( 943 ) 
We have not yet used this apparatus, but as we have already 
mentioned we made use of the less direct method of determining 
Wig ANd Ora, Separatedly for the same temperature, sometimes adding 
determinations of the difference 07, — Ova, 
To determine Qi, and era, a glass reservoir of known volume is 
filled with the phase to be investigated, while a small quantity of 
the coexisting phase is allowed to be present. This reservoir is joined 
to a narrow glass capillary, which allows it to be immersed in a 
bath of liquefied gas in one of the eryostats of the cryogenic laboratory. 
The glass capillary is continued by a narrow steel capillary that 
‘an be closed by a tap. Since the measurement of the quantities of 
gas filling the densimeter is, as will be seen in § 3, made in a 
volumenometer, it would seem desirable to make use of different 
reservoirs if the densities of the liquid and vapour are to be determined 
with the same degree of accuracy; but for the purpose of our 
measurements this was not necessary. To determine a point on the 
diameter it is not necessary to obtain such a high percentage accuracy 
in the density of the vapour as in that of the liquid; it is sufficient 
if the quantity filling the reservoir used for both determinations is 
known with the same absolute accuracy in each case. Hence the 
same reservoir may be used for determining eu, and Qrap. 
As the determination of the vapour density necessitates an accurate 
knowledge of the volume of liquid that is left in equilibrium with 
the vapour, the reservoir of the densimeter terminates at its lower 
extremity m a graduated appendix «. Unfortunately, the capillary 
chosen for this appendix was so narrow that in the course of the 
measurements if was necessary to take the level of the liquid in the 
conical portion of the capillary above the graduation. In these cir- 
cumstances it was rather difficult to obtain this correction !). 
The shape of the reservoir was so chosen that the method of 
constant mass as well as that of constant volume could be employed. 
With this end in view the reservoir consisted in part of a graduated 
stem d.—d,; thus the apparatus formed a dilatometer provided with 
a very narrow capillary dy and a tap /,, as well as an appendix 
d,. While the tap £, remains closed, the liquid meniscus that at the 
1) This was done in the following way: Copies were constructed from pieces 
of the same kind of glass to exactly the same external dimensions. and of such 
internal dimensions as to be optically identical. These copies were then subjected 
to successive grindings and the internal dimensions were taken after each grinding; 
the volumes were then determined by integration. We gratefully acknowledge our 
indebtedness to Mr. G. Horsr for the care which he devoted to this part of the 
jevestigation. 
