38 Prof. J. H. Poynting on 



and k, and we obtain as an approximation 



or 



(3) 



CO — CQ—p-YT 



At 0° 



w '- v=s .087, V'=209037, 



and the pressure required to lower the melting-point t° is 

 AA - OQ millim. by the well-known formula. Substituting in 



'VViOO 



equation (3), we get 



co-co' = '0Ut, 

 or 



-= -=- = '044 millim. of mercury, , . (4) 



which is KirchhofFs result. 



If the temperature be much below 0° C, we cannot make 

 these approximations without further examination, as the terms 

 containing k and k' in (2) may rise into importance. 



It may be noticed that (2) could be used as an equation to 

 determine p 9 the pressure required to produce a fall of the 

 melting-point to T, if there were any accurate experimental 

 method of measuring «r and v/. 



Effect of Pressure on the Melting-point. 



If we are right in regarding the change from the solid to 

 the liquid state as an exchange phenomenon in which the rate 

 of exchange is indicated by the vapour-tension, we ought to 

 be able to show that the pressure which lowers the melting- 

 point to a certain temperature will so alter the rate at which 

 the two states of the substance give off molecules from their 

 surfaces, that at that temperature there will be an equilibrium 

 of exchange. That is, we ought to be able to show that pressure 

 alters the vapour-tensions of the two states, but alters them by 

 different amounts, so that the equality of vapour- tensions now 

 occurs at the new melting-point. 



Now in the ordinary case, where the vapour-tension is mea- 

 sured we have the substance only under the pressure of its 

 own vapour ; but in the rise or fall of a liquid in a capillary 

 tube we may have a substance in contact with its own vapour 

 when the substance is at a very different pressure from the 

 vapour in contact with it. 



Sir William Thomson has shown (Proc. Eoy. Soc. Edinb. 

 1870, vol. vii. p. 63 ; Maxwell's ' Heat,' 1877, p. 287) that if 



