Evidence is presented that the previously observed decrease of the Vmax of hepatic microsomal demethylation of dimethylnitrosamine (DMN), following pretreatment of rats with 3-methylcholanthrene (MC), is not due to increase in the rate of breakdown but to decrease of de novo synthesis. Determinations of Vmax at time intervals in the transition from the high steady-state level induced by a carbohydrate-devoid casein diet, down to the low steady-state level of carbohydrate-containing basal diet, yielded two consecutive slopes; descent from the basal diet level to the lower steady-state level following pretreatment with MC yielded one slope. Plotting these slopes against the initial Vmax values gave a typical exponential curve (or straight line if the logs of slopes are used) indicating that the rate of enzyme decay in the MC-treated animals is not greater than that expected from normal enzyme catabolism. A multiplicity of effector sites appears to be involved in the repressor action of different structural types; for example, repression by MC (46.6%) and by phenobarbital (23.9%) in combination are approximately additive (62.0%), rather than competitive, indicating that the two agents act at different sites. A P-450 type cytochrome is involved in the demethylation of DMN. DMN-demethylase is inhibited by carbon monoxide, but the susceptibility to CO is far greater than that observed previously with 3,4-benzopyrene hydroxylation; inhibition of DMN-demethylase as a function of CO concentration follows typical enzyme kinetics. However, while both phenobarbital and MC powerfully repress the DMN-demethylase, we have confirmed that they are strong inducers of the synthesis of P-450 and P-448, respectively, as estimated from the difference spectra.