666 
PROFESSOR W. RAMSAY AND DR. J. SHIELDS ON THE 
upper dome of liquid ; and with a narrower tube, the temperature above Avhich the 
bubble could not be induced to remain stationary was much higher. In no case did 
it ascend until it had been made approximately spherical by compression. 
Now, had there been an angle of contact, it would have been necessary to compress 
the bubble until its shape had become lenticular, i.e., until its vertical diameter had 
become less than its horizontal diameter. This was never observed, and we conclude 
that there is, therefore, no angle of contact between a liquid which wets glass and 
the glass, if its surface is in presence only of its own vapour. But we have obtained 
evidence that the capillary rise in air differs to a small extent from that in the vapour 
of the liquid, and, under these circumstances, there may well be an angle of contact. 
For example, with ethyl oxide, the read height in air was 37‘40 millims., while in 
contact with its own vapour the rise was 36'5 millims. ; with methyl formate, the 
numbers are 40'9 and 39‘9 ; with carbon tetrachloride, 27’5 and 25‘4; with benzene, 
519 and 49-6. 
Such discrepancies might be due to the angle of contact existing in presence of air, 
but not in presence of vapour; or it might be due to a film of moisture on the surface 
of the liquids. But it is noteworthy that those liquids which are most easily volatile 
at ordinary temperatures give measurements in air which most nearly corroborate 
the rectilinear relation. The surface-energies, moreover, calculated from the rise of a 
liquid in contact with its own vapour, lie on the curve prolonged to — 89°'8 ; while 
those calculated froin the capillary rise in air lie off the curves to a greater or less 
extent. 
The reduction of the read heights in the capillary tube to dynes per linear centi¬ 
metre was effected in the usual manner, by means of the equation 
7 = 2 {p — 
where r is the radius of the tube, k the height to which the liquid is raised, g the 
gravitation constant, p the density of the liquid, and tr that of the vapour. 
In the accompanying Tables the experimental results are stated. Column 1 shows 
the temperatures at which measurements were made; column 2, the observed capil¬ 
lary rise in the narrow tul)e which is corrected in column 3 (Ag), so as to eliminate 
the effect of the barometer tube in which the measurements were made ; column 4 
(Ag) reproduces these numbers smoothed, so as to eliminate errors of experiment; 
p and cr are the weights of 1 cub centim. of the liquid and the gas respectively ; and 
y is the surface tension, calculated by the equation already mentioned. 
