AbstractIn this paper we present a rigorous argument for the validity of a new technique for obtaining greatly increased resolution of the chromatograms of a gel permeation chromatograph. The method is phenomenological in that no assumptions are made about the separation process, although certain simple assumptions are made about the characteristics of this process. Instead, the method makes use of the observable behavior of narrow polymer fractions of presumably known composition. From a pragmatic point of view, this is an important aspect of the method, since the results predicted by the method depend only upon what the chromatograph is observed to do, and not upon what it is assumed to do. This method has been implemented in FORTRAN IV for the IBM 7094 computer.