1. The rates of diazepam (DZ) metabolism to the primary metabolites 3-hydroxydiazepam, 4'-hydroxydiazepam and nordiazepam were studied in vitro using rat hepatic microsomes and hepatocytes. 4'-hydroxydiazepam had the largest intrinsic clearance (Vmax/Km ratio, CL(int)) in both microsomes and hepatocytes representing 49 and 70% of total metabolism respectively. Whereas the contribution of 3-hydroxydiazepam was similar in both systems (21-24%), the N-demethylation pathway was greater in microsomes (27%) than hepatocytes (9%). 2. The pharmacokinetics of DZ were determined in vivo using the intraportal route to avoid blood flow limitations due to the high clearance of DZ. No dose dependency was observed in either clearance or steady state volume of distribution, which were estimated to be 38 ml/min/SRW (where SRW is a standard rat weight of 250 g) and 1.3 L/SRW respectively. Blood binding of DZ was concentration independent, the unbound fraction being 0.22. 3. Scaling factors were used to relate the in vitro CL(int) to the in vivo unbound clearance. Hepatocytes (123 ml/min/SRW) produced a more realistic prediction for the in vivo value (174 ml/min/SRW) than microsomes (41 ml/min/SRW). This situation is believed to arise from the quantitative differences in the three metabolic pathways in the two in vitro systems. It is speculated that end product inhibition is responsible for reduced total metabolism in microsomes whereas hepatocytes operate kinetically in a manner close to in vivo.