THE COMBINING VOLUMES OF HYDROGEN AND OXYGEN. 
425 
of the paper. They may be tabulated here conveniently under five heads, as errors 
due to :— 
(1) Length of time required for the gas in the pipette to cool to 0° C. 
(2) Variation in the temperature of the iced system on different occasions. 
(3) Rise in temperature in parts of the iced system before completion of setting. 
(4) Variation in temperature of exposed portion of manometer column. 
(5) Relative temperature change in the mercury in the manometer and dead-space 
below the level of the point. 
Comment on the nature of the error under each head may be made as follows :—- 
(1) Systematic error, which was estimated in the first four series and eliminated in 
the fifth. 
(2) Probably very small errors : the point was frequently tested by repeating a 
measurement on the following day, when results agreeing to within a few cubic 
millimetres could be obtained, provided that the pipette had been in ice for a sufficient 
length of time. 
(3) Negligible errors : experiments prove that the system remained at 0° C. for a 
much longer period than was actually required for the setting. 
( 4 ) Small hap-hazard errors, estimated and corrected for in series 3 and 4 and 
eliminated in series 5. 
(5) Small hap-hazard errors, except perhaps in series 5. 
A very large proportion of the time occupied by this research was spent in investi¬ 
gating errors of temperatures and, although we have no reason to regard our final 
result as seriously affected by such errors, we are inclined to believe that if the whole 
of the measuring apparatus down to the point where the manometer joins the tubing 
below the dead-space had been in direct contact with melting ice, a better concordance 
would have been obtained. 
One further point suggested itself as worth examination. With the measuring 
apparatus used only a small variation was possible in the volume of hydrogen taken 
for synthesis. In the first three series, where this variation was greatest, there is no 
indication of any connection between the quantity of hydrogen taken and the result. 
Of the thirty-two experiments considered, the mean of those sixteen in which the 
hydrogen excess was greatest agrees to 1 part in 100,000 with the mean of the other 
sixteen. 
The Atomic Weight of Hydrogen. 
If the ratio of the combining volumes is 2'00288 the atomic weight of hydrogen is 
1‘00772, when Mokley’s values for the densities, namely, 0'089873 and 1'42900, are 
taken *- 
VOL. ccxvi.—A. 
3 M 
