OF GASES AT CONSTANT VOLUME. 
955 
(3 and 5 of Part 1.) are completely inappreciable. The final correction therefore to 
be applied to each experiment is got by subtraction of the minus corrections I. and II., 
and addition of the plus correction YII. Each of the tables of experiments contains 
this amount at the foot, it being understood that the mean of each series of 
experiments is treated in all cases as the experiment to be corrected. 
I have not tabulated the figures used to derive the density of the gas, as they can 
be verified comparatively easily from the table of corrections already given. Thus 
the mean density p of table XIII. is the quotient obtained when the total weight of 
gas in the sphere is divided by the mean volume of the sphere. The first quantity 
is got by adding to W, recorded at the head of each table of experiments, the weight 
of a volume of carbon dioxide equal to the initial volume of the sphere and at atmo¬ 
spheric pressure. For evidently this last quantity of gas, while it does not enter into 
the estimate of the mass producing the observed calorimetric eftect—as it remains in 
the sphei'e during the comparison of the inactive vessels—yet must be considered 
in estimating the actual density of the gas. For the foregoing series, I. to X., a 
mass = 0T587 gramme of gas is added in each case to W ; calculated on a volume ot' 
86 T2 cub. centims., a temperature of 16°'7 C., and a pressure of 760 millims., the 
approximate volume, temperature, a)id pressure obtaining. 
The mean volume is obtained from the table for corrections (XIY.) by adding to the 
volume at and Pj the increase of volume due to the rise of pressure ■§ (P., — P^), 
and also the increase of volume due to the rise of temperature I {to — q). 
Finally, the mean pressure obtaining during experiment is evidently that exerted 
by the total mass of gas confined in the volume obtaining at mean tenqjerature 
augmented by the elastic distension due to the mean pressure sought. As, for 
calculation of errors, it is necessary to estimate the initial and final pressures clue to W 
(not to the total mass), it is convenient and sufficiently accurate to add to the mean 
pressure due to W, the pressure due to the mass added, as above, in ascerraining the 
total mass. We may calculate this pressure on Andiiews’ coefficient 0'0037 for 
change of pressure at constant volume between 20° and 100°. The pressure so 
estimated is found to be 1T5 atmosphere. I may observe, however, that I departed 
only so far from accuracy as to add the one atmosphere, as the degree of accuracy 
attained in estimating P^ and Pg, and the mean pressure due to W, did not warrant 
addition of small quantities. 
The higher pressures were ascertained from Amagat’s recently published tables of 
the isothermals of carbon dioxide (‘Annales de Chimie et de Physic^ue,’ 6th series, 
vol. 29). A chart of densities (as abscissse) and pressures (as ordinates) was con¬ 
structed. The mass afibrding the Y in Amagat’s tables, is that of unit volume of 
carbon dioxide at 0° and 760 i.e., 0'0019767 gramme. Hence 
„ F X 0 0019767 
(PV) 
6 F 2 
