Studies were carried out on conscious female non-pregnant (NP) and pregnant (P; third-trimester) dogs (n 16; eight animals per group) to define the role of the liver in mixed meal disposition with arteriovenous difference and tracer techniques. Hepatic and hindlimb substrate disposal was assessed for 390 min during and after an intragastric mixed meal infusion labelled with [14C]glucose. The P dogs exhibited postprandial hyperglycaemia compared with NP dogs (area under the curve (AUC; change from basal over 390 min) of arterial plasma glucose: 86 680 (sem 12 140) and 187 990 (sem 33 990) mg/l in NP and P dogs, respectively; P < 0·05). Plasma insulin concentrations did not differ significantly between the groups (AUC: 88 230 (sem 16 314) and 69 750 (sem 19 512) pmol/l in NP and P dogs, respectively). Net hepatic glucose uptake totalled 3691 (sem 508) v. 5081 (sem 1145) mg/100 g liver in NP and P dogs, respectively (P = 0·38). The AUC of glucose oxidation by the gut and hindlimb were not different in NP and P dogs, but hepatic glucose oxidation (84 (sem 13) v. 206 (sem 30) mg/100 g liver) and glycogen synthesis (0·4 (sem 0·5) v. 26 (sem 0·7) g/100 g liver) were greater in P dogs (P < 0·05). The proportion of hepatic glycogen deposited via the direct pathway did not differ between the groups. Hindlimb glucose uptake and skeletal muscle glycogen synthesis was similar between the groups, although final glycogen concentrations were higher in NP dogs (9·6 (sem 0·6) v. 70 (sem 0·6) mg/g muscle; P < 0·05). Thus, hepatic glucose oxidation and glycogen storage were augmented in late pregnancy. Enhanced hepatic glycogen storage following a meal probably facilitates the maintenance of an adequate glucose supply to maternal and fetal tissues during the post-absorptive period.