The cathode reaction is one of the most seriously limiting factors in a microbial fuel cell (MFC). The critical dissolved oxygen (DO) concentration of a platinum-loaded graphite electrode was reported as 2.2 mg/l, about 10-fold higher than an aerobic bacterium. A series of MFCs were run with the cathode compartment inoculated with activated sludge (biotic) or not (abiotic) on platinum-loaded or bare graphite electrodes. At the beginning of the operation, the current values from MFCs with a biocathode and abiotic cathode were $2.3{pm}0.1$ and $2.6{pm}0.2mA$, respectively, at the air-saturated water supply in the cathode. The current from MFCs with an abiotic cathode did not change, but that of MFCs with a biotic cathode increased to 3.0 mA after 8 weeks. The coulomb efficiency was 59.6% in the MFCs with a biotic cathode, much higher than the value of 15.6% of the abiotic cathode. When the DO supply was reduced, the current from MFCs with an abiotic cathode decreased more sharply than in those with a biotic cathode. When the respiratory inhibitor azide was added to the catholyte, the current decreased in MFCs with a biotic cathode but did not change in MFCs with an abiotic cathode. The power density was higher in MFCs with a biotic cathode ($430W/m^3$ cathode compartment) than the abiotic cathode MFC ($257W/m^3$ cathode compartment). Electron microscopic observation revealed nanowire structures in biofilms that developed on both the anode and on the biocathode. These results show that an electron-consuming bacterial consortium can be used as a cathode catalyst to improve the cathode reaction.