PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 657 
wire, with a sudden reversal of the current. Sometimes the incandescence was 
assisted by a spirit-lamp flame applied to the middle part of the wire, and the ends 
were kept cool by wet threads. Sometimes in a long wire with a current through it 
not quite strong enough to keep it at a red heat, a small part was made incan- 
descent by a slight application of heat as nearly as possible at one point, by a spirit- 
lamp flame. Still there was never observed the slightest motion of the incandescent 
part, when the current was suddenly reversed, and I concluded that whatever had 
been observed in the way of different heating effects of the positive and negative 
electrodes, must have been owing to peculiar agencies of the current in passing 
between metal and rarefied air, or to some other cause than thermal convection in 
metals ; and I saw that more powerful tests would be required to bring out the result 
I looked for. 
20. I next made experiments on a conductor of bar iron bent into two equal 
upward vertical branches on each side of the horizontal part, which was kept im- 
mersed in a vessel of hot oil, while the upper ends of the vertical branches were kept 
cool by streams of cold water. Vessels of water were applied round the two vertical 
branches, as calorimetric arrangements to test heat evolved or absorbed in them by 
the agency of a current sent down one and up the other from a nitric acid battery of 
sixteen small iron cells, arranged as a single element. 
The current was sent first for half an hour in one direction, then half an hour in 
the contrary direction ; and so on, with a reversal every half-hour. The water round 
the two vertical branches was kept constantly stirred, and thermometers in fixed 
positions in them were observed at frequent intervals during the experiments, which 
were each continued for about two hours. A comparison of all the readings taken 
showed a rather higher mean temperature in the branch down which the current was 
passing than in the other ; indicating, differentially, a cooling effect in the branch 
through which the current passes from the hot middle, and a heating effect in the 
other. This experiment appeared to show that ‘^the resinous electricity” carries 
heat with it in an iron conductor; but the irregular variations of temperature in each 
thermometer were so much greater than the differential effect deduced, that I could 
not consider the conclusion satisfactorily established. 
^§21 to 29. Unsuccessful attempts with large bar conductors. 
21. There were diflSculties connected with the arrangements of the calorimetric 
vessel, which made me judge that it would be better, instead of testing the average 
temperature of two portions of the conductor, each extending the whole way from 
the hot middle to the cold ends, to simply test the temperature of as nearly as pos- 
sible one point midway between the hot and cold on each side ; and it appeared that 
the heating could be more easily applied and better regulated by a source of heat at 
the middle of a straight horizontal conductor, than by the plan I had followed in the 
arrangement just described. I therefore got bars of copper and iron, with holes to 
admit the bulbs of sensitive thermometers, made to the following dimensions : — 
