having marked actions on living cells and tissues, and as poisons for which it may eventually be possible to find chemical antidotes. With these possibilities in mind we have undertaken to isolate, in a pure state, diphtheria toxin which is, up to the present time, the best known and most thoroughly studied of all the bacterial toxins. Precipitation by acid. Of the various methods of purifying diphtheria toxin so far investigated, precipitation by acid seems, on the whole, to have yielded the purest products. Watson and Wallace (1924), Sedallian and Gaumont (1927) and Locke and Maine (1928, 1930) have precipitated the toxin with glacial acetic or hydrochloric acids, removing 98 to 99 per cent of the nitrogenous impurities. Koulikoff and Smirnoff (1927) obtained by the precipitation of toxin with hydrochloric acid at pH 4.7 a product reported to have a "lethal dose" of 0.0006 to 0.00028 mgm. dry weight. Koulikoff and Smirnoff (1927) and Locke and Maine (1928, 1930) have shown that the iso-electric point of the toxin is between pH 4.7 and 5.2, whereas the point of maxim precipitation of inert protein material in the toxic filtrates was between pH 3.8 and 4.6. The work of Schmidt and his col-347