Fisher( 2 ) and subsequently Stavitsky ( 7 ) have adapted the tannic acid haemagglutination technique for measuring small amounts of antiproteins as worked out by Boyden ( l ) , for estimating diphtheria antitoxin in human, rabbit, and horse sera. The method consists of sensitizing tannic acid treated red blood corpuscles with diphtheria toxin or toxoid, and agglutinating them with diphtheria antitoxic sera.Landy et al. ( 5 ) using the same method investigated the relationship between the diphtheria antitoxin titres expressed in international units as measured by the intracutaneous test, and the haemagglutination titres, using a large number of human sera and a small number of rabbit sera. They found the relationship of the logarithms of the international antitoxin units per ml and of the haemagglutination titre expressed as logarithm of the reciprocal serum dilution, to be linear. However, the slope of the line was found to deviate from 1, indicating that the titres are not directly proportional throughout the whole range.The established way of measuring the concentration of antitoxin in a serum by the intracutaneous method, and expressing it in terms of the international reference antitoxin-which is a highly avid antitoxin -is, as pointed out particularly by Jerne (4), far from being ideal. He

No abstract