In the early stages of iron oxide growth on mild steel in pH 11 LiOH solution at 300C, the oxide is largely oriented and grows at a rate and with a morphology dependent on the substrate grain orientation. After a certain film thickness is attained, however, the epitaxial oxide degenerates, leaving in its place a much thinner, fine-grained, randomly oriented 'base film' partially covered by large solution-grown Fe3O4 crystals. Experimental evidence indicates that the transformation is induced by stress generated within the epitaxial oxide as the film thickens. The two-phased oxide structure which exists after completion of the epitaxial oxide degeneration persists through oxidation periods of at least 300 hours. Observations of its growth behavior lead to the postulation of a growth mechanism in which iron passes through the base film into solution, subsequently to precipitate forming the upper surface crystals. It is possible, using this hypothesis, to derive a theoretical rate law, which predicts that, in agreement with the experimentally determined corrosion rate data, the corrosion rate should decrease in a logarithmic fashion.