450 
SIE B. C. BEODIE ON THE ACTION OF ELECTEICITY ON GASES. 
are absorbed by that solution. The value of the hyposulphite employed for the titration 
was estimated with the greatest care ; and the difference of these two numbers certainly 
does not depend on any error in the mode of conducting the experiment. Moreover 
every chemist who has experimented upon this subject has observed a difference in the 
same direction. The question is as to the cause of this difference. 
It appears, on the face of the results given in the preceding Table, that this difference 
is proportional, or nearly proportional, to the “ iodine-titre,” which, again, is itself pro- 
portional to the quantity of oxygen employed in the experiment. 
It was ascertained by Andrews*, and the point has been amply confirmed by Meissner, 
that when the electrized oxygen is passed through an acid solution of iodide of potas- 
sium this discrepancy no longer exists, but the increment of weight agrees with the 
“ titre.” The same is true when the gas is passed through the solution of neutral iodide 
until the passage of the ozone is no longer arrested by it, in which case the whole of 
the iodine is converted into the form of iodic acid f. Two causes may be indicated, 
both of which tend to create such a discrepancy: — ("1) the solution of oxygen in the 
alkaline solution of iodide of potassium; (2) the formation of oxides of nitrogen in 
the induction-tube from traces of atmospheric air mixed with the oxygen employed in 
the experiment. I have ascertained by direct experiment that when atmospheric air, 
carefully dried, is submitted in the induction-tube to the electric action, very considerable 
quantities of the oxides of nitrogen are formed. Thus in one experiment in which 2900 
cub. centims. of air were thus operated upon, and then passed through a bulb containing 
a solution of caustic potash, 0'0086 gramme of nitrogen was after the experiment 
detected in the alkaline solution in the form of nitric and nitrous acids, the nitrogen 
being carefully estimated in the form of ammonia. Now, assuming the iodine-titre to 
have amounted in the previous experiments to 4 per cent, of the total gas, in the first 
experiment (which exhibits the greatest divergence) 1257 cub. centims. of oxygen must 
have been operated on in the induction-tube. If the gas contained only 1 vol. in 1000 of 
nitrogen, 1000 cub. centims. would contain 0'00125 gramme of nitrogen, corresponding 
to *0034 gramme of the teroxide of nitrogen, which is more than the whole difference. 
Andrews attributes the discrepancy referred to to the presence of carbonic acid in 
oxygen formed by electrolysis J ; but, as has been clearly shown by Baumert§, the 
occurrence of carbonic acid was simply a peculiarity in the experiments of Andrews, and, 
moreover, the gas employed by Meissner was derived from chlorate of potash, and 
specially freed from carbonic acid || . In the case of gas collected, as in the present 
experiments, over large quantities of sulphuric acid, such traces as might be present, 
either of the oxides of nitrogen or of carbonic acid, would doubtless be removed by 
solution in that liquid. 
There is still something here to be explained; but I have not pursued the subject 
* Philosophical Transactions, 1856, vol. cxlvi. p. 8. f Meissner, p. 78, Table B. 
7 Philosophical Transactions, 1856, vol. cxlvi. p. 4. § Annalen der Physik, vol. xcix. p. 91. 
|| Meissner, p. 17. 
