Theory of Heat to the Study of Volatile Liquids. 485 



tity 0*041, or very nearly. For liquids so different as mercury 

 and sulphurous acid, the difference is only 0*003, or a deviation 

 of less than -fa. 



Having verified our hypothesis concerning Carnot's cycle 

 and that of volatile liquids, we return to our hypothesis relative 

 to the cohesion of liquids. We suppose that atomic cohesion is 

 constant for all substances affecting the liquid state. Let a de- 

 note the weight of a liquid atom, and F the cohesion which 

 connects it with those surrounding it ; let P be the pressure 

 at the temperature t° ; and let us choose for all liquids a con- 

 stant pressure P = 760 millims., which corresponds to the boil- 

 ing-point. 



Since the external pressure, exerted by the vapour on the 

 liquid, is constant, the heat-effort and its work of disgregation 

 is exactly represented by the internal latent heat at that tempe- 

 rature. Now, if F is the force acting according to any law 

 through a space equal for all liquids, the work of disgregation 

 to separate into the gaseous state an atom of liquid will be Fk 

 ; a constant. 



The number of atoms contained in a weight 1 is - ; but a 



must be taken at its value for the temperature t — that is to say, 

 the density of the atom variable according to the temperature 

 of the boiling-point. Integrating the elementary work for the 

 weight of 1 kilogramme, the sum will represent the internal 

 latent heat of the liquid at the temperature t°, and we have the 

 fundamental equation 



K(274 + Q 



X ~ ax 274 ' {VL) 



whence we deduce 



3^=R 5 (VII.) 



274 + £ ' v ' 



and making the constant factor 274 to vanish for all liquids, 

 we arrive at 



27^7 = K > < VIIL > 



an equation connecting the atomic weight with the tempera- 

 ture and the internal latent heat. 



Before making the numerical verification, we will remark 

 that the external latent heat is sensibly proportional to a ; so 

 that the equation (VIII.) will refer with sufficient accuracy 

 either to the total latent heat, or to the internal latent heat 

 alone. In the following Table we give the products of the 

 atomic weight by the two latent heats. 



