LARGER CHEMICAL ERRORS. 55 



It will be necessary to say a few words about this process 

 by way of a general protest against its common use. 



The silver chloride volumetric process is most valuable in 

 technical analyses, and nothing here to be said is intended 

 to reflect upon the method of Gay-Lussac and Alohr when 

 restricted to technical problems. 



But we must insist that this method is unfit for the work 

 of atomic weight determinations. See True Atomic Weights, 

 pp. 121-128. 



In order to keep all the silver in the precipitate, there 

 must be an excess of the chloride. But this solution then 

 reacts again with a drop of silver solution. 



If the solution gives no further reaction with silver, it will 

 again react upon the addition of a chloride. 



Clearly, the silver chloride precipitate is held down by 

 an excess of the soluble chloride; the amount of silver is in 

 no fixed proportion to the amount of chloride in the liquid. 



These facts were fully presented by Mulder, and have 

 been admitted by Stas, who supposed that the mean between 

 the silver and chloride limit marks the true compound. 



We have no inclination to consider fine spun imagination 

 such as this one of Stas or the apparent " ion "-philosophy 

 of Hoitsema presented in Ostwald's Zeitschrift (XX, 272- 

 282; 1896). 



It is sufficient for us to know that the chlorine and the 

 silver are not present in fixed, definite proportion in this 

 process, but vary very greatly. 



We are tired of being called to facts, when the facts are 

 imaginations and dreams in the head of so-called exact 

 chemists. 



Chemists must cease to take the fancies of Stas and his 

 school as facts. 



We shall not discuss this point, but insist on the facts. 



It was Pelouze, about 1845, who introduced the volumetric 

 silver chloride process into atomic weight determinations. 



This ready method was unfortunately used extensively by 

 Dumas up to the time when Stas began his pretentious work. 



