To quantify the extent to which exogenous glucose is used directly or indirectly for hepatic glycogen synthesis, fasted rats were given [U-14C,3-3H] glucose intragastrically, intravenously, or as a component of a solid diet eaten ad libitum. In all cases liver glycogen was deposited at high linear rates over a 3-h period. Portal vein glucose levels seldom exceeded 8 mM. At a time when the specific activities of 3H and 14C in circulating glucose were identical with those in the administered material their values in newly synthesized glycogen were reduced by 72-88% and 50-65%, respectively. An intragastric load of unlabeled glucose sufficient to suppress completely hepatic glucose output greatly stimulated the incorporation of intravenously infused [14C]bicarbonate, [14C]lactate, [14C]alanine, and [14C] glutamine into liver glycogen. Using an improved assay the ability of liver homogenates to phosphorylate glucose at concentrations of 5-10 mM was found to be far short of what would be needed if glucose were used directly to support hepatic glycogen synthesis in vivo. These data support the notion that in the rat a major fraction of liver glycogen deposited in response to exogenous carbohydrate is formed by a pathway involving glucose leads to C3 unit leads to glycogen, although the site of the initial steps in the sequence is not yet known. The limited capacity of the liver to utilize intact glucose for glycogen synthesis might reside in its limited capacity to phosphorylate the sugar at physiological concentrations.