Polymer erosion of matrices of similarly substituted hydroxypropyl methylcellulose (HPMC) polymers was examined, and drug release in terms of diffusion and erosion contributions was characterized, focusing on matrices containing either polymer alone or a drug content of 25% level with no added excipients. A novel approach was utilized to separate diffusional and erosional contributions to drug release. Diffusional drug release was determined by fitting release data versus (time)0.45, and the drug release due to erosion was quantified by subtracting the percent predicted for diffusional drug release from the total drug release at each specific time point. Drug release resulting from polymer erosion was linear versus time and was found to be a function of the number average molecular weight of the polymer. In contrast, diffusional release rates were comparable for all HPMC grades studied and, thus, were independent of number average molecular weight of the polymers studied. Under stirring conditions of 10-100 rpm as well as static condition, the detachment of individual polymer chains at the matrix surface occurred at a faster rate relative to diffusion away from the matrix surface. The erosion study indicated that polymer diffusion of the HPMC polymer chains through the aqueous diffusion layer was the rate-limiting step for polymer erosion. In general, polymer erosion was found to be inversely related to the polymer number average molecular weight. A scaling law was used to relate polymer erosion rate with the respective polymer number average molecular weight. Similar relationships were obtained for matrices with and without drug at a stirring rate of 100 rpm.