672 
PROFESSOR W. RAMSAY AND DR. J. SHIELDS ON THE 
keep a considerable c[uantity—say, 100 centims.—of nitrous oxide liquid for several 
hours; and if the liquid be filtered into the tube, to remove particles of ice, it is 
absolutely limpid and clear; and the radiation of heat through the walls is so slow 
that it is only occasionally necessary to wipe off the dew which deposits. Readings 
may be made with a cathetonieter with perfect ease. It is, of course, necessary to 
have the nitrous oxide at a level considerably higher than that of the liquid in the 
capillary tube, so as to ensure that the surface of the liquid in the capillary tube 
shall be at the temperature — 89°'8, and some time must be given to allow the 
liquid to reach the low temperature. The rise is a slow one ; but in half-an-hour the 
position of the liquid in the capillary tube is permanent. 
The defect of this method is that readings must be taken at different parts of 
the capillary tube, for there is no means of adjusting the level of the liquid so that 
its surface is always at a definite point in the capillary. The differences which we 
have observed between the actual values of surface energy and the calculated values, 
which amount to about 2 per cent., are probably due to this cause, and to a slight 
error in determining the diameter of the tube ; it was found by weighing the 
mercury which filled a known length, while the diameter of the other tube was found 
by the much surer method of direct reading. 
Description of Plates 24 and 25. 
Plate 24 shows the height of ascent of liquid in the capillary tube, in millimetres, 
mapped against temperature as ordinates. Two points call for special notice. First, the 
continuous lines, indicating the rise, end abruptly at the respective critical points. In 
no case is there any change of curvature as the critical point is approached. Hence, 
although observations near the critical temperature are difficult to make, the correct¬ 
ness of readings from the curve is very jirobable, inasmuch as the curve could not 
follow any other course than that drawn. This enabled us to calculate the surface 
energy of chlorobenzene, a liquid which was not examined above 333°, and, moreover, 
it confirms the observations at lower temjieratures of that and the other liquids. 
Second, it is to be remarked that the lines showing the difference of heights with 
temperature of acetic acid and of the two alcohols are much more curved than those 
representing the other substances. 
The broken lines represent the surface tension in d 3 Uies. Here it is to be noticed 
that acetic acid and the alcohols give lines of double curvature, while the other 
substances give continuous curves. 
It is difficult to explain the behaviour of methyl formate below 20°. Eeadings 
taken on two separate days at — 89°'8 gave practically the same height, viz., 39'8 
and 40d millims. rise, in the wider capillary ; this corresponds to 54'5 and 54'9 
