DETERMINATION OF ATOMIC WEIGHTS 517 



would be directly proportional to its pressure. In no known 

 case, however, does this relationship hold ; and the extent to 

 which the measured densities deviate from the "theoretical" 

 can be used in deducing the value of Ao for the gas. Measure- 

 ments have been made by Baume (16) of the densities of sulphur 

 dioxide, methyl ether and methyl chloride at o° C. and at 

 pressures varying from 760 mm. to 311 mm. He has expressed 

 his results by means of the equation 



pv = 1 + m(i - -j-j (") 



in which L p and L, denote the weights of a litre of gas at o° C. 

 and under the pressures p and 1 atmos. respectively, m being 

 a constant. The pressure is expressed in atmospheres and at 

 N.T.P. the product pv = 1. This method of extrapolation is 

 similar in principle to that expressed in equation (7). 

 The following values were obtained : 



According to Baume (16) and also to Guye (18), we have 



m = A' 

 This conclusion, however, is erroneous and is due to the fact 

 that Baume, in his paper, defines Ap*, in two different ways 



which are not equivalent. As a matter of fact, we have from 



equation (n), 



p,v, s= 1 and p Vo = I + m, 



and since, from equation (2), 



A = I -pxV./poVo, 



it follows that 



Ao=i -(. + m)-' = -^- (12) 



The above values accordingly lead to the coefficients : 



so, . . . . A = 0*02325 



(CH 3 ) 2 . . . =0*02587 



CH3C1 = 0*02167 



which are much lower than those adopted by the Geneva 

 experimenters ; further, the agreement between the values for 

 sulphur dioxide obtained by this method and by that used by 

 Jacquerod and Scheuer (p. 5 16) vanishes. It will be noted, 

 however, that the figure for sulphur dioxide, obtained by 



