PRESSURE ON THE TEMPERATURE OP VOLATILIZATION OF SOLIDS. 
39 
warmed, and the water which had melted from its interior surface transferred to A. 
The bulb A was then surrounded by a freezing mixture, and by gently warming B, 
all ice in contact with the sides was removed. The bulb B was now surrounded with 
boiling water, and the temperatures registered by the thermometers in A and B were 
carefully observed, while the temperature of the freezing mixture was altered from 
time to time. 
Fig. 2. 
A. Bulb, named condenser, placed in freezing-mixture. 
B. Pear-shaped bulb, containing a block of ice, C, frozen round thermometer. 
D. D'. Narrower tnbes fused on to the wider tubes, through which the thermometers 
passed, secured by wired indiarubber connexions. 
E. Exit tube for steam, closed by indiarubber tube and screw-clip. 
F. Hot bath of water or paraffin. 
G. Freezing mixture of hydrochloric acid and ice. 
During freezing the phenomenon of supersaturation was nearly always observed. 
The temperature fell occasionally as low as —11°, while the water was still liquid. A 
sudden formation of ice then occurred, and the temperature rose to 0°. After remain - 
ing constant for some time while ice was being formed on the sides of the bulb, we 
were surprised to find the temperature fall below 0° (on one occasion as low as —10°), 
before the whole of the water was frozen. This we afterwards ascertained to be 
owing to the thermometer transferring heat to the ice with which it was in contact, 
and which was cooled by the freezing mixture, for the temperature of the water had 
not fallen below 0°. 
5. The pressure in the apparatus could be increased only by raising the tempera¬ 
ture of the condenser or by admitting air. The former has the effect of increasing the 
total pressure of water-vapour, and also of the air already contained in the apparatus, 
and can be calculated from the tables of tension of vapour in contact with ice, given 
by Regnault. An attempt was made to measure the pressure of the air present by 
inverting the apparatus and transferring all the air to one bulb, and then reading the 
difference in level of the water in the two limbs; but as the apparatus itself is such a 
delicate air-thermometer, it was found impossible to obtain satisfactory results. The 
amount of air was therefore calculated by a method which will appear hereafter. 
