978 
DR. J. JOLY ON THE SPECIFIC HEATS 
the thermal expansion of the spheres, as indicated, introduced slight differences of 
density into observations over variable ranges of temperature, this effect is so small, 
as may be seen by referring to the tabular statement of experiments, that experiments 
upon the one mass of gas may be considered, without introducing sensible error, as 
being experiments also upon gas of the one constant density. Each of these lines 
is determined by measuring the initial temperature, above 0°, of experiment hori¬ 
zontally ; the condensation in grammes of water (ct)— after all corrections—vertically. 
The temperatures are numbered from the origin 0° to 100°. At 100°, as initial tempe¬ 
rature, there could be no condensation, hence all the lines must pass through this 
point. As the condensations are set off vertically at the initial temperature, of 
each series of experiments, the slope of the line joining this point to the point 100° 
on the axis of temperature affords the mean specific heat over the range to 100°. 
Hence joining all the points so fixed gives a graphic representation of the behaviour 
of the gas at the particular density to which the experiments apply when the lower 
limit of temperature is varying. The dotted lines upon the diagram to Part II. 
give these same experiments, for the most part, but plotted as specific heats against 
absolute density. 
Looking at the lowest couple of lines, p — 0’0456 and p = O'OSOO, we see observa¬ 
tions connected by a straight line—laid down in fact through the experiments by aid 
of a straight-edge, and it follows from this—the slope of the line being uniform 
throughout—that at these densities the variations of specific heat, as the range is 
varied from about 10°-100°, 35°-100°, 56°-100°, and 78°-100°, is so small as to have 
escaped experimental detection or be non-existent. For although the lower of the two 
densities lies at the limit to which observations could be carried with the arrange¬ 
ments used, and are therefore unsteady in some degree, the higher density allowed 
of considerable accuracy in the observations, and the uniformity of the line obtained 
for it, taken in conjunction with the teaching of the lines referring to higher 
densities—which show a curvature increasing wdth the density—appears to render it 
a safe conclusion that at densities below 0’08 the variation of the specific heat with 
temperature, over the limits 10° to 100°, is very small. Not probably non-existent, 
however, as there is no reason to suppose any discontinuity in the physical properties 
of the gas, as its density increases to that of the third line. O'1240. 
This line. O'1240, shows that throughout the specific heat is very slightly variable. 
The sharp upward curvature at the colder end is due, however, to the presence of a 
small quantity of the liquid carbon dioxide not evaporated till a temperature of 8° C. 
was reached, whereas the mean initial temperature was 6°'39 (Table XII.). The 
upper lines more strongly repeat this behaviour. When, also, the actual specific heats 
are plotted against density, as in the diagram to Part II., it is seen that the dotted 
line carrying the experiments over the range 35°-100° lies at the higher densities 
well below that for the ordinary range, 12°-100°, and to possess an upward curvature 
with increasing density. Again, the line for experiments between 57° and 100° lies 
