1. Five anto-horse sera which did not contain antibodies for horse-serum crystalbumin have been shown by absorption experiments to contain horse-serum crystalbumin.
2. All five sera precipitated when mixes with other anti-horse sera containing anti-crystalbumin, precipitation being due to the presence of horse-serum crystalbumin in one antiserum and of anti-crystalbumin in the other.
3. Antigen and its homologous antibody were never found together in the serum of an animal, and contradictory results in other experiments are probably due to impure multiple antigens.
4. It is concluded that anti-horse sera may contain some of the antigenic components of injected horse serum together with antibodies to other antigens of the horse serum, but not homologous antigen and antibody. Consequently, the ‘mutual’ precipitation of anti-horse sera is due to the presence of a number of antigens in horse serum, one or more of which, present in one anti-horse serum, in the absence of its homologous antibody, may precipitate when mixed with another anti-horse serum which contains the homologous antibody.
5. The use in these experiments of a single α-procedure optimum as the indicator of an antigen-antibody reaction illustrates a method enabling the investigation of problems involving single antigen-antibody reactions, even though the available reagents consist of mistures of antigens and mixtures of antibodies.