of 31 in as Geraes, Brazil. 31 



if not identical type. In the Agua Limpa schist, moreover, 

 the manganese-bearing element is spessartine as in the ore 

 bodies, thus giving greater plausibility to the hypothesis that 

 the relation between these last and the above mentioned rocks 

 may be a genetic one. If thus related, the ore bodies present 

 strong analogies with those of magnetic, titaniferous and 

 chromic iron ores that are now generally considered as mag- 

 matic segregations in various types of eruptive, and, all things 

 considered, this hypothesis seems the most plausible one for 

 the manganese ores here discussed. The fact that though 

 carefully looked for, no rock types that could be positively, or 

 even presumably, identified as clastic could be found near the 

 ore bodies, makes the alternative hypothesis of a clastic origin 

 a difficult one to apply. The occurrence of distinct traces of 

 nickel and cobalt in the unaltered ore of the Morro da Mina 

 tunnel may perhaps be also an argument for an eruptive origin, 

 though too much stress cannot, without farther study, be laid 

 upon it, since Mr. Scott reports these elements under totally 

 different conditions near Miguel Eurnier in a small patch of 

 secondary manganese oxide that is undoubtedly a secretionary 

 deposit in a decomposed clastic schist. 



Through the kindness of Dr. Francisco da Paula Oliveira, 

 director of the geological section in the National Museum of 

 Rio de Janeiro, I have had the opportunity of examining a 

 specimen from a manganese ore body enclosed in granite near 

 Queimados, in the interior of the state of Bahia, that represents 

 another interesting phase of this rock type. The rock is per- 

 fectly fresh, showing about equal proportions of large-sized 

 garnets and of pyroxene with a diallage-like cleavage. The 

 garnets, which attain a size of five millimeters or more, are of a 

 light yellow color, becoming perfectly white in thin sections, 

 and give a strong manganese reaction. The pyroxene, when 

 free from staining by manganese oxide, is colorless and glassy 

 and gives an extremely abundant reaction for manganese, which 

 is evidently much more abundant than iron. Aside from a 

 colorless glassy amphibole intergrown with the pyroxene, no 

 other constituents could be recognized and the rock is essen- 

 tially composed of manganese-garnet and manganese-pyroxene. 

 From such a rock through the replacement of the easily decom- 

 posable pyroxene by secondary quartz the type of quartz-garnet 

 rock of Piquiry and Morro da Mina might be produced.* With 



* This type might, however, be expected to leave some trace of its lime and mag- 

 nesia in the form of secondary asbestus, and in this case the most probable repre- 

 sentative is the quartz-garnet rock with asbestus, while the similar rock without 

 the latter mineral may be suspected to have come from an original type in which 

 the bisilicate element was perhaps rhodonite, a mineral that has been looked for 

 in vain although it seems natural that it should occur in such an association of 

 manganese-bearing rocks. 



