They find three oxides of potassium; the lowest degree is obtained by exposing potassium to atmospheric air in a small bottle, with a common cork; a gradual oxidation takes place; a blueish grey brittle product is obtained; there does not appear however, to be any proper limit to this oxidation besides that which they admit as characterizing the second degree or potash, which degree of oxidation may always be immediately obtained by placing potassium in contact with water. This I think should be called the protoxide and considered as 1 atom of potassium, and 1 of oxygen; before this point it is potassium and potash mixed or perhaps combined. Besides these there is another obtained by burning potassium in oxygen gas at an elevated temperature; this oxide is yellow, fusible by heat, and crystallizes in lamina on cooling; it contains three times as much oxygen as potash; put into water it is suddenly decomposed, giving out ⅔ of the oxygen in gas and becoming potash. Very probably an oxide containing twice as much oxygen as potash might be formed with some mark of discrimination, by uniting 18 parts potassium with 56 of yellow oxide, but this has not yet been done. According to these conclusions the weights of the oxides of potassium may be stated as under.--Potassium 35, protoxide or potash 42, deutoxide (supposed to exist) 49, and the yellow or tritoxide 56. Hence we have Potassium. Oxygen. Protoxide (potash) 100 + 20 } Gay Lussac & Thenard 19 } Davy Deutoxide 100 + 40 (unknown) Tritoxide 100 + 60 Gay Lussac & Thenard One feels unwilling to admit of a _tritoxide_, (and that perhaps the only one existing,) when the deutoxide is unknown, were it not upon good authority. The obscurity on this subject may be removed by future experiments. It may be proper to add that Gay Lussac and Thenard concur with Davy in assigning a much greater saturating power to potassium and sodium than to the fused hydrates of potash and soda of equal weights. From the table, Recherches, Tom. 2, p. 214, it may be deduced that 35 potassium require as much sulphuric acid to saturate them as 50 or more of the hydrate of potash; and that 21 sodium are equivalent to 36 or 37 hydrate of sodium. If these results are accurate, the weights of potassium and sodium, considered as hydrurets, cannot be as we have deduced them at pages 486 and 503, VOL. 1, namely, 43 and 29 respectively, but 35 and 21, as at page 262. 16. _Oxides of sodium._