THE ATOMIC THEORY 303 



Quantity used to 

 precipitate 100 parts 

 Method of Preparation. of silver. 



Chloride sublimed at ordinary pressure. Ratio I 49*597 

 determined at ordinary temperature . . .1 49*593 



I 49*597 



Chloride sublimed at ordinary pressure. Ratio I ? J?^ 

 determined at 100 . . . . . . j ^Q.. Q7 



Chloride sublimed in a vacuum. Ratio determined / 49*598 

 at ordinary temperature \ 49*592 



Mean .... 49 '997 



In these experiments the average deviation from the mean 

 was only 0*004 P er cent. 



The conclusion may be drawn that the law of fixed pro- 

 portions is exact up to the extreme limit of experimental 

 investigation. 



Soddy and Hyman (Trans. Chem. Soc., 1914, 105, 

 1402 1408) have shown recently that different samples of 

 " lead " may unite with quantities of chlorine which differ 

 by one part in 225, a result that has been confirmed by 

 Richards and Lembert (/. Amer. Chem. Soc., 1914, 36, 

 1329 1344). The difference is attributed to the presence 

 in the metal of two elements, which are even more difficult 

 to separate than NEODYMIUM and PRASEODYMIUM, which 

 were regarded for a long time as a single element under the 

 name of DIDYMIUM. The exceptional difficulty of resolving 

 such mixtures cannot be regarded as affecting the validity of 

 the law of fixed proportions, as it applies exclusively to pure 

 chemical compounds. 



C. MULTIPLE PROPORTIONS. 



The law of multiple proportions. The LAW OF 

 MULTIPLE PROPORTIONS states that : 



" If several compounds be formed, the fixed proportions in 

 which two elements combine together are in simple integral 

 ratios to one another" 



