On the Temperatures of Conductors of Electrical Currents. 4i7 



a thermo-electric current developed at the joint A_, but one deve- 

 loped at the joint C^ the most highly heated portion of the bis- 

 muth conductor AC : for the unequal heating of bismuth con- 

 ductors see Edinb. Journal, vol. xxxvii. p. 300. The direction 

 of this current is ACI, from left to right ; in metallic contact 

 with it there is another circuit AEF, also from left to right ; 

 now where there are two wires for conducting an electrical cur- 

 rent offering different degrees of resistance to its passage, the 

 electricity passes along that which offers the least obstruction ; 

 by the path ACI it would require to pass through the battery, 

 while by the path AEF it has a continuous metallic circuit to 

 pass along through the galvanometer ; by this route it finds the 

 least resistance to return to A, so that in the cross the same 

 action on a galvanometer attached to the wires EG can be 

 obtained from heating at C and developing a current moving in 

 the direction ACI, as by cooling at A and generating a current 

 in the direction AEF, both being from left to right. 



In my experiments published in 1844, I mention that the 

 nearest approach to the production of cold by electricity was 

 attained with a bismuth and antimony couple, the bismuth bar 

 projecting a little beyond the joint for the purpose of immersing 

 it in water. On passing a feeble current across the joint from 

 the antimony to the bismuth, the water had its temperature 

 raised 12°; then, on reversing its direction to make it cross the 

 joint from the bismuth to the antimony, the temperature of the 

 water was only raised 2°. The following repetition of the expe- 

 riment will show a remarkable effect on the temperatures of 

 metallic conductors. I cannot make the experiment go the 

 length of producing cold, but I can show the power of an elec- 

 trical current quickly to reduce the temperature of a heated joint to 

 a point which is stable so long as the current remains unchanged. 



In the annexed dia- 

 gram AB is a bar '>f an- 

 timony, CD a bar of bis- 

 muth soldered to the an- 

 timony bar at B a little 

 within the extremity of 

 the bar at C. At C the 

 bulb of a small thermo- 

 meter CE is applied in 

 contact with a cavity in 

 the point of the bar and 

 enveloped with a little 

 cotton wool, merely suf- 

 ficient to cover the sur- 

 face of the bulb. When this apparatus had been handled, it 



