OF ARTS AND SCIENCES. 45 



bar. Tlionison attributed tliis to the fact that in some metals the cur- 

 rent evolved heat in passing from hot to cold parts ; in others it 

 absorbed heat. And the difference between the heat absorbed in one 

 metal, and that evolved in the other, he supposed to be one cause of the 

 thermo-electric current. 



Another explanation is that the current of electricity changes the 

 thermal conductivity of the bar, in some metals mcreasing the con- 

 ductivity, and diminishing it in others. 



It is well known that the electric current modifies the physical 

 properties of metals along which it passes. It changes their cohesion, 

 in some cases increasing it, in others diminishing it. The elasticity 

 of metals is also changed under the influence of electricity. What 

 is more probable than that the thermal conductivity should also be 

 changed ? 



There are many objections to the view taken by Thomson. The 

 numbers expressing the Thomson effect bear no apparent relation to 

 the thermo-electric current ; and, moreover, the effect is entirely too 

 small to produce even the weakest thermo-electric current. 



The fact that the formulte deduced on Thomson's hypothesis agree 

 with experimental results is of but little importance. Tlie thermo-elec- 

 tric curves determined experimentally are, approximately, parabolas- 

 Thomson's equation is also that of a parabola. But any theory, based 

 on the sujDposition that the heat effects are proportional to the current 

 strength, will give the equation of a parabola. 



It was mentioned that the current (9 -|- /S) (7 was imperceptible. 

 In the few expei'iments that were made the current G was very feeble, 

 not much greater than the thermo-electric current I, which was given 

 by an element of German silver alid iron. The resistance of the cir- 

 cuit was about two hundred ohms. Consequently, the current equiva- 

 lent to the extremely small amount of heat (9 -j- aS) C might readily 

 have produced no apparent effect. 



I have attempted to show that the Peltier and Thomson effects can- 

 not be the whole cause of the thermo-electric current. The true cause 

 is yet to be discovered. 



Many unsuccessful attempts have been made to find relations between 

 the strength of the thermo-electric current and the physical properties 

 of the metals of the thermo element. To completely solve the 

 problem, however, we must know the way in which the physical 



