46 Mr. H. Tomlinson on the Thermoelectrical Properties 



slight bending stress it suddenly unbent. Corresponding 

 sudden changes in the opposite direction occurred on cooling. 

 When, however, the wire was permanently strained and 

 relieved from stress the equally sudden changes which ensued 

 at the critical temperature were opposite in direction to the 

 changes produced when the wire was under stress ; namely, 

 a permanently twisted wire free from torsional stress would 

 suddenly twist further when, on heating, a temperature of 

 bright red was reached, and a permanently bent wire would 

 suddenly bend further. Likewise the sudden changes which 

 took place on cooling were in opposite directions for stress 

 and permanent strain : the permanently twisted wire suddenly 

 untwisted, and the permanently bent wire unbent. 



Those sudden changes were regarded as probably indicating 

 equally sudden changes in the molecular architecture of iron ; 

 and experimental evidence was brought forward to prove 

 that the temperature at which the above-mentioned phenomena 

 occurred was not the same as that at which iron suddenly 

 loses its magnetic properties. 



Again, Sir W. Thomson* arrived at the remarkable con- 

 clusion that when a permanent strain is left after the with- 

 drawal of the stress producing it, the residual thermoelectrical 

 effect is the reverse of the thermoelectrical effect which is 

 induced by the stress and which subsists as long as the stress 

 acts. 



The author was led by the above-mentioned considerations 

 to attempt to ascertain how far the thermoelectrical properties 

 of iron when under temporary stress or permanent strain 

 might be affected by raising the temperature to a bright red. 



Experiment I. 



A piece of well-annealed iron wire one millimetre in 

 diameter was subjected throughout half its length to a great 

 many turns of permanent torsion. It was then supported on 

 the ring of a retort-stand, being insulated from the latter 

 by paper, with the junction of the twisted and untwisted 

 portions in the centre of the ring. The free ends of the 

 iron wire were connected by silk-covered copper wire with 

 the terminals of a delicate reflecting-galvanometer of about 

 7 B.A. units resistance; the two junctions between the iron 

 and copper being tied together and well wrapped up in tissue- 

 paper, which served also to insulate them from each other. 

 The junction of the twisted and untwisted portions of the iron 

 wire was now heated slightly by a Bunsen's burner ; this 



* « Electro-dynamic Qualities of Metals," Phil. Trans. Part IV. 1856. 



