Contact-Angle of Liquids and Solids. 179 



15-28 at 21°*5, and again 15*39 at 16°'7, and, by the method 

 of this paper, 15*17 at 23°*2. The values of a 2 obtained on 

 the same days by the " k " method, in the first two cases with 

 bubbles formed under a flat plate, and in the third under the 

 lens as described, are respectively 15*18, 14*93, and 14*96. 

 The evidence afforded by these consistent differences of the 

 results by the two methods is strengthened by the fact that 

 experiments on the rise of water in capillary tubes by Gay- 

 Lussac*, Frankenheim and Sondhausf, and BrunnerJ, give 

 for a 2 at 20° the value 14*80, and those of Volkmann§ give 

 14*90. Assuming that the "</" values are correct, these 

 values are such as would follow from the method of capillary 

 tubes if there were a small finite contact-angle, and if in the 

 computations the contact-angle w T ere assumed to be zero. 

 The differences in the values of a 2 are, how r ever, so small, and 

 the difficulty of obtaining surfaces of glass really clean so 

 great, that, while these experiments point to a finite contact- 

 angle between water and glass as probable, yet 1 do not feel 

 that they demonstrate its existence. 



Ethyl- Alcohol. — The results for ethyl-alcohol have been 

 presented fully, although those of April 6 and April 23 evi- 

 dently contain some error. I am satisfied that the magnitude 

 of the u k" results on those days is not due to any constant 

 error of observation, and suggest the following as its probable 

 explanation. The observations of those days were made, as 

 has been described, after the centre of the lens' surface was 

 cleared of liquid by evaporation. It seems to me probable 

 that, owing to evaporation at the circle of contact, in the way 

 discussed later, the water taken up by the alcohol from the air 

 blown into the bubble was left behind on the lens by the more 

 rapid evaporation of the alcohol, as explained by Prof. J. 

 Thomson ||, and that there was thus formed in the upper por- 

 tion of the surface of the bubble a film of liquid with a higher 

 surface-tension than that of the pure alcohol. The bubble 

 was thus made higher than it would have been if the surface- 

 tension had been throughout that of pure alcohol, and the 

 value of a 2 , calculated from its height, was greater than the 

 true value. Since precipitation of alcohol- vapour was going 

 on upon the cooler lower surface of the bubble, that portion of 

 the bubble was not similarly affected, and the constant obtained 

 by the " q " method was that of pure alcohol. This explana- 



* Laplace, Mec. Cel. Suppl. a la Th. de VAct. Cap. p. 54. 

 t Pogg. Ann. lxxii. p. 211 (1847). 

 X Berlin Berichte, 1846, p. 181. 

 § Wied. Ann. xvii. p. 353 (1882). 

 || Maxwell, ' Theory of Heat,' p. 293. 



N2 



