206 SCIENCE PROGRESS. 



working it as well as the neglect of certain precautions 

 which could well have been taken, and ought to have been 

 taken in an attempt to settle such an important constant as 

 the present ; such precautions as to weighing with counter- 

 poises of equal volume, for example, seem to have been 

 neglected. 



The method was to determine, firstly, the weight of 

 hydrogen given off from unit weight of aluminium when 

 dissolved in strong potash solution ; secondly, by supplying 

 oxygen to a small combustion chamber so as to burn the 

 hydrogen evolved from a known weight of aluminium, and 

 collect all the water formed in the apparatus, one gets thus 

 the gain of the equivalent amount of oxygen to the hydro- 

 gen and to the aluminium. The only corrections not of 

 the simplest order were due to the oxygen and hydrogen 

 remaining in the apparatus or which had to be evolved after 

 the combustion had ceased. It was not found possible to 

 burn all the hydrogen evolved completely as the current 

 became so very slow when a very little aluminium remained 

 undissolved. The aluminium did not require to be perfectly 

 pure as long as it gave off no other gas than hydrogen. It 

 was found that 162 "3705 grams of aluminium gave off 

 1 8*1778 grams of hydrogen giving the ratio 



Hydrogen 



p — = 0*111902 + "000015 



Aluminium 



as the mean of twenty-one experiments. 



The weight of oxygen required to combine with 86*9358 

 grains of aluminium (or rather with the hydrogen evolved 

 by its solution in potash) was found to be yyiSyG grams 

 from which we get the ratio 



°*W^ = -88787 ±00001 S 

 Aluminium 



from which two results we get 



O -88787 



5- = — — = 7 '9345 



Ho '11190 



or — = i?-86qo + "oo22 

 H D ~ 



We seem to have every reason now to regard it as com- 

 pletely proved that the atomic weight of oxygen is 15*87 to 



