A general approach to analyze associations of the type mA + nB = AmBn at sedimentation equilibrium under ideal conditions was derived. A simple transform to generate a function Q was defined in which the optical measurement (absorbance or refractive index increment) at a given radial distance is divided by the value that would result if no association occurred. Explicit calculations for the cases of m = 1, n = 1 and m = 2, n = 1 were used to simulate centrifuge experiments. Nonlinear least-squares regression of Q vs. the optical measurement showed that accurate estimation of parameters could be obtained in selected instances for both types of association only if the appropriate model was used. Additionally, the effect of random experimental and systematic error (e.g., the presence of contaminants) was evaluated.