of Iron when under Stress or Strain. 47 



caused a deflection of the galvanometer, indicating a current 

 from the strained to the unstrained portion through the hot 

 junction, as already found by Sir William Thomson. The 

 flame of the burner, which was placed underneath the junc- 

 tion, was at first kept small, but was afterwards increased 

 by degrees. At about dull red temperature the deflection 

 nearly ceased to increase, but as soon as the wire assumed 

 a bright red colour, the light reflected from the mirror of the 

 galvanometer went off the scale. The adjusting magnet 

 was therefore used to bring the spot of light back on to 

 the scale ;in the hope of getting a fairly steady deflection; 

 but this was found to be impossible, the light frequently 

 changing its position by fits and starts and sometimes going 

 over to the other side of the scale. The reason of this un- 

 steadiness will be explained later on. 



The experiment was repeated with a second piece of the 

 same wire similarly treated, with this difference, however, 

 that after slightly heating the junction to observe the 

 direction of the deflection, a large resistance was introduced 

 into the galvanometer circuit. By this means the variations 

 of the sudden deflection again produced at the temperature 

 of bright red were much diminished, and a moment of fair 

 steadiness was seized to remove the burner and allow the wire 

 to cool. Almost immediately after the removal of the burner 

 the light on the scale, after pausing slightly, went suddenly 

 in the opposite direction. 



The experiment was repeated some six or seven times with 

 different pieces of the same wire, but always with like results. 



Experiments similar to the above were tried with other 

 pieces of the same iron wire, to which permanent strain was 

 imparted by hammering transversely, by stretching longi- 

 tudinally, or by bending the wire into coils of small diameter. 

 In all these experiments the therm oelectrical currents pro- 

 duced by moderate heating were, as Sir W. Thomson has 

 already found, from the strained to the unstrained through 

 the hot junction, and in all there was a very sudden increase 

 of the current when, on being heated, the wire became bright 

 red, and an equally sudden decrease when the wire was 

 cooled. 



Experiment II. 



Two pieces of the same wire were arranged as in fig. 1 (p. 48). 

 A x B 1} A 2 B 2 are two wires brazed together at C and looped at 

 AijBx and A 2 , B 2 ; from B x and B 2 the wires continue to D, 

 where they are joined to the connecting wires of the galva- 

 nometer G ; D was well wrapped up in tissue-paper. B x Si, 



