PHYSICS: F. G. KEYES 
327 
This fact was observed and discussed by Ramsay and Young for ether 
and carbon dioxide. Amagat thought, from an examination of his own 
data, that the 'law' was nearly exact in the case of many substances, 
certainly not true, however, in the case of the alcohols and water. Un- 
fortunately it is necessary to have very accurate pressure measure- 
ments, to test the rule and moreover the measurements have always 
been made along isotherms instead of along isobars. This makes it 
necessary to obtain the constant volume data by graphical means with 
the Hability of introducing a trend into the experimental data. It is 
comparatively easy to measure temperatures and pressures, but volume 
measurements under pressure are extremely difficult. Every substance 
which may be used to serve as a container for the experimental substance 
expands with temperature and pressure increase and the expansion 
may be moreover a compHcated function of the time. 
There can scarcely be any doubt that the high pressure gauge used 
by Amagat for the measurements above 1000 atm. was more in error 
than his low pressure gauge. . The values of pv for nitrogen were plotted 
accordingly for the several temperatures from 0° to 200°, with v as one 
coordinate in the region up to 1000 atm. Values of pv were read off at 
the different temperatures for various values of the volume thus yield- 
ing material for calculating the constant volume, pressures, and tem- 
peratures. These values were laid off on coordinate paper and the best 
straight line passed through the data. The equations of the several 
straight lines was corrected subsequently by calculating the deviations 
and constructing an additional graph by means of which a deviation 
equation could be evaluated and also any trend in the deviations detected. 
The equations obtained were all of the form, p = [^T — and by 
comparison with equation (1) ^ = R/{v — b) and $ = a/{v — iy whence 
d = V — R/^. The 6 values may be treated best by using log 3 and 1/v 
as the coordinates of the graph since it is propsed to test the functional 
form 8 = The <J> values were tested by writing v = ^| a/ ^ ^ -\- 1. 
It is seen that v should be plotted with 1/ as abscissae. The graph 
for nitrogen is given in figure one and for argon in figure two. It should 
be stated that the data given by Amagat above 1000 atm. give values 
of log b which join on to the values in figure two, but begin to differ 
slightly in trend as the pressure becomes greater. In fact a straight 
line may be passed through the 1-1000 atm. series and the 1000-3000 
atm. series with the lines intersecting at about 1000 atm. The effect 
is due undoubtedly to the inaccuracy of the high pressure gauge. 
The graphs show the functional form of b as related to volume changes 
is very accurately satisfied, while argon shows no trend. In the graph 
