206 PROCEEDINGS OF THE AMERICAN ACADEMY 



thirty times a second by an interrupter, the thermometer showed a 

 marked rise in temperature. Grove * surrounded a horseshoe electro- 

 magnet with water at a constant temperature, phiced a thermo element 

 upon the armature, which was carefully covered with flannel, and found 

 an increase of temperature upon interrupting or reversing the electrical 

 current. He found no effect in non-magnetic metals. Villari f endeav- 

 ored to show the same effect from transversal magnetism, by passing 

 an interrupted current through two wires, one of lead, and the other 

 of iron or steel. These wires were bent in the middle, placed in glass 

 cylinders, so that their ends projected through corks at each end of 

 the cylinders. The glass cylinders were filled with absolute alcohol, 

 and were provided with capillary tubes, and thus served as spirit 

 thermometers. The cylinders were placed parallel to each other. 

 The circuit was arranged so that the interrupted current preserved 

 always the same direction in the lead wire, while at each interruption 

 it was reversed through the iron or steel wire. The lead wire showed 

 no increase of heat, while the iron wire exhibited a marked rise of 

 temperature. These various experiments can be explained on the 

 theory that the heating is due to induction currents generated in the 

 mass of the iron by its sudden introduction into a magnetic field. 

 And the effect is greater in the iron and steel than in non-magnetic 

 metals, from the reason that the coefficient of induction is very greatly 

 increased by the substitution of an iron core in an electro-magnet for 

 a non-magnetic core. 



It is claimed by some investigators, however, that the heating is due 

 to a greater degree to the movinij or twisting of the molecules of the 

 iron in their beds while under the influence of the alternating and 

 magnetizing currents. Indeed, this view is apparently held by the ob- 

 servers Avhose experiments we have quoted. And a late paper, by W. 

 Siemens $ leads the reader to infer that the internal friction among the 

 molecules of the iron contributes largely to the heating due to rever- 

 sals of magnetism. The phenomena which support the view that the 

 heat observed when a bar of iron or steel is magnetized and demagnet- 

 ized is due to molecular action, are the following. The bar gives forth 

 a musical note, which is attributed to longitudinal vibrations of the 

 bar. A bar of steel is lengthened when magnetized. Magnetic 

 filings distributed upon the poles of the electro-magnet move tremu- 



* Pliil. Mag., XXXV. 153, 1849; Vo^s- Ann., Ixxviii. 567. 

 t Nuovo Cimento, Ser. II., Vol. IV., Nnv.-Dec, 1870. 

 j: Ann. der Pliysik und Cliemie, No. 12, 1881, pp. 635-656. 



