xiv THE ATOMIC THEORY 315 



ployed by Dalton, can only give the equivalents of the 

 elements and not their true atomic weights. Dalton's table 

 of atomic weights was thus really only a table of equivalents, 

 selected so as to give the simplest possible formulae to the 

 different compounds, but bearing little resemblance to the 

 real atomic weights as determined by the methods described 

 in the following chapter. The only check upon his arbitrary 

 selection of formulae was that a heavy gas was usually 

 represented as containing more atoms than a lighter one ; 

 for this reason he regarded carbonic oxide as a binary 

 compound, and carbonic anhydride as ternary, and therefore 

 calculated the atomic weight of carbon from the ratio 6 : 8 

 of carbon to oxygen in carbonic oxide instead of from the ratio 

 3 : 8 of the combining weights of the two elements in car- 

 bonic anhydride. 



Stas's determination of equivalents. A description has 

 already been given of the methods used by Stas in deter- 

 mining the combining weights of carbon with oxygen 

 (pp. 148 153). During a period of about 25 years, ending 

 in 1882, Stas determined with extreme care and accuracy 

 the combining weights of all the elements quoted in 

 the above table, with the exception of calcium, copper 

 and mercury. The method he adopted was as 

 follows : 



(i) Potassium chlorate, which contains six equivalents of 

 oxygen in combination with one of potassium and one of 

 chlorine, was converted into potassium chloride 



(a) By ignition 



(b) By heating with hydrochloric acid. 



It was found that 100 parts of potassium chlorate gave 

 60-846 parts of potassium chloride, so that 6x8 parts of 

 oxygen were combined with 48-^(100 - 60*846) x 60*846 = 

 74-592 parts of potassium chloride ; this number is, 

 therefore, the equivalent of potassium chloride. 



