410 
DR. F. P. BURT AND DR. E. C. EDGAR ON 
accuracy of one of the settings was suspected at the time on account of the light 
being had. 
The five columns of figures in Table II. correspond to those in Table I. 
Table II. 
1 . 
Hydrogen 
excess. 
2. 
Hydrogen 
residue. 
3. 
Difference. 
4. 
Ratio of 
combining 
volumes. 
5. 
Atomic weight 
of 
hydrogen. 
c.c. 
c.c. 
c.c. 
1 
2-026 
1-147 
0-879 
2-00286 
1-00772 
2 
3-331 
2-440 
0-891 
2-00290 
1-00774 
3 
3-687 
2-794 
0-893 
2-00291 
1-00774 
4 
2-233 
1 • 359 
0-874 
2-00285 
1-00771 
6 
5-641 
4-777 
0-864 
2-00281 
1-00769 
7 
3-871 
2-960 
0-911 
2-00297 
1-00777 
8 
4-738 
3-822 
0-916 
2-00298 
1-00778 
9 
3-926 
3-000 
0-926 
2-00302 
1-00779 
10 
2-835 
1-934 
0-901 
2-00293 
1-00775 
12 
2-958 
2-062 
0-896 
2-00292 
1-00774 
Mean .... 
2-00292 
1-00774 
±0-000014 
1 
Maximum deviation, 1 in 
10,000 
Before the nest series of determinations the palladium tube, which was showing 
some signs of weakness, was cut down and sent away to be repaired. 
Third Series. 
Hydrogen was made and purified as in series 2, except that on leaving the 
electrolysis vessel it passed through a tube containing palladium black. Oxygen was 
obtained from potassium permanganate as described on pp. 399—401. 
At this stage of the research we had a great deal of trouble owing to the appear¬ 
ance of condensible gas in the hydrogen left after explosion, accompanied by 
discrepant and much lower results. The quantity of gas condensed in the spiral, >/, 
varied Irorn 40 c.mm. to nearly 300 c.mm. in different experiments. It was completely 
soluble in a drop of potash or baryta solution, giving a white precipitate in the latter 
case. The source of the impurity was at first believed to be the oxygen generated 
from permanganate, since all other conditions were apparently the same as in 
series 2, but liquefaction and fractionation of much larger quantities of oxygen 
produced no improvement. 
