Data for long-term marine immersion and half-tide corrosion loss and pitting of chromium-bearing and stainless steels is systematically re-examined and re-interpreted to show that it is consistent with the model previously proposed for mild and low-alloy steel corrosion. Early corrosion of chromium-bearing steels is well known to be influenced by bacterial activity in the biofilm adjacent to the metal surface. However, the effect appears short-lived and typically the corrosion rate declines, presumably due to the buildup of corrosion products, either on the surface or within pits. The data considered herein show that after some two to four years of exposure there is usually a considerable increase in the instantaneous corrosion rate both for weight loss and for pitting. This is interpreted in terms of the previously proposed model for the corrosion of steels as the result of bacterial activity but now located primarily within pits and crevices. This interpretation implies that the early and the later corrosion processes involve different mechanisms and that corrosion loss and pit depth data are not all from the same homogeneous population. As a result, the prediction of long-term corrosion from short-term laboratory or field observations is of doubtful validity. This has important implications for the prediction of the effective life of chromium bearing and stainless steels in infrastructure applications.