l06 Scientijic Proceedings, Royal Dublin Society. 



Ao initial twist of 180° was just within tlie elastic limit of the wire, but 

 beyond that angle the wire was permanently deformed as indicated by tlie 

 curve becoming flat. 



Nagaoka' has described an interesting experiment designed to show the 

 Wiedemann effect to a large audience by means of capillary ripples on tlie 

 surface of mercury ; and important results have been obtained by Gray and 

 Wood^ on the effect of a magnetic field on the rate of subsidence of torsional 

 oscillations in iron. 



In order to carry out experiments to test the effect of stress on the 

 material as indicated in the first statement on page 102, another iron wire 

 of the same gauge, No. 16, was taken and carefully prepared and annealed. 

 One end of the wire was fixed to a support and stretched horizontally with a 

 string over a pulley having a weight of about 7 kilos, attached. The flame of 

 a good bunsen burner was tlien passed slowly over the wire, raising it to a 

 bright red heat ; by this means all the kinks were taken out of tlie wire. It 

 was then hung loosely in a horizontal direction and the bunsen again passed 

 slowly over it twice, which left the wire annealed or as soft as the nature of 

 the material would permit. 



The wire was then cleaned with emery cloth and placed in position in the 

 Earth's vertical field, 0-45 c.g.s. units, as indicated in fig. 1 ; a vibrator or 

 weight of a certain definite amount was attached, and tlie galvanometer put 

 in circuit. 



Then by means of the handles P the wire was ultimately put through a 

 complete cycle in steps of 20° each ; at every 20° mark in the scale on the 

 top of the guide-frame F a small hole was made to hold a stout pin which 

 was moved along at each step and served as a stop or guide for the handle P. 

 Thus the pin being in the 20° mark the handle P was moved from 0° to 20° 

 and the throw on the galvanometer read ; then tlie pin was moved into the 

 40° mark and the handle moved from 20° to 40°, and the throw on the 

 galvanometer scale again read off, and so on right round the complete 

 cycle. 



The throws or deflections on the galvanometer scale were read off at each 

 of the nine steps from 0° to 180°, and from this point the wire was put 

 through four complete cycles before reading the scale further — -in order to 

 bring the wire to a proper cyclic state. That being done, the throws on the 

 scale were again observed at each step from 180 to 0, and on the other side 

 from to 180, then back from 180 to 0, and finally from to 180. Then 

 summing up the galvanometer throws or transient currents, we have a measure 



' Nature, vol. Ixix., p. 487. - Proc. Roy. Soc, vol. Ixxiii., p. 286. 



