460 
Proceedings of the Royal Society of Edinburgh. [Sess. 
therefore be used ; but both platinum and quartz-glass are suitable in this 
respect, the devitrification and loss of strength in the case of the latter 
being very small below 1200° C. In quantitative volume experiments it 
is practically essential that the dilatometer be made of some transparent 
material, and, if not essential, it is desirable that the coefficient of expansion 
of the material of the dilatometer should be small. Platinum is not trans- 
parent, and its coefficient of expansion is seventeen times that of quartz- 
glass. The latter is therefore more suitable, and the dilatometers used were 
all made from quartz-glass, obtained in tubes of various sizes from Sieverts 
and Kuhn, Cassel, and made to the desired shape by a local glass-blower. 
Two properties of quartz-glass which caused difficulties in the course of my 
work were : (1) it was found impracticable to fuse metal electrodes into the 
walls of a quartz-glass tube, and (2), as Berthelot and others have found, 
quartz-glass is at high temperatures not impervious to gases. How these 
difficulties affected the work will appear later. 
As regards the form and dimensions of the dilatometers, both were 
limited to a large extent by the space available in the compression-cylinder, 
and details of both appear in the course of the following discussion. 
In experiments involving the measurement of volume changes under 
pressure an electrical method of detecting these changes has often been 
found the most suitable. To take two examples : Tait, in his experiments 
on compressibility, and later Richards, both used such a method. Tait’s 
method, which was successfully employed by Amagat, can best be ex- 
emplified by a figure. 
The piezometer bulb is filled to a definite point C * with the liquid under 
investigation. The bend CD is filled with mercury. The part BC has a 
number of platinum electrodes fused into its walls at given distances 
apart. These are all connected with one another on the outside by wires 
wound round BC, and the highest and lowest are connected by wires with 
a circuit containing a battery and a galvanometer. When the liquid in 
AB is compressed, the mercury rises in CB, and, as it passes each contact, 
there is a change in the deflection of the galvanometer. 
Richards’ method consists in having only two platinum contacts at A 
and at B,f which are connected direct with the galvanometer and battery 
circuit. Mercury again fills the part CBA. The liquid in CD is compressed 
until the mercury level, originally above A, falls below A, causing a break 
in the galvanometer circuit. A known weight of mercury is added at A 
and the experiment repeated. 
In my experiments a similar volume change had to be measured, so I 
* Fig. 36. t Fig. 3a. 
