\ 



■46 On the Dalionian Theory of [July, 



Weight of an atom, 



29. Sodium , 5*882 p 



30. Barytium 8*731 "i 



31. Strontium 5*900 



32. Calcium , 2*620 



33. Magnesium 1368 ' 



Gay-Lussac and Thenard, I think it follows clearly that pure 

 potash is a compound of 100 potassium and 20 oxygen. Gay- 

 Lussac and Thenard have shown that the peroxide of potassium 

 is composed of 100 potassium and 60 oxygen. (Recherches Phy- 

 sico-Chimiqiies, i. 132.) What they consider as the protoxide 

 I believe to be a mixture of potassium and potash. These data 

 are sufficient to constitute potash a binary compound, and the 

 peroxide a quaternary compound. Hence an atom of potash 

 must weigh five times an atom of oxygen. 



p From the experiments of Davy, and of Gay-Lussac and 

 Thenard, it appears that soda is a compound of 100 sodium and 

 34*1 oxygen, and the peroxide of sodium of 100 sodium and 

 51*1 oxygen. Now 34*1 is to 51*1 as 2 : 3. Hence it follows 

 that soda must be a compound of 1 atom of sodium and 2 atoms 

 of oxygen ; and the peroxide, of 1 atom of sodium and 3 atoms 

 of oxygen. These data give us the weight of an atom of sodium 

 as in the table. 



q We have no direct data for determining the weight of an 

 atom of barytium, calcium, and strontium. The numbers in 

 the table have been calculated from one of the canons given in 

 a preceding part of this essay ; namely, that the portion of 

 bases which saturate the same weight of acid contain the same 

 weight of oxygen. Now to saturate 100 of sulphuric acid, 15)4 

 of barytes, 72' 11 of lime, and 138 of strontian, are necessary. 

 And each of these quantities must contain 20 oxygen. Hence 

 if we suppose them to be compounds of 1 atom of base and 1 

 atom of oxygen we easily deduce the numbers in the table. 

 They agree very nearly with those given already by Davy and 

 Berzelius. 



r This number is determined in the same way as the three last. 

 We have not sufficient data for determining the weight of an 

 atom of the metallic bases of the earths proper. The weight of 

 an integrant particle of some of these earths themselves may be 

 determined from the salts into which they enter with tolerabie 

 precision ; and if we were to consider each of them as prot- 

 oxides, it would be easy to determine the weight of an atr.m of 

 the base of each. But I do not see any proof, or even analogy, 

 that can lead us to consider them as protoxides. It would serve 

 no purpose, therefore, to introduce such arbitrary numbers into 

 the table. Perhaps the same objection will be -made to the bases 



