A rapid, simple and reliable technique for determining the affinities of antibody subpopulations in a complex mixture is described. The principle of this method is that antigen cone can be represented as the amount of antigen immobilized on the polystyrene surface of a microwell containing a fixed vol of diluted antibody. Thus, by measuring the proportion of antibody bound to different wells coated with varying amounts of antigen, it is a straightforward matter to calculate an affinity distribution. We have verified that: (1) the amount of antigen bound to a polystyrene plate is proportional to the concn of antigen used for sensitization and follows a typical saturation curve; (2) the antibodies bound to plates sensitized with low concns of antigen are of higher affinity than those bound to plates sensitized with high concns of antigen; (3) an apparent affinity constant (aK) is defined as the reciprocal concn of free hapten required for 50% inhibition of antibody binding to immobilized antigen; (4) the aK determined by this method is in close agreement with the intrinsic affinity constant (K) measured by fluorescence quenching or the Farr assay; and (5) that during the course of immunization in vivo there is a clear shift to higher-affinity antibody subpopulations.