xv THE MOLECULAR THEORY 355 



expression in tables of molecular and atomic weights, based 

 largely upon physical measurements of density, e.g : 



H=i, C=i2, = i6, 8 = 32, Cl = 35i 



A glance at the two lists shows that the atomic weights 

 calculated from the vapour densities are either identical 

 with, or are simple multiples of, the equivalents determined 

 by analysis. The ratio of the atomic weight to the equiva- 

 lent is called the VALENCY of the element. As elements 

 combine together in simple atomic proportions the valency 

 must obviously be a simple number. In practice, therefore, 

 measurements of vapour density are used to determine the 

 valency of an element (i.e. the factor by which the equivalent 

 must be multiplied to give the atomic weight) rather than 

 the atomic weight itself. The accurate determination of an 

 Atomic weight therefore includes the following operations : 



(1) By means of a series of exact analyses, the equivalent 

 or combining weight of the element is determined with the 

 greatest accuracy possible. 



(2) Vapour-density measurements are made with as many 

 compounds as possible, but not necessarily attaining a very 

 high order of accuracy. From these the approximate value of 

 the atomic weight is easily deduced by Cannizzaro's method. 



(3) The ratio of the (approximate) atomic weight to the 

 (exact) equivalent approximates so nearly to an integer that 

 no doubt can exist as to the correct value of the valency. 

 The equivalent multiplied by the integral number repre- 

 senting the valency gives the exact atomic weight. 



These operations may be illustrated in the case of arsenic. 

 Exact analyses show that 24/99 parts of arsenic combine 

 with 8 parts of oxygen or with 35*46 parts of chlorine ; the 

 equivalent of arsenic is therefore 24*99. From the vapour- 

 density of its volatile compounds its atomic weight is seen 

 to be approximately 75. Its valency is therefore 3 and the 

 exact atomic weight is 3 x 24*99 = 74'97- 



A A 2 



