Variation in the Pressure of Saturated Vapours. 45 



various savants for the pressure of vapours in a state of satu- 

 ration. 



It must be remembered, however, that equation (5) gives 

 only an approximate value for the vapour-pressure, except 

 for the temperatures T and those adjacent to it, when 

 there can be no difference between the results of calculation 

 and experiment; but at temperatures far removed from T 

 the difference may be very considerable. Recollecting the 

 assumptions and inexactitudes which were allowed in de- 

 ducing the equation (5), we must acknowledge it to be 

 perfectly applicable : (1) when the vapour, saturating its 

 space, subjects itself to Boyle's and Gay-Lussac's laws, or 

 varies very slightly from them ; (2) if c—Ci does not vary 

 with the temperature ; (3) if the coefficient of expansion 

 of the liquid be exceedingly small, and the pressure under 

 which it occurs be not abnormally high ; and (4) if the 

 volume (w) of the liquid be so small compared with the 

 rolume (v) of the vapour formed from it that w may be 

 taken as equal to zero. The first three conditions may be 

 allowed within certain limits of temperature, but the fourth 

 3an never be observed with any degree of great accuracy ; in 

 the case of small pressures w can be neglected without essen- 

 tial error, but with great pressures this is not allowable. The 

 iifference between the results of theory and experiment in- 

 3reases in proportion to the divergence from the above 

 conditions, and in the end they frequently differ entirely. 



When the dependence of the pressure of a vapour upon 

 my magnitude whatever is discovered, then it is easy to 

 express the specific volume of a vapour by means of the 

 same magnitude. The equation 



^=DT 



serves for this purpose. Let us also take 



i? v = DT ; 

 whence 



Vq T $ 



lo g7 = lo gT ~ lo gf- 

 l o ± o Po 



On replacing log— by its magnitude in equation (5) we shall 



have Po 



. v T c-^r, T T-T "! mr /l 1\ 



lo S^ = ^T + -WL log T - m -T-J + AD(T-"T > 



