APPENDIX. 669 



tube filled with chloride of calcium, which renders it very dry. 

 The exhaustion being repeated in this way fifteen or twenty 

 times, the moisture which the oxide of copper so readily imbibes 

 is withdrawn, and the whole made dry. The decomposition is 

 then begun and completed in the way already explained. The 

 increase of weight of the chloride of calcium tube gives the quan- 

 tity of water formed, and the ninth part of the weight of water 

 is the amount of hydrogen contained in the substance under ana- 

 lysis. In general, the weight of hydrogen obtained exceeds a 

 little that of the hydrogen in the substance under analysis. That 

 is the consequence of the difficulty of depriving the oxide of cop- 

 per of all moisture. The excess is so much less when chromate 

 of lead is used: indeed, if the proper precautions be used, the 

 error in that case may be considered as evanescent. The whole 

 water is not absorbed by the chloride of calcium, a portion of it 

 is usually deposited in a liquid state in the small bulb at the end 

 of the chloride of calcium tube next the decomposing tube. 



As common cork imbibes moisture it cannot be used when we 

 wish to determine the hydrogen with very great accuracy. In 

 that case, the tubes should be ground into each other so as to be 

 air-tight. 



Liebig's potash tube answers so well for determining the car- 

 bon by the weight of the carbonic acid evolved, that no addi- 

 tional observations in that subject seem necessary. 



Liebig's method of determining the azote is somewhat differ- 

 ent from that described above, which is the method of Dumas. 

 He puts into the bottom of the decomposing tube a quantity of 

 hydrate of lime ; and after the combustion is at an end, the hy- 

 drate of lime is heated, and its water converted into steam, which 

 forces all the gas remaining in the apparatus into the gas tubes 

 standing over mercury. These are filled in succession, and the 

 ratio between the volume of carbonic acid and azotic gas being 

 determined, it is easy to calculate how much azote the substance 

 under analysis contained. 



But it must be acknowledged that both the process of Dumas 

 and of Liebig leaves considerable uncertainty, and that they af- 

 ford at best only approximations to the truth. A new method 

 has been recently proposed by Drs Varrentrapp and Will, which 

 is both of easier execution, and promises to be susceptible of 

 greater accuracy than any of the old methods.* 



* Ann. der Pharm. xxxix. 257. 



