242 Chloride of Aluminiurn and its Analysis. 



oxide of aluminium being considered by them and others as a sesqui- 

 oxide, from its isomorphism with the oxide of chromium, and the 

 peroxides of iron and manganese. With a view of ascertaining 

 whether the chloride of aluminium be a chloride or sesquichloride, 

 the following analysis was undertaken. It was thought also, that by 

 means of an accurate analysis of this compound, the atomic weights 

 of aluminium and alumina might be determined. The compounds 

 of alumina heretofore analyzed have been such that no certain de- 

 ductions could be made of the atomic weight, of either the metal or 

 oxide of aluminium. Thomson from an analysis of many minerals 

 and compounds in which he deduces atomic weights varying from 

 2.0580 to 2.3168, oxygen being unity, finds the mean to be 2.24205, 

 and as this is near 2.25 he assumes that number as the atomic weight 

 of alumina, in accordance M'ith what he considers a general principle, 

 viz. that the atomic weights of all bodies are multiples by whole 

 numbers of the atomic weight of hydrogen.* ^ 



Berzelius gives the atomic weight of aluminium 1.71166, oxygen 

 being unity .f The numbers of Thomson and Berzelius for alumi- 

 nium being so different and in no simple ratio to each other, neither 

 is to be depended on as accurate, unless confirmed by accurate anal- 

 ysis. 



Lest there should be some unpercelved cause of error leading to 

 erroneous deductions, i\\] the steps cf the analysis are given below, 

 so that if there be error, it may be detected. 



The balance used weighs easily ^{^ grain, and is sensible to ^^^ 

 grain. 



Analysis of the Chloride of Aluminium. 



A. Some of the chloride of aluminium in fine crystals, when 

 first taken from the tube in which it was made, was put into a small 

 dry glass tube and hermetically sealed. The chloride had been kept 

 thus for three years, and had undergone no change in its appearance. 

 The tube when vv^iped dry and clean, weighed with its contents 5.027 

 grammes. The tube was then slightly cut with a file, the dust from 

 filing being made to fall into the scale cf the balance, then'broken in 

 two where it had been scratched by the file, and thrown into some 

 pure distilled water. An intense action with a hissing sound like that 

 of hot iron under water, continued until the chloride was entirely de- 



* Thomson's Elements of Chem. i. p. 306. 



t Gaudin, Ann. de Ch. et de Physique, T. Hi. p. 132. 



