CHANGE FROM THE LIQUID TO THE SOLID STATE ON YAPOUR-PRESSURE. 471 
considerable length of tube. Another was the use of a glycerine barometer and 
gauge. A third was to suspend the barometer and the gauge from the arms of a 
balance, while the lower ends dipped into two troughs communicating with each other. 
The diameters of the tubes being known, the difference in level could be ascertained 
by a difference in weight. This last method promised success, but it was abandoned 
in favour of a much more simple one. 
26. The principle of this method is to observe the temperatures of both ice and 
water while at the same pressure. Absolute accuracy in reading pressure is, as will 
hereafter be shown, not essential. The apparatus by which this was accomplished is 
shown in Plate 18, fig. 2. 
A and B are the two thermometers, graduated in tenth-degrees, the bulbs of which 
are covered with cotton wool. The arrangements for introducing liquid, so as to 
moisten the cotton wool, are similar to those already described. The two vertical 
tubes dipped in water, the temperature of which was maintained at 35°-40°. The 
condenser differed but slightly from that previously used. The altered position of the 
exit tube C was found to be more favourable to condensation. The arrangement of 
the gauge and barometer is also shown in the figure. In order to ensure that the 
diameter of the gauge was the same as that of the barometer, the portion of the gauge 
D and the top of the barometer tube E were taken from adjacent portions of the same 
tube. The divisions on the mirror scale F were extremely fine, and an attempt was 
made to read to half-tenths of millimeters, with what success will be seen hereafter. In 
reading pressures, the level of the mercury in the trough was frequently altered, so as 
to avoid reading only at one part of the scale. As a rule, the mean of several readings 
is given. In order to facilitate exhaustion, the gauge was connected by means of a 
T-tube with a Carre’s air-pump, as well as with a Sprengel pump, the former serving 
as a rough and the latter as a fine adjustment. In these experiments, after exhaustion 
by the Carre pump, the lead tube connecting that pump with branch G of the 
T-tube was removed, and this junction served to admit air when desired. The india- 
rubber tubing, which was specially made for the purpose by Messrs. Thornton, of 
Edinburgh, was very thick-walled, so as to avoid collapse on reduction of pressure, 
and was impervious to air. The india-rubber corks through which the thermo¬ 
meters passed were coated with paraffin. These precautions prevented any leakage 
whatsoever. 
As before, a freezing mixture of pounded ice and hydrochloric acid surrounded the 
condenser. 
27. Three thermometers were used, A and C by Negretti and Zambra, B by 
Cetti. Their zero-points were first determined, and a careful comparison of the scale 
below 0° was made by placing them in position, freezing water on the cotton wool 
covering both bulbs, and altering pressure. In this comparison both thermometers 
were under the same pressure, and the volatilizing point of ice being solely dependent 
on the pressure, both thermometers were under precisely the same conditions, and 
