[ 171 ] 



XXIII.— A THERMO-ELECTRIC CURRENT IN SINGLE CON- 

 DUCTORS. By FRED. T. TROUTON, B.A. 



[Eead, March 24, 1886.] 



If a flame be placed under an iron wire in circuit with a galvano- 

 meter, and be so moved along the wire that the part in the flame is 

 always white-hot, a current is indicated which flows in the direction 

 the flame is carried. The electromotive force is generally in the 

 fourth decimal place. In looking for an explanation of this, it 

 was observed that in front of the flame the fall in the temperature 

 along the wire is more rapid or steeper than behind it. So that, if 

 a difference in the rate of transference of heat in opposite direc- 

 tions in a wire causes a current, as some have supposed, there would 

 be one in this case. The rate of flow of heat is greatest in the 

 direction the flame is moved, for the fall in temperature along the 

 wire is most rapid in that direction. The current, then, and the 

 greatest flow of heat are in the same direction. The amount of the 

 current would thus obviously depend on the difference of the gra- 

 dients in temperature in either direction along the wire. By making 

 the gradient in front as steep as possible, and that behind the flame 

 more gradual, we should expect an increase in the current. Or 

 again, by making the gradient behind the flame steeper, by cooling 

 it more rapidly ^than the air can, say by applying water, we should 

 get a decrease in the current, and even a reversal if the gradient 

 became steeper than in front. It was with no small surprise, then, 

 that the opposite was observed on trying the experiment. For, 

 cooling with water behind the flame as it moved along was found 

 to increase the current. That this could not be due to chemical 

 action was ascertained by applying various substances to cool the 

 wire. Thus, whether bodies of a reducing or oxydizing nature 

 were employed the result was always the same. 



