536 Mr. 0. J. Lodge on a Mechanical Illustration 



than before, yet, since the impacts of one are more violent 

 than those of the other, they will still have some irregularity — 

 those of the hotter metal A going faster from A towards B 

 and slower in the reverse direction, while those of B go slowly 

 from B to A and taster back again from A to B. On the 

 whole, then, we have a double layer of buttons oscillating 

 quicker in one direction, viz. A B, than the other*; and snch 

 buttons, if threaded on a cord, would have a tendency to drive 

 it in the direction A B, as we have already seen, § 28. 



§ 30. At a junction of two metals, or of unequally hot 

 pieces of one metal, we appear then to have found an electro- 

 motive force acting in one direction ; but there is something 

 further to be noticed yet. We found in the former paper 

 that in any complete model it was necessary to have not one 

 cord but two, arranged side by side and connected so that 

 however one moves the other must move with the same velo- 

 city in the opposite direction — one representing a current of 

 positive, the other the simultaneous equal opposite current of 

 negative electricity. Now the electromotive force whose exist- 

 ence we have been recently imagining, is one that will act on 

 both these cords ; and if it acted on both equally they would be 

 both urged equally in the same direction, and consequently 

 could not move at all. It is necessary, therefore, to assume 

 that the bite of the buttons on one of these cords is greater 

 than their bite on the other ; but whether it is the positive or 

 the negative cord on which the greatest force is exerted in a 

 given case will depend on the nature of the metal, and possibly 

 also on its temperature — though we have no evidence as re- 

 gards the latter point, and I shall not assume it. 



What w T e have now assumed amounts to this, that the 

 quantity r, called the specific resistance of a substance, is made 



* Although we should have at such a junction molecules travelling 

 quicker in one direction than the other, there "will be no excess of 

 momentum either way as was to he expected in dielectrics and elec- 

 trolytes (see § 8 footnote and § 16) ; for there will be a large number 

 of molecules side by side in such a layer, and, since each takes a longer 

 time over its slow journey than over its quick one, there will be at any 

 instant more molecules moving slowly than quickly, and their excess in 

 number will make up for their defect in velocity, so that the average 

 momentum each way will be the same. But it is a suggestion of Mr. G. 

 Johnstone Stoney's, which he made at the Physical Society last Novem- 

 ber, in connexion with his theory of Mr. Crookes's radiometer, that 

 though mv may equal m'v' it by no means follows that lmv 2 =^?7i'v' 2 : 

 there will, in fact, in the above arrangement be an excess of kinetic 

 energy in favour of the quickly moving molecules, *". e. from A to B. This 

 is energy which can be expended in driving the cord. But if the cord 

 be forcibly held so that it cannot move, the energy will still, I suppose, 

 be transferred from A to B ; but it will not now be doing useful work, it 

 will be degraded into mere diffusion of heat. 



