54 THE ERRORS OF PRECISION. 



This has been fully shown in our True Atomic Weights, 

 1894, and in the Comptes Rendus, T. 116, pp. 431-433; 1893. 

 We shall again demonstrate it in this work, but shall not 

 enter into any detail at this point. 



When the silver nitrate produced per unit of weight of 

 the metal quite notably varies with the amount of silver 

 operated upon, and even constantly differs according as the 

 nitrate is " dried " or " fused," the process used is simply 

 not fit for the determination of atomic weights. 



Methods of a dubious value are those wet way processes 

 in which a precipitate is separated and weighed. All 

 chemists understand that these methods are not quite exact 

 for atomic weight determinations. 



Neither barium sulphate nor silver chloride is absolutely 

 insoluble in the liquid used, nor absolutely free from foreign 

 matter. 



Processes of this kind must be expected to give values 

 not quite exact. 



False Methods. 



Even common acidimetric tests have been applied for 

 atomic weight determinations, by Julius Thomsen of Copen- 

 hagen, and Richards of Harvard. 



Richards tried to determine the sulphuric acid left after 

 the electrolysis of blue vitriol by this method and got won- 

 derful results, since he overlooked that a part had changed 

 to persulphuric acid, which has only half the saturating 

 capacity. See True Atomic Weights, pp. 135-136. 



Here also must be mentioned the volumetric process of 

 Rimbach (1893) on borax with hydrogen chloride, using 

 methyl orange as indicator. 



It is really strange that chemists can so far forget the 

 fundamental requirements of atomic weight determinations 

 as to think of volumetric processes of this kind. 



But there is one wet way process which has caused many 

 errors to take deep root in the chemistry of atomic weight 

 determination. Together with the Stasian syntheses of the 

 nitrates, the volumetric silver chloride process has muddled 

 this part of the science for almost forty years. 



