In the fasted to fed transition liver glycogen derives mainly from gluconeogenic precursors. Why glucose is not used efficiently as a direct precursor of glycogen has become a controversial issue, in part because of disagreement over the question of how well liver can phosphorylate glucose under conditions prevailing postprandially. To try to resolve the matter the relative merits of two recently described assays, one spectrophotometric (A), the other isotopic (B), for monitoring rates of glucose phosphorylation in the high speed supernatant fraction of liver have been rigorously evaluated. A third method, also isotopic (C), was developed for use with unfractionated as well as fractionated liver homogenates. Using fasted rats and mice from different nutritional backgrounds the glucose-phosphorylating capacity of liver extracts was measured and compared with rates of hepatic glycogen synthesis observed during refeeding. Two of the assays (A and C) provided reliable data at all concentrations of glucose tested (5-100 mM), while method B exhibited shortcomings at lower substrate concentrations. The results clearly establish that in both rats and mice the ability of the liver to phosphorylate glucose at physiological concentrations is sufficient to support only 25-30% of postprandial glycogen synthesis. A limited capacity for glucose phosphorylation probably accounts for the fact that two-thirds of glycogen synthesized with refeeding after a fast is formed by the indirect (gluconeogenic) pathway.