224 SCIENCE PROGRESS 



values which agree very well among themselves, they may differ 

 by 3 per cent, from those of other workers. It is also note- 

 worthy that the capillary rise results are generally lower than 

 those obtained by other methods. 



Before describing the method of measurement, it will be 

 well to consider the causes of the discrepancies pointed out 

 above. Willard Gibbs has shown that, in order to keep the 

 potential energy of the system at a minimum, a solute which 

 diminishes the surface tension will be concentrated in the surface 

 layer, while a solute which raises the surface tension will have a 

 lower concentration in the surface layer than in the main bulk of 

 the liquid. Now the surface layer in a capillary tube has very 

 small dimensions, and thus a very minute trace of impurity, 

 which lowers the surface tension, is capable of giving a high 

 concentration in this very small surface, and in consequence 

 may produce a lowering of the surface tension altogether out of 

 proportion to its mass. 



The angle of contact has been generally assumed to be zero 

 in the case of liquids which wet glass, and cos « is then equal to 

 I ; since cos 2°3o' is 0-999 and cos 8° is 0-990, the errors which 

 would be introduced by taking these angles as zero would be 

 O'l and I per cent, respectively. In the case of a liquid such 

 as water, there seems to be an element of doubt as to whether 

 we are justified in taking the angle of contact to be zero. Some 

 writers, e.g. Ferguson (Science Progress, No. 35, 1915) and 

 Matthews (/. Phys. Chem., vol. xx, p. 554) think not ; while 

 others, e.g. Lord Rayleigh {Science Papers, vol. iii, p. 393)) 

 Bancroft (/. Phys. Chem., vol. xxi, p. 407), and Richards 

 and Coombs (/. Amer. Chem. Soc, vol. xxxvii, p. 1656) 

 either definitely consider the angle to be zero, or to be too small 

 to be of any practical importance. While the weight of the 

 evidence seems to be in favour of the latter view, further 

 experimental investigation of the point is undoubtedly 

 necessary. 



The height h given in the formula represents the distance 

 of the meniscus in the tube above the flat surface in the outer 

 vessel. It is probable that one of the chief reasons for the 

 variations between the results of different observers lies in the 

 fact that they have assumed the surface in the outer tube to 

 be flat, which is not the case. The question has been con- 

 sidered by Richards and Coombs {loc. cit.) ; they showed that 

 in a tube 2 cm. in diameter, with a thin rod in the centre, water 

 rose 0*31 mm. higher than in tubes on either side without the 

 central rod ; they found experimentally that a tube 3-8 cm. 

 in diameter gave a flat surface at its centre, all the capillary rise 

 taking place at the edges. Lord Rayleigh (Proc. Roy. Soc, 191 5) 

 calculated that the minimum diameter for a flat surface would 



