Disappearance of Gas in the Electric Discharge. 239 

 appreciably, and we may put p-2 = 0. Then 



-A-oH-A): • • • • (5) 



and if D is small compared with ft we shall have 



A=iQ; (6) 



that is to say, we shall find for a half its normal value, which 

 is the value actually found. 



Bat the assumptions need justification. Some measure- 

 ments o£ ft were attempted by filling the vessel with C0 2 , 

 burning the filament without a discharge, and observing the 

 rate of production of CO. Quantitative results were not 

 obtained, but it was shown that the reduction is exceedingly 

 rapid and that the reverse reaction might probably be regarded 

 as always in equilibrium. On the other hand, the assumption 

 that D is small compared with ft is not permissible, for much 

 of the C0 2 undoubtedly accumulates in the cooled trap. 

 Indeed, the whole research was suggested by the observa- 

 tion in the experiments of the first paper (confirmed in 

 these) that the C0 2 found in the trap was the whole of that 

 to be expected from the reaction 2CO = C0 2 + C. If any 

 appreciable part of the C0 2 were reduced by the hot tung- 

 sten, the oxygen of that part would be lost permanently and 

 the C0 2 in the trap would be less than half the CO. This 

 observation indicates that D is great compared with ft ; and, 

 if it is, the normal value of a, namely 2, ought to be obtained. 



Nevertheless, a shortening of the tube connecting the 

 vessel to the trap produced a notable increase in the value of 

 & obtained : the low values must, therefore, be due to an 

 insufficiently rapid removal of the C0 2 formed. The only 

 explanation appears to be that the conversion of CO into 

 C0 2 is reversed by the discharge itself, and that this reversal 

 is a complete reversal, bringing the carbon again into the 

 reaction and producing from one volume of C0 2 two volumes 

 of CO. Confirmation was obtained by filling the vessel with 

 CO and passing the discharge till almost all the gas was 

 absorbed ; then removing the liquid air (mercury was ex- 

 cluded by a second trap separated by a tap) and passing the 

 discharge through the C0 2 evolved from the warmed trap. 

 By this process nearly the whole of the original CO could he 

 regained. It appears, therefore, that the discharge not only 

 reverses the reaction 2C0 = C0 2 + 0. but also reverses (or 

 prevents) the reduction of C0 2 by hot tungsten. The value 



