Amongst other techniques, dynamic light scattering may be used to obtain molar mass distributions. The first step in this process consists in the Laplace inversion of the time correlation function that was measured by dynamic light scattering. This inversion gives a distribution of diffusion coefficients. In order to convert this distribution into the corresponding molar mass distribution, a relationship between diffusion coefficient and molar mass of monodisperse fractions has to be known. Such a relationship can be derived for linear and star-branched macromolecules from measurements of polydisperse systems, since the polydispersity of the distributions does not change with the molar mass. The problem is more involved with randomly branched materials, since in these cases the polydispersity increases strongly as the point of gelation is approached. A procedure is suggested for deriving the diffusion-molar mass dependence of monodisperse samples from polydisperse systems. After an outline of this background the method is applied to the three selected systems (i) radically polymerized linear PMMA, (ii) a star-branched microgel where monodisperse arms are attached to a microgel centre and (iii) a randomly branched poly(dicyanate) sample based on bisphenol A. The results are compared with the combined column chromatography SEC/LALLS/VISC. Good agreement was found up to molar masses of about 10 millions g/mol, but systematic deviations occurred in the high molar mass region. These differences result from the limitations of size permeation chromatography. Finally it is shown that the size distribution can be determind by this method, even for associates.