90 Profs. J. A. Fleming and J. Dewar. On the 



with certain conclusions of Messrs. Laws and Warren (see ' Pro- 

 ceedings of the American Academy of Sciences,' vol. 30, p. 490). 

 These observers made a series of experiments on a material which 

 was practically a very soft steel, and employing a differential watfc 

 meter, measured the hysteresis loss in the iron at varying and 

 increasing temperatures up to 600 or 700. They found that the 

 hysteresis loss in this material did not begin to decrease sensibly 

 until about 150 C. ; after that it decreased regularly in accordance 

 with the simple linear function of the temperature. In one experi- 

 ment which they tried with the same material cooled to 78 C. in 

 solid carbonic acid and ether, they found no difference between the 

 hysteresis loss of this soft steel at that temperature and at the ordinary 

 temperatures. Our results, which have been carried to the much 

 lower temperature of liquid air, indicate that in the case of soft 

 annealed Swedish iron the hysteresis loss is not changed by cooling 

 from ordinary temperatures to the temperature of liquid air. As we 

 know that the hysteresis loss of soft iron decreases when the tem- 

 perature is increased, from the ordinary experience with transformers 

 in commercial use, the matter that requires further investigation is to 

 discover the temperature at which the hysteresis loss sensibly changes 

 and begins to diminish. 



Experiments on Unannealed Swedish Iron. 



We have also carried out a series of experiments of the same 

 character with unannealed iron and steel. A ring coil was con- 

 structed of sheet iron of the same quality as that forming the core of 

 the soft iron transformer above described, but no special pains were 

 taken to anneal the iron, and as it was " hardened " in a magnetic 

 sense by being bent into shape, this difference in quality showed 

 itself in the magnetic observations. A ring coil was constructed of 

 the following dimensions : The thickness of the strip was 031 cm., 

 width of the strip 1/24 cm., the ring was formed by 23-| layers of 

 this sheet iron wound up closely into the form of a ring. The out- 

 side diameter of this ring was 4 cm., the inside diameter 2' 13 cm., 

 the cross-section of the iron in the ring was therefore 0'9032 sq. cm., 

 and the mean perimeter of the ring 9*62 cm. This iron ring was not 

 annealed in any way, but it was simply wound over with silk ribbon, 

 and then had placed upon it two coils of wire. The primary coil 

 consisted of 150 turns of No. 26 wire, having a resistance of 

 0'383 ohm, and the secondary coil consisted of 240 turns of No. 36 

 wire having a resistance of 8*092 ohms. As the diameter of cross- 

 section of the ring was not very small compared with the mean 

 diameter of the ring, it was necessary to calculate by a proper 

 integration the mean value of the mean magnetising force in terms 





