414 
DR. F. P. BURT AND DR. E. C. EDGAR ON 
Hydrogen was prepared and purified as in series 3 : oxygen was obtained both by 
electrolysis and by heating permanganate, so that large quantities of gas were 
rapidly available for fractionation. Electrolytic oxygen, accumulated from time to 
time in a 2|--litre storage bulb, entered the main train at the T-piece (Z, fig. 3), after 
being dried over calcium chloride, and phosphoric oxide. 
In this series, and in the preceding one, an attempt was made to measure the 
temperature of the mercury in the short exposed portion of the manometer column 
that projected below the bath (K to L, fig. 4). For this purpose, a small thermometer, 
graduated in tenths of a degree, was fastened by fine copper wire to the manometer 
tube, and the whole was wrapped in cotton-wool. The temperature registered by 
this thermometer was probably proportional to, if not identical with, the temperature 
of the mercury in this region of the manometer. The readings varied with the room 
temperature and length of time during which the system had been in ice. The main 
controlling factor was the temperature of the room. Since the distance between the 
two glass points to which the mercury meniscuses were set 
was constant, any rise in temperature of the mercury in the 
manometer implied a diminution of pressure and corresponding 
increase in volume of the gas in the pipette. An empirical 
value for the temperature effect was obtained by filling the 
pipette with gas, making a setting and noting the tempera¬ 
ture of the thermometer, and then repeating the measurement 
when the temperature was a degree or so higher. From the 
weight of the few drops of mercury that had to be withdrawn 
from the volume-adjuster to set the points on the second 
occasion, it was possible to estimate the increase in the volume 
occupied by the gas per unit rise of temperature as registered 
by the thermometer. 
The mean of several very concordant results was 21 c.mm. 
per degree. Now the length of the exposed portion of the manometer was about 
4 cm., and a change of 1° C. in the temperature of this length of mercury could not 
produce so large an effect. In fig. 6 the two columns of mercury, AA and BB, only 
equalise each other when their temperature is the same. All this region was packed 
in cotton-wool, the tubing below the dead-space being the more thickly coated. The 
mercury in A would be more sensitive to temperature changes than the larger bulk 
of better insulated mercury in B, so that part of the measured effect may have been 
due to relative temperature change in this part of the system. 
In any case, the effect on the results is very small, since the room temperature 
seldom varied half a degree during a determination. The application of the correction 
to the results of series 3 improves the concordance a little without altering the 
mean. 
In series 4 the temperature variations were greater. In order to protect the 
Fig 6 
