which liecomes a violet-coloured Gas hj Heat. 9 



the solution with the substance, considerable quantities are ob- 

 tained. By pouring off the mother-Hquor and evaporating it a 

 httle, more of the crystals fail down. 



All these crystals, if precipitated from a solution not too 

 much saturated, are of tlie same kind ; they are little soluble in 

 water, have a taste analogous to that of ihe hyper-oxvmuriates 

 of potassa, scintillate when throwii upon burning coals, and 

 form a defiag;ratiag mixture when mixed with ci'.arcoal. When 

 fused, they give off abundance of oxygen gas, and become the 

 same substance as that formed by the action of potassium on 

 the new substance. 



If the liquor which has ceased to afford the crystals be eva- 

 porated to dryness, it yields a considerable quantity of a sub- 

 stance which is not capable of detonating with combustible 

 bodies, and which is the same as that afforded by the com^ 

 bination of the substance with potassium. 



It is evident then that the oxygen contained in the potassa is 

 newly combined by the action of the new substance, and two 

 compounds formed ; one consisting of potassium and the new 

 substance, the other of potassiimi, the new substance, and the 

 oxygen contained in the potassa. 



By passing the vapour of the substance over dry red hot po- 

 tassa formed from potassium, oxvgen is expelled, and it appears 

 that oxygen cannot remain in the triple compound at a heat 

 above the red heat. 



By dissolving the substance in solutions of soda and baryta 

 similar results are obtained, and in each case two compounds 

 are formed. The oxygen is condcnse;d in one, and the other 

 consists simply of the new substance and a metal. 



To separate entirely the deflagrating salt from the soluble 

 salt is not easy, there always remains in the mother-liquor a 

 little of tbe dellagrating compound ; but by separating the first 

 crystals from solution of potassa not too strong, the deflagrating 

 compound of oxygen, potassium, and the substance is obtained, 

 apparently pure. 



As the new substance combines with potassium and the me- 

 tals with much less energy than chlorine, it occurred to me, 

 that it vvoidd probably be expelled from its combinations by that 

 body ; and this I have found to be the case in all the experi- 

 ments I have made. When the compound of the new substance 

 and potassium is heated in contact with chlorine, potassane 

 (nmriate of potassa) is formed, the violet gas appears, but soon 

 combines with chlorine, and they form togetber the peculiar acid 

 compound 1 have before described ; but towards the end of the 

 process, as the proportion of chlorine diminishes, tlic violet-co- 

 joured gas again appears. 



When 



