of the Oxides of Manganese. 27 



of manganese. This result, in the accuracy of which M. Arf- 

 wedson does not place implicit confidence, would fix the equi- 

 valent of manganese at 28*61. He prepared the chloride by 

 placing the carbonate of manganese in a spherical cavity blown 

 in a barometer tube, transmitted over it a current of muriatic 

 acid gas, and heated the carbonate by means of a spirit-lamp 

 as soon as the atmospheric air was expelled from the tube. 

 As it is difficult by this, as well as by Dr. Davy's process, to 

 procure a perfectly pure chloride of manganese, I had recourse 

 to the following method. A solution of the muriate of man- 

 ganese was evaporated to dryness, the heat being carefully rer 

 gulated so as not to decompose any of the salt, and the dry 

 compound was placed in a spherical cavity in the middle of a 

 barometer tube about six inches long. Muriatic acid gas was 

 then transmitted through the tube, and heat applied by the 

 flame of a spirit-lamp. The chloride entered into perfect fu- 

 sion at a low red heat, and on cooling yielded a highly cry- 

 stalline lamellated mass of a beautiful pink colour. Every 

 trace of acid and moisture was expelled by heat ; and while 

 the tube was still hot, its extremities were closed by corks, so 

 that the chloride might be weighed without attracting mois- 

 ture from the air. In the sense above explained it was quite 

 neutral. Of this chloride 12*47 grains were dissolved in di- 

 stilled water, and formed a colourless solution without any 

 residue. The muriatic acid was thrown down by the nitrate 

 of silver, and yielded 28*4-2 grains of the fused chloride of sil- 

 ver, equivalent to 7*008 grains of chlorine. Consequently the 

 chloride of manganese consists of 



Manganese 5*462 28*06 



Chlorine 7*008 36 



It follows from the preceding researches, that 28 is the true 

 atomic weight of metallic manganese, and 36 the equivalent of 

 that oxide of manganese which forms definite compounds with 

 acids, and which I regard as the real protoxide of the metal. 

 It is consequently composed of 28 parts of manganese and 

 8 parts of oxygen. These numbers agree with the atomic 

 weight of manganese as stated by Dr. Thomson, but not with 

 that given by Berzelius, who fixes it at 28*463. This estimate 

 is made from an analysis of M. Arfwedson, who finds that the 

 deutoxide of manganese is composed of 1 00 parts of the metal 

 and 42*16 parts of oxygen ; but it will appear from the sequel 

 of this paper that the real quantity of oxygen united with 100 

 parts of manganese to constitute the deutoxide is 42*857 and 

 not 42*16 as Arfwedson supposes. 



On the Protoxide of Manganese. — By this term I mean the 

 salifiable base of manganese, the only oxide of the metal which 



E 2 appears 



