268 SCIENCE PROGRESS. 



ing equal numbers of molecules, R a constant, and T tem- 

 perature reckoned upwards from absolute zero, where pres- 

 sure disappears. Ramsay and Shields have pointed out that 

 there exists an analogous equation for liquids. If we repre- 

 sent the surface tension of a liquid by y, the surface on which 

 equal numbers of molecules lie by s, a constant analogous to 

 R by k, and temperature measured downwards from the 

 critical temperature, where the surface tension becomes 

 zero, by r, then the analogous equation for liquids is, 

 ys = kr. To this equation, however, a slight correction in 

 the shape of another constant, d, must be added, as the line 

 representing the product of surface tension into molecular 

 surface has its origin, not at the critical temperature, but 

 usually about d = 6° below it. The equation, therefore, 

 becomes ys = k (t — a). 



As gases have never been investigated at temperatures 

 closely approaching the absolute zero we have no means of 

 knowing whether a correction, similar to d, should be applied 

 or not. 



All the terms in the above equation are either known or 

 susceptible to measurement. Eb'tvos determined the surface 

 tension by an optical method, Ramsay and Shields by mea- 

 suring the capillary rise in a tube of narrow bore. 



In order to obtain good results several refinements of 

 the ordinary methods are necessary, for the details of which 

 the original memoirs should be consulted. In a good many 

 instances the surface tension has been determined for a 

 range between the ordinary temperature and the critical 

 temperature ; in some cases even as low as — 90°C. 



Assuming that the distribution of the molecules is the 

 same on the surface of a liquid as in the interior, the next 

 term in the equation, s, the surface on which equal numbers 

 of molecules lie, is determined as follows. If z/ is the 

 specific volume of the liquid and M its molecular weight, 

 then Mv is its molecular volume, that is to say, the volume 

 in cubic centimetres occupied by a gram-molecule of the 

 liquid. Now, let us conceive a cube with a volume equal 

 to the molecular volume of the liquid ; the cube root of the 

 molecular volume, on the above assumption, obviously gives 



