500 PROCEEDINGS OP THE AMERICAN ACADEMY. 



Inspection of the plots shows (Figs. 4 and 5) that for this sample of 

 steel a moderate elevation of temperature up to about 270° has but a 

 small influence on the hysteresis loss ; that at this point a change takes 

 place, and the diminution of hysteresis, with rise of temperature, is 

 much more rapid, the decrease being such that at 675° the loss is but 

 13 per cent of its value at ordinary temperatures. After heating 

 above the inflection point, the specimen shows on cooling a marked 

 decrease of the hysteresis loss from its value at the same temperature 

 on the ascending curve. If the specimen be heated again, a change of 

 direction of the curve takes place, but at a lower temperature. The 

 second plot (Fig. 5) shows an increase of the hysteresis loss after the 

 point of inflection has been reached. As only two heatings were made, 

 we were unable to verify these observations, the results of which became 

 apparent only when they were reduced. In addition to the experi- 

 ment at high temperatures, the following observations were made. A 

 specimen was made up and the loss measured at the ordinary tempera- 

 tures ; a mean of three measurements gave .289 Watts as Wh. The 

 specimen was then covered with a mixture of solid carbonic acid and 

 ether ; a mean of three measurements showed Wh as .288 Watts, 

 showing that there is practically no change of hysteresis for a diminu- 

 tion of temperature to about — 80°. 



Between the ascending and descending observations of the second 

 heating, readings were taken for the purpose of finding whether the 

 temperature at which the iron becomes markedly magnetic on coolino- 

 is the same as that at which the sudden diminution of the mas:netic 

 properties occurs on heating. Simultaneous readings of the B gal- 

 vanometer and pyrometer were taken, the current remaining constant, 

 no adjustment for maximum was made. The readings, together with 

 some from the descending curve of the first heating, are given below, 

 and, after being corrected for zero, are plotted in Figure 6. 



