combinations of Oxyrnuriatic Gas and Oxygene, &c. 15 
ensure a complete neutralization, and the tray heated to red- 
ness : there was no additional increase of weight. 
In the few experiments which I have made on the action of 
sodium and soda on oxyrnuriatic gas, the phenomena appeared 
precisely analogous ; but sodium, as might have been ex- 
pected, absorbed nearly twice as much oxyrnuriatic gas as 
potassium. 
When common salt that has been ignited, is heated with 
potassium, there is an immediate decomposition, and by 
giving the mixture a red heat, pure sodium is obtained ; and 
this process affords an easy mode, and the one I have always 
lately adopted for procuring that metal. No hydrogene is 
disengaged in this operation, and two parts of potassium I find 
produces rather more than one of sodium. 
From the series of proportions that I have communicated in 
my last paper, it is evident that 1 grain of potassium ought to 
absorb 1.08 cubical inches of oxyrnuriatic acid; and that the 
potash formed from one grain of potassium ought to decom- 
pose about 2.16 cubical inches of muriatic acid gas ; and these 
estimations agree very nearly with the result of experiments. 
The estimation of the composition of soda, as deduced from 
the experiments in the last Bakerian lecture, is 25.4 of 
oxygene to 74 .6 of metal, and this would give the number 
representing the proportion in which sodium combines with 
bodies 22. ;* from which it is evident, that a grain of sodium 
* Or if soda be considered as deutoxide, which seems probable from the experi- 
ments detailed page 4, 44 ; and on this supposion, the salts of soda must be conceived 
to contain double proportions of acid. On either datum the proportion of oxygene in 
water must be taken as 7,5, and that of hydrogene as 1 , though other numbers might 
be found as divisors or multiples of those which would equally harmonise with 
the general doctrine of definite proportions. In my last communication to the 
Society, I have quoted Mr. Dalton as the original Author of the hypothesis. 
