I OF ARTS AND SCIENCES. 355 



of whose saturated vapors has been satisfactorily deter- 

 mined; and the indication is clearly that the cohesive at- 

 traction varies inversely as some power of the distance 

 between four and five. 



The constancy of this law may be established by com- 

 paring the tables for latent heats and surface tensions. 

 The latter is easily seen, as Sir William Thomson and others 

 have pointed out, to represent, numerically, a large frac- 

 tion of the Avork necessary to volatilize a film of the 

 thickness of one molecule; and neglecting this thickness, 

 which does not var}' through wide limits, we shall find 

 that the surface tension and latent heat are approximately 

 proportional. That is, not only the equivalent of the work 

 necessary to separate a liquid into the thinnest possible 

 laminae (or films) is approximately proportional to that 

 necessary to volatilize it completely, but also both of 

 these quantities are determined by the attraction of the 

 molecules for each other, which could not be the case 

 unless the law of variation of this attraction were the 

 same, or nearly the same, for all liquids. 



In no matter what way the cohesive pressure is calcu- 

 lated, we find that the latent heat varies very nearly in 

 proportion; thus the total latent heat of steam is 536 

 units, of alcohol vapor 202 units, and of ether vapor 91 

 units, corresponding to the order of the cohesive pressures 

 alreadv calculated, and also to the order of their surface 

 tensions, 81, 26 and 19 (dynes per centimetre). 



These points will be more carefully considered after an 

 analytical investigation of the actual relations which must 

 subsist and those which probably subsist between these 

 various quantities; at present all the evidence of the facts 

 is to show that the cohesive attraction between two mole- 

 cules varies as the fourth or fifth power of their distance, 

 inversely, and not as any power materially greater or less. 



There are various points of view which would make it 

 seem improbable that any such general truth can be estab- 



