174 Scientific Proceedings, Royal Dublin Society. 



perature when at rest, which is shown on either side in the figure 

 by the dotted lines of equal length. The arrow shows the direc- 

 tion the flame is moving. 



If water be applied to one side of a heated part of the wire, 

 there is a very rapid cooling and a current flows, due to the higher 

 temperature on the other side, until the junction slowly travels 

 in to the point at the normal temperature, when the current ceases, 

 as was described above. 1 Again, on stopping the water, the current 

 flows for a while in the opposite direction. 



So far it has been assumed that the iron, as the flame moves on, 

 cools and returns to its original state ; but it does not do so com- 

 pletely. However, if the flame be passed several times over the 

 same part, the iron seems after that to undergo no further altera- 

 tion. There is a permanent heterogeneousness or alteration found 

 from the place where the heating by the flame began to where it 

 ended, similar to that between the wire and another metal. Either 

 of the ends of this gives a current on heating. In some cases, 

 especially in steel, the currents were easily observed, even at 100° C. 

 The current, as in the case of the temporary alteration in the wire, 

 flows from the altered to the unaltered metal at the hot junction. 

 It follows from this, that the first time the flame is moved along 

 the wire the current is somewhat greater than subsequently, 

 it being the sum of both effects ; though afterwards it appears not 

 to alter sensibly in amount on repeated heatings. That currents 

 due to permanent alteration in the structure of metals could be 

 obtained, was long ago shown by Magnus. 



Of other metals examined, nickel acts like iron ; copper, silver, 

 and platinum appear not to — that is, an alteration once made in 

 their structure remains on cooling ; while iron and nickel return 

 partly to their original state. This difference may be owing to 

 the more or less pasty condition iron and nickel assume at tempe- 

 ratures considerably below their melting points; and, probably, 

 both copper and silver raised to temperatures just beneath their 

 melting points would behave like iron or nickel. The difficulty of 



1 That Le Eonx did not observe these currents map be due to his not employing as 

 high a temperature. 



