1. Addition of N-hydroxy-2-acetylaminofluorene, N-hydroxy-4-acetylaminobiphenyl, N-hydroxy-4-chloroacetanilide, and N-hydroxyphenacetin to suspensions of liver microsomes from rabbits and rats treated with phenobarbitone, gave type I binding spectra followed by development of a Soret peak at 455 nm and an increase of absorbance between 400 and 310 nm. The Soret peak was evidence for the formation of a productadduct of cytochrome P-450 with a metabolite, and the increase in the u.v. region indicated the accumulation of a metabolite.2. A distinct lag phase was observed in the kinetics of the Soret peak formation with N-arylacetohydroxamic acids in the absence of NADPH or NADH; the peak was formed instantaneously with arylhydroxylamines or nitrosoarenes, but not with arylamines, N-arylacetamides, nitroarenes or bisazoxyarenes. In the presence of NADPH, arylamines and nitroarenes too caused a small increase in absorbance at 455 nm. This is an indication that the ligand is either the nitrosoarene itself or closely related to it, since it can also be formed following N-hydroxylation of the amino group or reduction of the nitro group. Nitrosoarenes are formed from the N-arylacetohydroxamic acids via N-deacetylation and dehydrogenation of the arylhydroxylamines in the presence of oxygen.3. Paraoxon (10−6 M) and exclusion of oxygen inhibited the formation of the Soret peak and the increase of absorbance between 400 and 310 nm, demonstrating the involvement of microsomal N-arylamidase(s) and oxygen in the sequence of reactions.4. Qualitative and quantitative analysis of the products accumulated during a 30 min incubation of N-arylacetohydroxamic acids with microsomal and cytosolic enzymes of rabbit and rat liver in the absence and presence of NADPH has shown differences in the metabolic pattern, caused by differences in the rate of metabolism of the substrates and the reactivity of the intermediates. The order of metabolism was similar to that of the optical experiments.5. In the presence of oxygen, arylhydroxylamines are oxidized to nitrosoarenes which either react with an excess of arylhydroxylamine to bisazoxyarenes or accumulate. The conversion rate differed among the arylhydroxylamines, rate constants being 2.0 × 10−3 M0.15 s−1 for conversion of N-hydroxy-2-aminofluorene into 2,2′-bisazoxyfluorene, 8.2 × 10−4 s−1 for conversion of N-hydroxy-4-aminobiphenyl into 4,4′-bisazoxybiphenyl, 7.2 × 10−4 s−1 for conversion of N-hydroxy-4-chloroaniline into 4-nitrosochlorobenzene, and 1.5 × 10−3 M0.3 s−1 for conversion of N-hydroxy-4-phenetidine into 4-nitrosophenetole. An apparent reaction order of 1 was found for conversion of N-hydroxy-4-aminobiphenyl and N-hydroxy-4-chloroaniline, and of 0.85 and 0.7 for conversion of N-hydroxy-2-aminofluorene and N-hydroxy-4-phenetidine, respectively.6. Measurement of the utilization of oxygen during the conversion of N-hydroxy-4-chloracetanilide and N-hydroxy-4-chloroaniline into 4-nitrosochlorobenzene and/or 4,4′-bisazoxychlorobenzene has shown that the [substrate]/[O2] ratio was affected by substrate and protein concentration, and that the initial velocity of oxygen uptake decreased in the presence of microsomes. This is evidence for a change of the reaction mechanism caused by the microsomes acting as a diluent of the nitrosoarene.7. Neither the differences in the rate of metabolism of the N-arylacetohydroxamic acids nor the differences in the rate of oxidation of the arylhydroxylamines, but the clear-cut differences in the reactivity of the nitrosoarenes, between the carcinogenic polycyclic and the non-carcinogenic monocyclic compounds, might explain the differences in toxicity.