DETERMINATION OF ATOMIC WEIGHTS 527 



01 the reduced temperature x an ^ deduces two equations to 

 represent the values of mp c and np* as functions of %, within 

 the limits of experimental error. These equations are 



io 4 . mp c = i8-8sx Ox 3 - ^2 X + 2 V5 X - 1) < 28 ) 



and 



io 4 . np 2 c = 3-5x 3 (x - ( 2 9) 



which represent the results of Leduc's experiments on the com- 

 pressibilities of gases with great accuracy. 1 



The molecular volume method is applied in the following 

 manner. Given the density d of a gas at temperature T and 

 pressure p and given its critical temperature T c and critical 

 pressure p c : required its molecular weight. Since p c is given 

 and x, which equals T c /T, is also known, equations (28) and (29) 

 enable m and n to be calculated. Equation (27) then gives the 

 value of (f>, as e, which equals p/p c , is also known. It is then 

 necessary to be able to calculate <f> , } the value for oxygen at 

 the same temperature T and pressur ' p and the required 

 molecular weight follows from equation (24). In practice, T 

 and p are 273 and 1 atmos. respectively. 



Leduc found that the theorem of corresponding states could 

 not be applied to certain gases, i.e. that equations (28) and (29), 

 which give the correct values of m and n for a large number 

 of gases, do not give correct results in these particular cases. 

 These exceptional gases are ammonia, phosphine, hydrogen 

 sulphide and methylic ether. On the other hand, equations 

 (28) and (29) derived from data relating to substances gaseous 

 at the ordinary temperature and pressure may be successfully 

 applied to the calculation of A*, for toluene vapour at 129-6° C., 

 the result being in excellent agreement with that obtained from 

 the experimental data of Ramsay and Steele (8). 



Since equations (28) and (29) rest largely upon the com- 

 pressibility data of Leduc and Sacerdote and upon critical 

 constants which often differ a little from those hitherto employed 

 in this article, a detailed statement of the numerical results 

 obtained by this method is unnecessary, as the values would 

 not be directly comparable with those already deduced by other 

 methods. It is obvious that the results obtained by this 



1 Leduc deduced an expression for A*, in terms of m and n and then sought 

 equations for m and n which would enable him to reproduce his experimental 

 values of A*. 



34 



