PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 699 
to explain them away by attributing them to coatings of oxide formed on the metals, 
or to other causes supposed with equally little reason to exercise sensible disturbing 
influences ; but the descriptions, given by the original observers, of their experiments 
leave no room for such doubts. It is certainly not easy to get the inversion between 
copper and iron (with such specimens as I have tried) by the heat of a spirit-lamp, 
applied as described by Becquerel to one junction while the other is left cool ; but 
I readily obtained it by raising the other junction somewhat in temperature, with the 
first still kept at a red heat. Probably if the atmospheric temperature had been 
higher, or if a somewhat more intense red heat had been obtained from the spirit- 
lamp, I should at once have obtained the result simply in the manner described by 
the previous observers. 
81. The easiest way to verify the thermo-electric inversion of iron and copper is to 
take a piece of iron wire a foot or two long, and twist firmly round its ends two 
copper wires connected with the electrodes of any ordinary astatic needle-galvano- 
meter. Then first heat one of the junctions with the hand, or by holding it at some 
height over a flame ; and note the deflection, which will be found such as to indicate 
a current from copper to iron through the hot junction. Again, heat both junctions 
in flame, or in sand at any temperature above 300° Cent., and withdraw one a little 
from the hottest place, so that while both junctions are at temperatures above 300° 
Cent., that which was heated in the first experiment may be still decidedly hotter than 
the other. The deflection will now be found to be the reverse of what it was before, 
and will be such as to indicate a current from iron to copper through hot. The 
reversal of the current may be very strikingly exhibited by allowing the two junctions 
gradually to cool, while ensuring that the same one remains always somewhat above 
the other in temperature. When the mean of the temperatures of the two junctions 
falls below 280° Cent, or thereabouts, the primitive deflection will be again observed. 
All these phenomena are observed indifferently whether the copper wires be simply 
twisted on the ends of the iron wire, or brazed to them, or tied to them by thin 
platinum or iron wire. 
82. Similar phenomena may be observed without the necessity of going to so high 
temperatures, by soldering galvanometer electrodes of copper to the ends of a double 
platinum and iron wire, and treating this compound circuit in the pig. 12 . 
manner just described, only with a more moderate application of 
heat. If the platinum wire be very thin in comparison with the 
iron one connected along with it, the circumstances will be but 
little altered from those observed when iron simply is used. By 
taking a thicker platinum wire, or several thin ones together, in 
connexion with the same iron wire, or by using a thinner iron wire 
and the same platinum, the neutralization and reversal may be 
shown with temperatures below the boiling-point. Most specimens of platinum wire 
thus applied reduce the neutral point of copper and the compound platinum and iron 
MDCCCLVI. 4 z 
