A comparison of the s- and w-detections of molecular weight, a necessary condition for precise determination of interdetector volume in size exclusion chromatography of polymers with the dual light scattering--concentration detection, is examined. From the theoretical analysis for a polymer with the log-normal molecular weight distribution it follows that the error in determination of the local calibration, i.e., of a logarithmic dependence of molecular weight on elution volume obtained by the s-detection, from the known dependence of radius of gyration on elution volume and of molecular weight on the radius of gyration, remains linear and is slightly shifted with respect to that obtained by the w-detection (dual detection giving the weight-average molecular weight) towards higher molecular weights, the difference being below the experimental error. It also follows that the error in the slope of local calibration found by the w-detection is given by the ratio of the error in interdetector volume and the variance of the elution curve. This is demonstrated on several polymer samples with symmetric elution curves of polymer samples differing in polydispersity indices. The range of this rule depends on the broadness and symmetry of elution curves. The precision of the interdetector volume determination depends therefore strongly on the Mw/Mn ratio of the sample.