Pure iron powder, reduced in hydrogen, was left for a week in oxygen‐free, distilled water at 257° C. in silica, pyrex, and soft glass flasks through which hydrogen bubbled continuously. After three days a black deposit began to form on the flask at the water surface and at the tip of the inlet tube. Magnetic analysis showed that the deposit contained magnetite in all of the flasks, and x‐ray analysis showed its presence in two flasks. Magnetite is therefore the final product of corrosion in the absence of oxygen, as found by a number of others. The reason iron with a smooth surface does not react is that a smooth iron surface has a higher overvoltage than the voltage corresponding to the small free energy decrease of the reaction, while the rough surface of the powder does not.