314 0. A. Derby — Magnetite Ore Districts in Brazil. 



situ intercalated in schist in which the pyroxene predomin- 

 ates over the magnetite. 



At other points flaggy layers were found in which a white 

 element is equally abundant with the pyroxene. This in well 

 preserved specimens proves to be nepheline and the rock stand- 

 ing alone might be called a schistose nephelinite. In this type 

 the iron minerals are rare, while perofskite, apatite, and pri- 

 mary biotite are constant, olivine a frequent accessory. 



The peculiar rock above described might very properly be 

 called a pyroxenite if it were not for the objection that this 

 would add still another to the very heterogeneous (as regards 

 geological relations) reunion of types to which that name has 

 been applied.* In view of the evident necessity of subdividing 

 according to geological origin, the so-called pyroxenites, it 

 seems necessary to distinguish this by a distinct name, and it 

 will here be called jacujnrangite. The rocks included under 

 this title are allied to the nepheline-bearing series and present 

 the various types of pure magnetite, magnetite with accessory 

 pyroxene, pyroxene with accessory magnetite, and pyroxene 

 and nepheline with biotite and olivine as accessory or (in the 

 case of the former at least) essential elements. All these types 

 are most intimately associated as parts of the same mass and 

 the gradual passage from one to the other has been most satis- 

 factorily proven. The most constant and characteristic ele- 

 ment is a violet titaniferous pyroxene. Another type of 

 pyroxene-nepheline rock (nephelinite) characterized by green 

 pyroxene and abundant biotite occurs in the district, but has 

 only been seen rarely in loose masses and is presumed to come 

 from dykes and to have an intimate relation to the jacupirangite. 

 At Ipanema a rock of similar mineralogical composition cuts 

 the Carboniferous strata. 



• The relations of the jacupirangite to the other rocks of the 

 district are not clear. The nepheline-bearing varieties are 

 suggestive of an eruptive origin and the phenomena observed 

 about a long narrow ridge of limestone of presumed Cambrian 

 age, that rises near the center of the jacupirangite area, appear 

 to confirm this suggestion, although no actual contact could be 

 observed. The limestone is a white, coarsely crystalline mar- 

 ble, heavily charged in places with crystals of magnetite and 



The mica shows the optical properties of biotite but chemically (at least in the 

 decomposed state in which it could be obtained for analysis), it is essentially a 

 hydrated silicate of iron and alumina. 



* In Brazil alone three or four distinct groups of pyroxenite have been recog- 

 nized. Ai>ide from the one here described, there is one associated with the lime- 

 stone of the Cambrian (?) series and undoubtedly derived from it through meta- 

 morphism. Another is similarly associated with limestones of the gneiss series 

 and still another forms dykes in gneiss. The calcareous groups are rich in scapo- 

 lite and apparently correspond with a part of the pyroxenites of Barrois and 

 Lacroix, but none agree with the original type as established by Dana for rocks 

 on the Hudson, whose genetic relations seem to be with the peridotites or norites. 



