Recent studies have questioned the efficiency with which administered glucose generates hepatic glycogen through the direct nonrecycling route compared with the indirect route from glucose recycled through glycolysis followed by gluconeogenesis. Using fasted and refed rats, we examined the relative access of infused (l-3H]-and [U-'4Cjglucose by way of these two pathways to liver glycogen and to hepatic glucuronic acid, the latter recovered from the urine as the glucuronide conjugated with administered acetaminophen. In fasted animals and during early refeeding, extensive dilution of administered [3HJ-and []4C~glucose recovered in glycogen showed that 60470% of the labeled glucose had undergone recycling by the indirect route. As refeeding progressed with substantial glycogen deposition, the contribution of the recycling pathway to glycogen and glucuronic acid diminished considerably. Thus, there is a shift in pathways of hepatic glucose utilization as liver glycogen accumulates.Consequently, the ratio of 3H/14C in glucuronic acid was closely correlated with the glycogen content of the liver at sacrifice, indicating that this ratio may prove useful as a noninvasive indicator of liver glycogen concentration. Since glycogen and glucuronic acid are derived by single reactions from UDP-glucose, they should show a common labeling pattern with 3H and 14C under various nutritional conditions. However, detailed analysis of their labeling patterns showed a striking divergence, implying that there must be compartmentation of the UDP-glucose pools leading to each of these end products, either because they are made in separate compartments within the same cell or because each is made in different hepatocyte populations. We favor the former explanation because galactose secreted in glycoproteins shows 3H and d4C labeling patterns similar to those of glucuronic acid conjugated with acetaminophen, and both of these conjugations occur in the endoplasmic reticulum of the liver, whereas most glycogen is present in the cytosol.A growing body of evidence (1) suggests that the intact glucose molecule is not directlyt used as such by the liver of animals refed after a fast, despite the net entry of ingested glucose carbon into the liver for glycogen deposition, lipid synthesis, etc. The proposed resolution of this "glucose paradox" is that administered glucose is first degraded to triose metabolites, presumably in the peripheral tissues, which return to the liver in the form of lactate, alanine, and pyruvate to regenerate glucose phosphates and sugar nucleotides through the gluconeogenic pathway (Fig. 1) Fig. 1) and thus lowering the 3H/14C ratio of the glucose deposited in liver glycogen. This preferential loss of 3H from administered glucose was observed in only one such study (7) but not in others (16-19), suggesting direct utilization ofglucose for glycogen deposition. However, when the specific activities of the glucosyl residues of glycogen in such studies were compared with those of the parent blood glucose, a loss of u...