464 Proceedings of the Royal Society of Edinburgh. [Sess. 
and the dilatometer is introduced in such a manner that the upper end of 
the bulb AB lies about 5 mm. from the lower end of the plug S (fig. 2), the 
whole dilatometer thus being roughly in the same position as shown for the 
thermoelement T in the figure. The dilatometer is held in a fixed vertical 
position by means of a horizontal copper platform screwed into A (fig. 2), 
and also by a number of wires which bind it tightly to the outside of N. 
The quartz-glass-covered thermoelement is similarly introduced and fastened 
to N. The thermoelement terminals and the wires HIT are connected to 
binding screws which connect with the lower ends of the conductors e. The 
plug is then inserted in the compression-cylinder and all the outer connections 
made. By a few strokes of the hand-pump the pressure is raised to 
approximately 300 atmospheres. The heating current is then gradually 
switched in. The heating gradually raises the pressure inside the cylinder ; 
and after a little practice it is possible to arrange the initial pressure such 
that the final pressure desired is just attained at the desired temperature, 
or at least attainable within the range of action of the regulating screw. 
By means of this regulating screw the pressure could be kept constant to 
within 10 atmospheres. At this stage two experimenters were usually 
required — one to regulate the pressure and temperature, and the other to read 
the temperature and mark the galvanometer deflections. The temperature 
at which contact was made, and the temperature at which contact was broken, 
were read on an average three times each. The normal duration of such a 
series for one value of the pressure was about 10 minutes, as it was not 
advisable to continue heating longer, lest damage should be done to the 
insulations. The cylinder was accordingly left to cool for about half an 
hour, the pressure being meanwhile kept high. After that time the pressure 
was adjusted at a different value and the observations repeated. At the 
end the plug is again removed and the dilatometer taken out. If every- 
thing is in order, the mercury in the bulb is collected and weighed, as a 
check on the original volume. The dilatometer is then ready to be refilled. 
Each separate dilatometer bulb could be used only a few times, owing both 
to the material being attacked at the high temperature and to the fact that 
some gas — probably hydrogen — seems gradually to occlude itself in the walls 
of the bulb. After a few experiments, even at comparatively low pressures 
and temperatures, this gas forces its way into the interior of the bulb and 
reduces the value of the observations. Whenever it began to be trouble- 
some, the old bulb was cut off and a new one fused on without interfering 
with the other parts of the dilatometer. 
To give an idea of the various quantities involved, I will quote the 
details of a single series. 
