THE ERRORS OF PRECISION. 



last trace of the vapor of mercury by a gold leaf. The 

 results obtained by this dry way process are among the very 

 best in the annals of chemistry, as we shall find. 



Here the analytical ratio is the weight of the mercury 

 collected, divided by the weight of the oxide taken. 



Another class of dry way processes we have in simple 

 ignition or dissociation. 



The purity of the material used is generally depending 

 upon the process of crystallization and careful re-crystalli- 

 zation. 



Thus pure blue vitriol will leave the fixed black oxide of 

 copper, as practiced by Richards. 



Ammonium Alum leaves a fixed residue of Alumina 

 (Mallet). 



The remarkable finely crystallized Chloro and Bromo 

 Platinates leave upon ignition pure platinum direct or after 

 washing the residue with water according as the Ammonium 

 or Potassium Salt has been used. Most excellent determina- 

 tions have been made in this line by Seubert and Halberstadt. 



The related Potassium Bromo-Aurate has furnished, by 

 Kriiss, the most accurate determination of the atomic 

 weight of gold. 



The ignition of purest Iceland Spar gave us the most 

 reliable determination of the atomic weight of calcium 

 (Erdmann and Marchand). 



We shall find that the ignition of the purest native 

 magnesite (from Frankenstein) has given us, in the determi- 

 nations- of Scheerer, really the true atomic weight of mag- 

 nesium. 



All these dry way processes are simple, direct, complete, 

 and permit accurate weighings. They are necessarily the 

 most reliable, although modern chemists have not estimated 

 these processes properly. 



The starting material, the compound used, is often either 

 some native or artificial crystal. 



The purity of the starting material thus is dependent upon 

 the crystallizing power. 



Crystallized minerals, such as quartz and calcite and 



