78 
PROFESSOR W. RAI^ISAY AND DR. M. W. TRAVERS 
confined a known volume of gas. The volume-tube wms transferred to the pressure 
api^aratus, and measurements were made. 
The high-pressure manometer had previously been compared with the low-pressure 
manometer, and the latter, with an open tube manometer. The volume-tube and the 
high-pressure gauge were jacketed in the first instance with water at atmospheric 
temperature, w’hich, of course, was accurately known ; and the heights of mercury 
in the air-gauge and in the volume-tube were read, both during ascending and during 
descending pressure. The results were quite concordant. The readings of the air- 
naucre wnre corrected for deviations from Boyle’s law, accordino’ to Amagat’s 
results for air, in comparison with directly read pressures; and the readings of the 
volume-tube wnre translated into real volumes, and, by assuming Gay-Lussac’s law, 
corrected to what they would have been at 11’2° C., which happened to be the 
tem 2 :)erature at which readings were taken with the gas first investigated, helium. 
As the highest temperature recorded did not exceed 15°, it was assumed that no 
actual variation from Gay-Lussac’s law would influence the results, within such a 
small interval of temperature as that between 15° and 11'2°. 
For a higher temperature, the boiling-point of quinoline under atmospheric 
pressure, about 237’3°, was chosen. The pressure of the atmosphere was not 
always quite normal; but the barometer was always read, and the necessaiy 
change in volume was calculated according to Gay-Lussac’s law for the small 
interval of temperature required—only 0T°. 
The pressures, volumes, and their j^roducts are given in the accompanying tables. 
The volumes are stated in terms of cubic centimetres occupied by the molecular 
weight of the gas taken in grammes at the pressures and temperatures specified. 
They are thus all comparable with each other, and with the corresponding constants 
for 28 grammes of atmospheric nitrogen, as measured ly Amagat with a direct-reading 
manometer. The basis of calculation has been taken as the volume occupied by 
32 grammes of oxygen. Data on this constant vary slightly among themselves. The 
mean of Kegnault’s, Bayleigh’s, Jolly’s and Leduc's determinations for the 
weight of a litre of oxygen is 1'42961 gramme; that of Bayleigh is 1 ■42952 ; and 
» 
of Leuuc, 1'42920. Taking Bayleigh’s number as occupying an intermediate 
position between the other two, 32 grammes of oxygen would occupy at 0°'and 760 
millims. 22,395 cub. centims. ; multiplying b}^ 0'760, the normal pressure as a 
fraction of a metre, the value of B.Y. is 17,012 metre-cuhic-centims. At 11 "2°, 
from Gay-Lussac’s law, this value is increased to 17,710; and at 237'3°, the other 
temperature under consideration, to 31,800. On the hy])othesis that the products 
of pressure and volume for a perfect gas remain constant, these values should 
represent B.V. at all pressures and volumes. We shall now see how fiir this 
condition is fulfilled by the gases in question. For the sake of convenient com- 
jiarison, Amagat’s results for atnios})heric nitrogen have lieen inserted, both in the 
tables, and in the diagram (Plate 2) representing the results. 
