1813.] a Chemical Knowledge of Manganese. 273 



By arranging the proportion of oxygen obtained by the diffe- 

 rent experimenters, as in the following table, we may readily 

 compare them together: — 



Oxygen In/ 



, —- * — 1 



Oxides. John. Berzelius, Davy. Bergman. 



1 — 7*0266 — 



2 15 1 1-0533 — 



3 25 28-107 26-58 25 



4 40 42-16 39-82 35 



S — 56-213 66-6 



All the experimenters agree very nearly in the composition 

 of the third oxide. Nor is the difference very great in the com- 

 position of the fourth. It is clear, from the experiments of 

 John, that it is with the green oxide, or second oxide of Ber- 

 zelius, that all the acids combine. Hence this is the oxide 

 which constitutes the base of all the salts. 



The first or protoxide of every metal must be a compound of 

 one atom of metal with one atom of oxygen. Hence its analysis 

 gives us the relative weights of an atom of the metal and an 

 atom of oxygen. If we consider Berzelius's first oxide as the 

 protoxide of manganese, it follows that the weight of an atom 

 of manganese is to the weight of an atom of oxygen as 100 : 

 7*0266. I have in my table, published in vol. ii. p. 42 of this 

 journal, adopted 1 to represent the weight of an atom of oxygen. 

 And an atom of manganese, on the supposition that Berzelius's 

 first oxide is accurate, would be 14*260. The oxides of manga- 

 nese given by Berzelius will be composed as follows : — 



1st Oxide of 1 atom manganese + 1 atom oxygen 



2d Oxide of 1 + 2 



3d Oxide of 1 + 4 



4th Oxide of 1 + 6 



5th Oxide of 1 + 8 



Hut if we suppose that the first oxide of Berzelius does not 

 exist (and his mode of obtaining it was far from satisfactory), 

 in that case the green oxide of John will be the real protoxide, 

 and the constitution of the respective oxides of manganese will 

 be ui> follows : — 



1. Protoxide, composed 1 atom metal + 1 atom oxygen. 



2. Deutoxide, 1 + 2 



J Vitoxide, 1 + 3 



1 . Peroxide, 1 + 4 



I confess this appears to me by far the most probable state- 

 • ■ t. In that case, tin. weight of an atohi of manganese will 

 Vol. II. N° IV. S 



