THE COMBINING VOLUMES OF HYDROGEN AND OXYGEN. 
417 
combine the gases on the first day and leave the wet hydrogen residue safely 
imprisoned between mercury surfaces in the explosion vessel over night. The 
residual hydrogen was measured on the following day. The order of operations at 
the beginning of each day was identical. First of all the pipette was cooled with ice ; 
then hydrogen was admitted, the upper bath iced and a setting made. During the 
rest of the day the ice in the pipette bath was never removed, but was pressed down 
and replenished immediately before each reading. On the second day, as soon as the 
first hydrogen setting was accomplished, the residue in the explosion vessel was dried, 
sparked and transferred to the pipette ; the ice in the baths was then replenished and 
the final setting was made. The last two measurements, therefore, were separated 
by a considerable interval of time. Of the five gas measurements required in each 
determination, only two, namely the first of each day, fell within the three-hour 
period preceding the establishment of equilibrium. Consequently, in both these 
measurements the quantity of hydrogen actually taken was less than the quantity 
assumed in the calculation. The error in each case makes the result too high. 
The accurate evaluation of the correction involves a knowledge of the time 
interval between the introduction of ice and the completion of the setting. Although 
the time had always been noted when a setting was made, the same precaution 
had, unfortunately, seldom been taken at the moment of iceing the bulb. There 
was little variation, however, in the time occupied by the intermediate operations, 
which in series 3 and 4 had become almost mechanical. In series 3, the most uniform 
as regards method of procedure, and the most concordant in results, the time interval 
as actually measured in several determinations was one and three-quarter hours, and 
the mean time for the whole series was certainly within a few minutes of this. 
It may be seen from the curve (fig. 7) that when the bulb had been in ice for an hour 
and three-quarters, the volume of the gas was still 18 c.mm. in excess of its final value, 
so that had the measurement been delayed till contraction was complete, an additional 
18 c.mm. of gas must have been introduced to achieve the setting. Since the calculation 
is based on this assumption, the nominal volume of hydrogen taken for synthesis 
must be diminished by 18 c.mm., and, by parity of reasoning, the nominal volume of the 
residue must be increased by the same amount. The net result is a diminution of 
36 c.mm. in the volume of hydrogen required to combine with 307 c.c. of oxygen. This 
implies a reduction of 6 parts in 100,000 in the values for the ratio and the atomic 
weight. 
In series 4, an attempt to estimate the above time intervals for each individual 
experiment led to the same ayerage reduction. In experiment 6, for example, the 
oxygen measurement, instead of being made at the end of the first day when the bulb 
had been in ice for many hours, was made on the following morning. The tempe¬ 
ratures of the oxygen and hydrogen were therefore more comparable, and the resulting 
correction was small. In several of the other determinations the correction was larger 
than in series 3. As regards series 1 and 2 a similar correction was necessary. The 
3 L 
VOL. CCXVI.—A. 
