21 
By partly exhausting the air at the pinch-cock end, which 
is then tightly closed, a partial vacuum is produced, and 
with my tube the pressure equalised itself after about six 
minutes, during which time a constant stream of hot air was 
passing through the capillary. With such an arrangement, 
and by always taking the precaution to cover the thermo- 
meter tube, whilst in use, with an open glass cap containing 
cotton wool, I found that a tube of this kind might be 
employed for an almost unlimited number of determinations. 
These determinations were made in the following manner : 
the liquid was placed in a small bottle ; the thermometer 
tube, after being sunk into the liquid and adjusted perpen- 
dicularly by a plumb line, was left for a short time. The 
height to which the liquid had risen was read off on the 
scale in millimetres with a magnifying glass, and the reading 
below the meniscus also taken. The difference gives the 
height in millimetres which the liquid has risen in the 
capillary. The tube is now raised a few millimetres, and 
the height again read off ; the process repeated once or twice 
and the mean height taken. With pure benzene, the 
following gives the height in millimetres of the column of 
liquid, determined at different times in the course of the 
present investigations : 
Height in millimetres. Temp. 
1 . 83-65 15 ° 
2 . 83-4 15 ° 
2 . 83-05 16 ° 
4 . 83-05 16 ° 
As is well known, in such determinations the temperature 
is a very important factor. As in the case just given, 1°C. 
corresponds to a difference of about *05 cm. I therefore care- 
fully noted the temperature of the room at the beginning of 
each reading, and took care to leave all the vessels and liquids 
there several hours before use. During the whole time the 
temperature remained between 15° and 16°. No precaution 
against evaporation was taken nor required, except in the 
