The mammalian mitochondrial electron transport chain catalyzes the oxidation of NADH at pH 8.0 and pH 6.5, and the oxidation of NADPH at pH 6.5. The pH-dependencies of the rate of steady-state oxidation of NADPH and NADH by Complex I as well as by its flavoprotein fraction have been extensively studied by the laboratory of Hatefi. One model to explain these pH-dependent oxidations was proposed by Bakker and Albracht (Biochim. Biophys. Acta 850 (1986) 413–422 and 423–428, modified by Van Belzen and Albracht (Biochim. Biophys Acta 974 (1989) 311–320), which predicts that Complex I is a heterodimer with protomer B, containing FMN and FeS clusters 1–4 in stiochiometric amounts, catalyzing NADH oxidation at pH 8, and Protomer A, containing FMN and FeS clusters 2, 4, catalyzing NAD(P)H oxidation at pH 6.5. A pH-dependent transfer of electrons from protomer A FeS clusters 2, 4 to protomer B FeS clusters 2, 4 is an obligate step in the oxidation of NAD(P)H at low pH. Strict interpretation of this model allows for only three types of inhibitor: one which inhibits all three oxidase activities (type 1); one which inhibits NADH oxidase, pH 8.0 (type 4) and a third which inhibits NAD(P)H oxidase, pH 6.5 (type 5). Another possibility is that there are three separate pathways of oxidation of NAD(P)H, which would allow for a total of seven different types of inhibitor, e.g., the three types above plus type 2 inhibiting NADH oxidase pH 8.0 and pH 6.5; type 3 inhibiting NADH oxidase pH 8.0, and NADPH oxidase pH 6.5; type 6 inhibiting NADH oxidase pH 6.5; and type 7 inhibiting NADPH oxidase pH 6.5. Using a series of thirteen inhibitors of Complex I activity and the chemical modification reagent ethoxyformic anhydride (EFA), four different inhibitor types were found: seven inhibitors of type 1, four inhibitors of type 2, one inhibitor of type 3 and one inhibitor of type 4. Treatment of submitochondrial particles (SMP) with EFA abolished NADH-dependent reduction of coenzyme Q at both pH 8.0 and 6.5, while inhibiting NADPH-dependent reduction of coenzyme Q at pH 6.5 by only 30%. These results do not support the heterodimer model of Complex I electron transport of Bakker and Albracht, but do support three separate electron flow pathways through complex 1 from reduced pyridine nucleotides to coenzyme Q. A new model of electron flow through Complex I based on these finding is proposed.