1888.] the Thermo-electric Properties of Bodies. 



225 



and by the chemical action of the oxygen ; both of these will, however, 

 produce permanent changes. That only part of the action is to be 

 explained in this way is shown by the previous experiments. It is, 

 however, further demonstrated by using two carbon rods in decom- 

 posing acidulated water ; after passing the current for one minute, 

 reverse it for one-tenth of a second and immediately join up to a 

 galvanometer. A short but violent deflection appears for the latter 

 contact, gradually falling to zero and passing to the other side, where it 

 remains for a considerable time, though with much decreased quantity. 



The same thing was obtained with platinum electrodes. The 

 second contact must be very short, or the former polarisation 

 disappears. I have not yet succeeded in obtaining more than one 

 reversal, although I have no doubt that more may be got with very 

 thick electrodes. 



Resistance. — Copper and iron absorb hydrogen and silver occludes 

 oxygen, but no change in their thermo-electric properties could be 

 detected. Carbonic oxide is absorbed by iron, and is said to produce 

 great changes in its properties. In this case, however, only the 

 resistance was measured. 



A piece of iron wire, about 3 yards in length, was twisted into 

 a spiral and placed in a porcelain tube; the ends projecting about 

 3 inches, were connected with one side .of a bridge and balanced 

 against an equal spiral of the same wire. After exhausting the 

 tube about 1 foot of the central portion was heated to a bright red- 

 ness ?nd then allowed to cool. Next day the resistance was measured, 

 and the experiment repeated twice. On the third heating, carbonic 

 oxide was allowed to enter the porcelain tube, and readings of the 

 resistance taken on cooling as before. This was also repeated. 



This series was again repeated with new wires, and lastly, the wire 

 was raised to a bright red in vacuo and allowed to cool, the object 

 being to remove the carbonic oxide gas in order that another measure- 

 ment might be taken after these repeated heatings. The resistance 

 fell, clearly proving that part of the previous increase was due to the 

 presence of the gas. No measurement of resistance was taken on 

 the same day that the wires were heated, but at least 15 hours were 

 allowed to elapse. 



First series of observations give the numbers thus : — 



Average of three measurements after heating in vacuo, 04 ohm. 

 )? jj a >j in car- 



bonic oxide, 0"41 „ 



With the new wire — 



Average of three measurements after heating in vacuo, 0'6S „ 

 » » » » in car- 



bonic oxide, 0'655 „ 

 After heating in vacuo to expel the gas, it fell to 0'642 „ 



