A modification of the rheumatoid latex fixation test which involves photometric measurements of the supernatant permitted plotting of agglutination curves and was used to characterize the "reactant" globulin. Both unaggregated electro-phoretically obtained human gamma-globulin and commercially prepared Cohn Fraction II were found to be suitable reactants. Heating Cohn Fraction II reduced the spontaneous precipitation of the coated latex. It was suggested that the heating process, in addition to slightly denaturing Cohn Fraction n, partially deaggregated it. Data were obtained which showed a relatively low and variable protein-binding capacity for latex, from 2.6 to 12.4 μg of gamma-globulin nitrogen/mg of latex. The variation in protein adsorption was attributed to heterogeneity of the gamma-globulin, since protein adsorbed to latex was firmly bound and did not exchange measurably with another protein added later. Despite the evidence that only a small fraction of the gamma-globulin added in the test procedure was adsorbed to latex, excess gamma-globulin did not inhibit the agglutination test. It was therefore postulated that a close apposition of gamma-globulin molecules on the latex surface increased the strength of rheumatoid factor attachment, and that complement and rheumatoid factor attached to the same sites on the aggregated gamma-globulin. Additional evidence given in favor of this concept included the fixation of complement by gamma-globulin coated latex and the presence in fresh serum of a heat-labile inhibitor. On the other hand, data were also presented suggesting slight differences between the specificity of complement and rheumatoid factor for sites on aggregated gamma-globulin.