The author points out that determination of the mean antitoxin titre is inadequate for ascertaining fine differences in the activity of diphtheria and tetanus toxoids, but that these differences become clearly recognizable when the relation is established between the antigen dose and the proportion of individuals whose antitoxin titre exceeds a given "threshold titre" or neutralizes a given amount of toxin.Furthermore, the author states that it is not feasible to assay crude and highly purified diphtheria toxoids against the same preparation, and that two reference preparations are therefore required.