If the heating curve of a coil or a machine were exponential the final temperature-rise could be calculated from the initial portion of the curve and the heating test thereby shortened. A large number of heating curves have been examined, and it is found that the final temperature, as calculated from the rate of increase of resistance, is always less than—often considerably less than—the actual value. The greater part of the paper is devoted to an analysis of the influence upon the shape of the curve, and thereby upon the ratio of the calculated to the actual rise, of such factors as the variation of (a) speed, (b)power generated in winding, (c)the temperature-rise per watt as affected by radiation and natural convection, (d)thermal conductivity, (e)specific heat, (f)point of maximum temperature and (g) heating time-constant.