Data are presented on fretting corrosion of mild steel using weight loss as a measure of damage. Relative slip was measured both by use of stroboscopic light and strain gages. Load was applied pneumatically. Humidity of ambient air was found to be a sensitive variable requiring control, a fact first discovered through discrepancies of weight-loss data obtained in winter compared with those obtained in summer. Fretting corrosion of mild steel in moist air is only 55 to 65 per cent of the weight loss in dry air, depending on duration of test. Because of this large difference, all tests were carried out in dry air. At the same time, it was observed that fretting corrosion is appreciably greater below room temperature than above room temperature (up to 150 C). Weight losses of specimens fretted at 50 C are approximately 50 per cent losses at 0 C. Rate of fretting corrosion in air is found to be constant with time after an initial run-in period during which the rate is temporarily higher. In moist or dry nitrogen, weight losses are much less, although not zero. The greater the relative slip, the greater is fretting damage, and in complete absence of slip the data indicate that no weight loss occurs. Increased pressure or load is found to increase fretting damage. Weight loss is greater the lower the frequency for the same number of test cycles. The frequency effect increases with relative slip, and in nitrogen the frequency effect disappears.