Wells, Graham and Havens,1have described the precipitation of diphtheria toxoid with aluminum potassium sulphate, following the method of Glenny and Barr.2Havens and Wells3have recently reported further work with completely precipitated toxoid. To the toxoid, prepared in the usual manner, by detoxification with formaldehyde, alum was added until no further precipitation occurred. The coarse, flocculent precipitate was washed twice with 0.85 per cent sodium chloride solution and the washed precipitate was then made up to such a volume with 0.85 per cent sodium chloride solution that 1 cc. contained the desired number of flocculating units. This treatment resulted in a purified product, as evidenced by the fact that nitrogen determinations showed an average loss of from 80 to 85 per cent of the original protein content. Antigenic tests in guinea-pigs indicated that a high degree of immunity is produced, a single injection of 5 units