SIR B. C. BRODIE ON THE ACTION OF ELECTRICITY ON GASES. 
463 
Table (continued). 
I. 
Degree of concentration of the solution of 
hydriodic acid employed. 
II. 
T. 
III. 
S. 
IV. 
-I- 
27- 1 gramme oi’ Iodine in 8 cub. centims 
15-5 
31*4 
2*02 
28. „ 
15-5 
30-9 
2-00 
29- „ „ 
12-5 
25-4 
2*03 
30. 1 gramme of Iodine in 12 cub. centims 
23 
42 
1-82 
31. „ » 
23 
42*4 
1-84 
32. 1 gramme of Iodine in 16 cub. centims 
15 
28-3 
1-88 
33. ,, „ 
15 
28-5 
1-88 
34. 1 gramme of Iodine in 24 cub. centims 
23 
37-7 
1-64 
35. 9, 59 
23 
38-8 
1-69 
36. 1 gramme of Iodine in 48 cub. centims 
20 
32-1 
1-61 
3/. „ )> 
20 
31*3 
1-57 
38. 1 srramme of Iodine in 64 cub. centims 
12-5 
18-9 
1-51 
39. 1 gramme of Iodine in 96 cub. centims 
20 
28-1 
1-40 
40. „ „ 
20 
28*8 
1-44 
41. 1 gramme of Iodine in 128 cub. centims 
12-5 
17-2 
1-40 
42. 1 gramme of Iodine in 192 cub. centims 
20 
22*4 
M2 
43. 1 gramme of Iodine in 256 cub. centims 
12-5 
14-2 
1*14 
It is evident on inspection of the Table that there is an appreciable diminution in the 
value of R in these experiments made with solutions of hydriodic acid containing less 
than one gramme of iodine in 16 cub. centims., and that in the last two experiments 
given, experiments 42 and 43, made with excessively dilute solutions, the oxidation in 
the solution of hydriodic acid closely approximates to the oxidation effected by the same 
gas in a solution of neutral iodide of potassium. This value, however, diminishes very 
gradually — the mean value of It in experiments 34 to 41 being 1-53, and the extreme 
values of It in experiments made the one with a solution of hydriodic acid containing 
1 gramme of iodine in 24 cub. centims. and the other 1 gramme of iodine in 128 cub. 
centims. respectively being 1'66 and 1-40, the difference of which, 0-26, lies almost 
within the errors of experiment. In experiments 1 to 33 no regular variation is seen in 
the value of R, the mean value of this ratio for the first nineteen experiments given in 
the Table being 2-04, and the mean value of the same ratio for the following fourteen 
experiments being 1-92, the mean of the thirty-three experiments taken together 
being T99. 
There are numerous causes of uncertainty in these experiments, several of which have 
already been pointed out; but it is impossible for us to apply any conjectural correction 
to the experiments, or to say how these causes have affected any particular experiment. 
With the view of determining the probable error of the result, I have applied to the 
system of experiments the method of least squares. Some of these experiments exhibit 
very considerable divergences from the mean result, but I have not ventured to reject 
any. A singular example of this divergence is exhibited in experiments 7 and 8, both 
