Our early 31P n.m.r. studies of compartmentation in suspensions of rat liver cells have been extended by following fructose-1-phosphate peaks, known to be in the cytosol, which gave the same pH as the P1 peak previously assigned to the cytosol. Gluconeogenesis has been followed from [13C]glycerol labelled at C1,3 or at C2 and from labelled [3-13C] alanine. With the glycerol substrate it was possible to follow the label into α-glycerophosphate and to determine its distribution in the glucose formed. To a first approximation (i.e. 90 %) the glucose label could be followed from its original glycerol position, e.g. [ 1,3-13C]glycerol to strongly labelled positions 1, 3, 4 and 6 of glucose. Slightly more than 10% of the label was scrambled (i.e. 10% movement of C2 to C1 and ca. 10% of C1 was lost, the remainder being unchanged). These are consistent with a flux through the pentose shunt, dominated by the transketolase pathway. With [3-13C]alanine, about 14 resonances are assigned to different carbons of the intermediates β-hydroxybutyrate, acetoacetate, lactate, pyruvate, glutamate, glutamine, asparate, as well as C2-alanine, while another 7 resonances are observed from the different anomeric carbons of glucose. The effects of thyroid hormone treatment of the rats upon numerous in vivo rates are clearly observed and will be illustrated.