470 
SIR B. C. BRODIE OX THE ACTION OF ELECTRICITY ON GASES. 
error; but there is no doubt that the mean value of — — i in these experiments is 
somewhat too high. I may especially notice two sources of error as constantly affecting 
these experiments in this direction for which I have not been able to apply a correction. 
When pure oxygen is passed through a solution of hyposulphite of soda, either alkaline 
or neutral, a minute but nevertheless real amount of oxidation occurs : this oxidation is 
far too minute to be estimated by measurement in the aspirator ; but I endeavoured to 
gain some notion of its magnitude in the following way. A small pipette of oxygen, being 
87 cub. centims. at 0°C. and 760 millims., was passed at a temperature of 14° C. in two 
experiments through an alkaline and a neutral solution of hyposulphite of soda respec- 
tively ; the hyposulphite of soda was “titred” with a standard iodine solution before and 
after the experiment. In the case of the alkaline solution, the difference between the 
two titrations amounted to 0-7, and in the case of the neutral solution to 0*8 division 
of the burette ; this would correspond in the previous experiments, in which about three 
times the amount of oxygen was passed through the solution, to an oxidation represented 
in the two experiments respectively by 2ff and 2 - 4 divisions of the burette, representing 
an absorption of about Off cub. centim. of oxygen by the alkaline solution of hyposulphite. 
The error from this source, although very small, affects every experiment in the same 
direction, and would appreciably affect the mean result — the value of the mean result, on 
the hypothesis that the contraction Y — Yj is in every case Off cub. centim. in excess, 
being, as I have ascertained by calculation, 1*02. I have not, however, applied this 
correction, as there is no positive evidence that the oxygen, as thus calculated from the 
difference of the two titrations, represents the oxygen actually employed in the oxidation 
of the hyposulphite. The point also is immaterial. Another source of error is found 
in the gas simply dissolved by the solution of hyposulphite, which would not be estimated 
by this method of titration. Some idea of the magnitude of the error from this source 
may be derived from the experiments given in Section II. (page 446) with neutral iodide 
of potassium, where the volumes before and after the experiment (much the same quan- 
tity of gas being employed as in the present experiments), which theoretically should 
be identical, show a difference in their mean value of 0 - 2 cub. centim., for which dif- 
ference we have an adequate cause in the absorption of 0 - 2 cub. centim. of oxygen by 
the neutral solution of iodide of potassium. Here, again, I have hesitated to apply a 
correction, which would be to a certain extent of an arbitrary character ; but this is un- 
doubtedly a real and constant cause of error operating in the same direction as the pre- 
ceding ; and if from these data we assume the contraction Y — V 1 to be in all cases in 
excess by 03 cub. centim., the mean value of It would, after applying this correction, be 
as nearly as possible TOO. The probable error of the result would not be appreciably 
affected. 
In the experiments of which the result is given in the following Table the electrized 
gas was passed through a solution of polysulphide of barium, made by boiling a solution 
of the neutral sulphide with an excess of sulphur out of contact of air. The passage of 
