Strains of four streptococcal types, 33, 41, 43, 52, and a nontypable strain, Ross, cross-reacted in precipitin and bactericidal tests. The homologous reactions, which determined the type, afforded the major protection and developed promptly and regularly in the serum of rabbits during immunization. The associated cross-reactions, on the other hand, appeared in the serum of certain rabbits only, were often not as strong as the associated homologous reactions, and required for their presence a longer period of immunization than the homologous reactions.

Agar gel analysis of the homologous precipitin reactions revealed, as would be expected, reactions of serological identity, while those cross-reactions which were strong enough to test in this way formed bands of precipitate which joined with spur formation on the side of the homologous reaction.

These experiments and others referred to in the text suggest that cross-protection, as demonstrated in bactericidal tests, is sufficiently widespread to be a factor in streptococcal immunity, if a corresponding protection occurs in vivo. Thus, streptococcal infection with one of the cross-reacting strains might confer, in addition to strong homologous protection, a certain amount of cross-protection.