LARGER CHEMICAL ERRORS. 53 



But the product is in excellently weighable condition, 

 and contains the metal wherewith boron was combined in 

 the substance, borax. 



Hence the value of the process depends entirely on the 

 completeness of the chemical reaction used. 



Sometimes it is almost impossible to obtain an exactly 

 weighable substance for the initial material, because of the 

 difficulty of removing definitely all water of crystallization 

 or some other secondary constituent. 



If now that compound permits the exact determination 

 of the element sought and some other of which the atomic 

 weight is known, good atomic weight determinations are 

 possible, though the original substance cannot be weighed. 



The most valuable application of this method we find in 

 the splendid work of E. Maumene" on silver acetate. True 

 Atomic Weights, p. 196; 1894. 



He determined the silver and carbon in this compound. 

 The silver as metallic residue, the carbon as dioxide. 



We shall find this the only strictly unobjectionable deter- 

 mination of the atomic weight of silver. 



The method used for the determination of the atomic 

 weight of uranium in the laboratory of Professor Armand 

 Gautier is of the same general character. In this case, the 

 atomic weight of uranium is expressed in that of nitrogen. 

 See pp. 35-36 above. 



Methods Giving Variable Results. 



We have not yet referred to the conversion of a weighed 

 amount of a pure metal into a definite salt by means of an 

 acid. 



In this way Berzelius produced lead sulphate, and Stas 

 obtained also the nitrate of lead and of silver. 



Stas has laid great stress upon these syntheses of silver 

 and lead nitrate. He even challenged the chemists of the 

 world to show that his results were not exact. True Atomic 

 Weights, p. 34. 



But his own data show that this method is not applicable 

 to atomic weight determinations for silver and lead. 



