Mark and Saito were the first to fractionate a high polymer (cellulose acetate) by chromatographic adsorption on blood carbon. They found that molecules of low molecular weight were adsorbed first, and that the mean molecular weight of the product which remained unadsorbed was higher than the original molecular weight. Levi and Giera confirmed this result, but did not succeed in fractionating Buna-S or polyisoprene, for these polymers were eluted by the solvent during the washing operation in the column. The present authors have carried out further experiments in this field with a study of three commercial synthetic elastomers, viz., GR-S (butadienestyrene copolymer), Perbunan-N (butadiene-acrylonitrile copolymer), and Visitanex (polyisobutylene). The polymer was adsorbed by starting with a poor solvent composed of a mixture of toluene and methanol ; the quantity of alcohol added was just below the threshold of precipitation. The adsorbent used was a mixture of 75 per cent of lamp black (80 square meters per gram) and 25 per cent of coarse active carbon. The latter served to prevent agglomeration of the lamp black. The adsorbent was divided into three layers, of 10 grams each. At the end of the tube was a filter of fritted glass. Filtration was carried out under pressure, the rate of flow thereby being maintained constant, The polymers were characterized by their intrinsic viscosities. The molecular weights which were estimated by means of the relation, found experimentally between the molecular weight and viscosity, are only approximate, for this relation holds true only for narrow fractions.