Summary: The serum-agar technic of Oudin was compared with the commonly used method of supernatant analysis for the determination of the homogeneity of serological systems. There was a complete correspondence between the results obtained by the two methods under the experimental conditions used. The conclusion of Oudin that the number of bands of precipitate formed indicate the least number of antigen-antibody systems that could possibly be present was substantiated by using mixtures of antigens and antisera of known immunological composition. The findings presented confirm previous observations that the movement of the bands is proportional to the square root of the time. Some limitations of the serum-agar technic and its possible applications were pointed out. A technic is described to correlate a given band of precipitate to a given antigen.