Abstract
There are several impudent assumptions in the determination of molecular weight distributions of high polymers by fractionation. Therefore the reliability of the molecular weight distribution is very problematical. We reconsider the theoretical basis of two fractionation methods, the one is Spencer's method and the other is the successive precipitation method (SPM). Solving simultaneous integral equations, a modified Spencer's method is proposed. Though it necessitates to measure the ratio of the volume of the solution phase to that of the precipitated phase, the results are more accurate than Spencer's method. A modified SPM is also proposed. The results are also compared with the numerical calculation made by other authors and found to be in accord with them.