IS On the Nature of Alkalis and Bar ths. 



At a red heat, even the purest glass is altered by the basil 

 of potash : the oxygen in the alkali of the glass seems to be 

 divided between the two bases, the basis of potash and the 

 alkaline basis in the glass, and oxides, in the first degree of 

 oxygenation, are the result. When the basis of potash is' 

 heated in tubes made of plate glass filled with the vapour of 

 naphtha, it first acts upon the small quantity of the oxides 

 of cobalt and manganese in the interior surface of the glass, 

 and a portion of alkali is formed. As the heat approaches 

 to redness, it begins to rise in vapour", and condenses in the 

 colder parts of the tube; but at the point where the heat is 

 strongest, a part of the vapour seems to penetrate the glass, 

 rendering it of a deep red brown colour;' and by repeatedly 

 distilling and heating the substance in a close tube of this 

 kind, it finally loses its metallic form, and a thick brown 

 crust, which slowly decomposes water, and which com- 

 bines with oxygen when exposed to air forming alkali, lines 

 the interior of the tube, and in many parts is found pene- 

 trating through its substance*. 



In my first experiments on the distillation of the basis of 

 potash, I had great difficulty in accounting for these pheno- 

 mena; but the knowledge of the substance it forms in its 

 first .degree of union with oxygen, afforded a satisfactory 

 explanation. 



[To.be continued,! 



II. On the Opinions that have prevailed respecting the Na- 

 ture of Alkalis and Earths. By a Correspondent. 



To Mr, Tillock. 

 sin, 

 X have not been a little amused by the manner In which 

 your correspondent O. begins his letter f. 



" Messrs. Davy, Berzelius, and Pont in, have only verified 



* This is the obvious explanation in the present state of our knowledge : 

 but it is more than probable that the silex of the glass likewise suffers some 

 change, and probably decomposition. This subject I hope to be able to re- 

 sume on another occasion. 



f Phil. Mag. vol. xxxi. p. 273. 



what 



