$2 Improvements in Physical Science [Jan- 
affinity similar to that pointed out by Mr. Phillips, and mentioned 
in a preceding part of this paper. I conceive that hydrogen has a 
greater affinity for oxygen than iron has, and that the decomposition 
af water by iron takes place only in a very limited degree. 
9. Oxides of Manganese.—Berzelius in his first papers on the 
atomic theory admitted five oxides of this metal ; but in his essay 
On the Cause of Chemical Proportions he reduces the number to 
four. (Annals, iii. 359.) Gay-Lussac admits only three ; namely, 
the oxide obtained by dissolving manganese in acids, the peroxide 
which is found native, and the deutoxide formed by exposing the 
peroxide to a red heat. (Ann. de Chim. et Phys. i. 38.) I think 
it more probable that only two oxides of manganese exist ; namely, 
the protoxide which dissolves in acids and forms salts, and the per- 
oxide which exists native in such abundance. The brown oxide is 
more probably a compound of the protoxide and peroxide. It is 
incapable of dissolving in acids; and, when treated with acids, is 
resolved into protoxide and peroxide. The constituents of these 
oxides have not yet been determined in a satisfactory manner. Pro- 
bably the protoxide is a compound of 100 metal + 20 oxygen, and 
the peroxide of 100 manganese + 40 oxygen. ‘This would make 
the weight of an atom of manganese 5. , 
10. Oxides of Tin.—Berzelius has concluded from his experi- 
ments on tin that this metal combines with three doses of oxygen, 
and forms three oxides, all capable of uniting with acids. But 
Gay-Lussac has shown that we have no proofs that any difference 
exists between the deutoxide and peroxide. (Ann. de Chim. et 
Phys, i. 40.) The conclusion that only two oxides of tin exist 
removes an anomaly from the atomic doctrine, and is, therefore, 
preferable. Protoxide of tin is composed of 
i Aci ak aka kee vetoes a, LOO. is cea ok ROE 
eres) es ace sb eas LOO t ae~ D 
and peroxide of 
Mts VECO aE ves gels 5 LOO” th. 45 Jina 
Oxyeen so... aati ules eb te | 
This gives us the weight of an atom of tin 7:352. If we con- 
sider it as 7°375, it will then be a multiple of the weight of an atom 
of hydrogen, and the difference will be too small to be detected by 
experiment. The protoxide would be a compound of 100 tin + 
13°56 oxygen, and the peroxide of 100 tin + 27°12 oxygen. 
11. Tantalum.—Berzelius has lately subjected this metal to a 
much more complete examination than either Hatchett or Ekeberg 
had subjected it to. J have given a sketch of the results which he 
obtained in a late number of the Annals, viii. 233. Its colour is 
nearly similar to that of iron. Its specific gravity, when in a cohe- 
rent mass, but not melted, is 5°6]. It is brittle. It is not acted 
upon by any acid hitherto tried. When heated, it burns feebly, 
and is converted into a greyish-white oxide. It detonates feebly 
with nitre, and is converted inta a snow-white oxide, which is com- 
