1. The release of 57Co-labelled vitamin B12 ([57Co]B12) and synthesis of transcobalamin II (TCII) by the isolated perfused rat liver were studied 10–42 days after the parenteral administration of a trace dose of 15 pmol (approximately 20 ng) of radioactive cyanocobalamin. 2. The rate of release of [57Co]B12 into plasma and bile was linear and constituted approximately 0.9% and 0.3% respectively of the initial hepatic radioactivity per hour of perfusion. 3. [57Co]B12 released into plasma was bound to TCII. Saturation of the total TCII by the addition of cyanocobalamin before perfusion resulted in the appearance of the hepatic [57Co]B12 in the free form. 4. These data were found to be compatible with the following observations in vivo: (i) rates of [57Co]B12 release as measured by urinary [57Co]B12 excretion after saturation of plasma binders with non-labelled cyanocobalamin; (ii) rates of biliary excretion of [57Co]B12. 5. Liver damage produced by hypoxaemia was associated with a fall in the rate of release of [57Co]B12. 6. TCII release occurred at a linear rate of almost twenty times that required for the binding of newly released hepatic vitamin B12. 7. Cycloheximide at a dose sufficient to inhibit release of TCII did not prevent the release of [57Co]B12 from the liver into plasma or bile. 8. Alteration of perfusate composition to contain either high plasma concentrations of vitamin B12 and low concentrations of unsaturated TCII or high plasma concentrations of vitamin B12 and high concentrations of unsaturated TCII had no effect on the rate of [57Co]B12 release into plasma or bile. 9. It is concluded that the fluxes of hepatic vitamin B12 and TCII are very rapid and that the release of vitamin B12 by the rat liver is controlled in the short term by factors other than the synthesis of TCII and the concentration of vitamin B12 or unsaturated transcobalamin in the plasma.