combinations of Oxymuriatic Gas and Oxygene, &c. 3 
to inflame, without heat in oxymuriatic gas. In all expert 
ments in which I fused the potassium upon glass, the retorts 
broke in pieces in consequence of the violence of the combus- 
tion, and even in two instances when I used the tray of platina. 
If oxymuriatic gas be used, not freed from vapour, or if the 
potassium has been previously exposed to the air, a little 
moisture always separates during the process of combustion. 
When pure potassium, and pure oxymuriatic gas are used, 
the result, as I have stated, is a mere binary compound, the 
same as muriate of potash, that has undergone ignition. 
The combustion of potassium and sodium in oxygene gas, 
is much less vivid than in oxymuriatic gas. From this pheno- 
menon, and from some others, I was inclined to believe that 
the attraction of these metals for oxygene is feebler, than 
their attraction for oxymuriatic gas. I made several experi- 
ments, which proved that this is the fact ; but before I enter 
upon a detail of them, it will be necessary to discuss more 
fully than I have yet attempted, the nature of the combina- 
tions of potassium and sodium with oxygene, and of potash 
and soda with water. 
I have stated in the last Bakerian Lecture, that potassium 
and sodium, when burnt in oxygene gas, produce potash and 
soda in a state of extreme dryness, and very difficult of fusion. 
In the experiments from which these conclusions are drawn, 
as I mentioned, I used trays of platina, and finding that this 
metal was oxidated in the operation, I heated the retort 
strongly, to expel any oxygene the platina might have ab- 
sorbed, and except in cases when this precaution w r as taken, 
I found the absorption of oxygene much greater than could 
be accounted for by the production of the alkalies. In all 
