of Minos Geraes, Brazil. 25 



In the specimens obtained from this tunnel the altered 

 patches present the earthy pitted appearance of the above 

 described ore from Sao Goncalo and like that give, on treat- 

 ment with acid, a residue of microscopic garnet, skeletons of 

 secondary silica in the form of garnet and a moderate amount 

 of exceedingly tine black sand giving a strong titanium reaction 

 and that is apparently ilmenite with a slight admixture of trans- 

 parent red grains without defined form that appear to be rutile. 

 The perfectly sound rock has a hardness between 6 and 7 and 

 a steely luster, but under the lens is seen to be minutely mot- 

 tled with white points corresponding to the pitting of the 

 decomposed portions. Occasional patches of oriented sheen 

 appear to be cleavage surfaces of large prismatic crystals. 

 Treated with hydrochloric acid, even in considerable lumps, 

 the black portion of the rock is readily and completely dis- 

 solved, leaving a considerable residue of white garnet with a 

 slight amount of ilocculent silica, ilmenite and rutile (?). Under 

 the microscope the rock is seen to consist of minute isolated 

 grains of garnet embedded in a much more abundant ground- 

 mass of manganese oxide of uniform steely luster, except in a 

 narrow zone around each garnet grain where it is coal black 

 and apparently softer. This black zone apparently comes from 

 a beginning of alteration in the garnet, though it may also be 

 due to an incipient hydratation of the manganese oxide of the 

 groundmass. In any case, the rock is essentially a mixture 

 of a hard prismatic, cleavable manganese oxide with a manga- 

 nese garnet, both being primary elements. An analysis kindly 

 made by Dr. Florence gave : 



Residue insoluble in hvdrochloric acid 20-78 



Si0 2 1 0-77 



Fe 2 3 5-79 



NiO + CoO 0-30 



MnO . 8-54 



Mn0 2 57-38 



P 2 5 0-08 



CaO _ 0-49 



MgO 0-20 



H 2 391 



98-24 



The deficiency in this analysis is perhaps due in part to the 

 presence of alkalies that were not determined. The manga- 

 nese oxides calculated as metal and oxygen and reduced to 100 

 give : l[n. 65*06 per cent, O 34/94: per cent, which corresponds 

 quite closely with the composition of polianite (Mn. 6T1 per 

 cent, O 36*9 per cent), and in view of the physical properties, 



