on the Nature ofcertcim Bodies. 343 



through the atmosphere, a'psorbs in different cases different 

 fjiiantities of oxygen and of moisture ; during the intervals 

 ol" the removal of the difit-rent portions of gas, likewise, 

 aome globules are lost. 



In instances when the heat has been more rapidly raised, 

 I have generally found n-.orf poiassium destroyed, and less 

 nitrogen in proportion in the aeriform products. \n such 

 cases, likewise, the !os& of weight has betn much srreater; 

 the gases have been always clouded, and the adaptors, after 

 being exposed to a moist air, emitted a smel) of ammonia ; 

 from which it seem? likely that snjali quantities of the dark 

 gray substance described in page 50 of thisvohune* are 

 sometimes carried over undecompounded in the operation. 



In some la!e experiments, I substituted for the iron tube, 

 a tube of copper, which had been bored from a solid piece, 

 and the sides of which were nearly a quarter of an inch in 

 thickness. Mv object i}j usins this tube was not ^nlv to 

 prevent the heat from being too rapidly coiT)nninicated lo 

 the fusible substance, bui likewise io be secure that no me- 

 tallic oxide was present ; for thouch tiie iron lubes had been 

 carefuliv cleaned, yet still it was possible that some oxide, 

 which could not be separated from the welded parts, might 

 exist, which of course would occasion the disappearance of 

 a certain quantity of potassium. 



I shall civc the results of one of the processes, which I 

 regard as most correct, made in the tube of copper. The 

 barometer was at 30'3 ; ihermometer wasat 59" Fahrenheit, 



The tube contained two cubical inches and half, and was 

 filled with hydrogen. 



Six grains of potassium, which had absorbed 13 cubical 

 inches of ammonia in a copper trav, were employed. 



The adaptors connected with the mercurial apparatus and 

 the stop-cocks, contained "7 of atmospherical air. 



The cas given off was collected in ixvo portions. 



The tirst portion wa=? equal to 1 1 cubical inches. It con- 

 tained '8 of amnujuia ; 11 of the residuum, detonated with 

 b «)f oxygen, It it 8. 



The second |x>nion equalled 9 cubical inches. They 

 contained no ammonia, to of this gas, with 8 of oxygen 

 detonated, left a residuum of 10. 



.There remained in the lube and adaptors I'l cubical inch 

 of gas. 



'J'he quantity of hydrogen produced bv the action of the 

 potassium, which had been regenerated, equalled 4* 3 cu- 

 bical inchci. 



• Pliil. Mag. vol. «xi'ti. page IS". 



Y4 lu 



