The effect of perfusion of an isolated rat liver on hepatic 3-hydroxy-3-methylglutaryl coenzyme A reductase was studied. In liver removed during the basal period of the diurnal cycle of enzyme activity, a 227 +/- 41% increase in enzyme activity occurred after 3 h of a plasma-free perfusion. This could be prevented by the addition of cycloheximide or pure cholesterol (dispersed with lecithin) to the perfusate. In contrast, the continuous addition of taurocholate or taurochenodeoxycholate, alone or in combination, at a variety of rates did not prevent the increase in enzyme activity. The added bile salts were efficiently extracted from the perfusate and excreted in the bile. The addition of these bile salts to a cholesterol-enriched perfusate did not alter the effect obtained with cholesterol alone. If the perfusate contained whole serum, the increase induced by perfusion in the basal period was smaller (88 +/- 27%) than with plasma-free perfusate. Again, the major bile salts of the rat failed to prevent the increase in enzyme activity induced by liver perfusion. If livers were removed and perfused at the height of the diurnal cycle of enzyme activity, the enzyme activity remained high (2 +/- 10% increase) rather than decreasing, as occurs in vivo. If cholesterol was added to these perfusions, a 52 +/- 4% decrease was induced. Bile salt addition induced no decrease. From the results it is concluded that the major bile salts are not direct regulators of hepatic cholesterol synthesis, but pure cholesterol, in the absence of bile salt or lipoprotein, is able to initiate the mechanism that represses hepatic 3-hydroxy-3-methylglutaryl coenzyme A reductase.