254 MR. 0. A. DERBY ON KErHELINE-ROCKS IN BRAZIL. 



tram- and pipe-liiics afford excellent opportunities for the study of 

 tlie lower marginal zone. Still more important is a high-level 

 ditch, cut for a provisional supply at a time of water famine, near 

 the GOO-metre level, for taking the waters of the Sfio Pedro over 

 the top of the ridge. This gives an almost ahsolutely continuous 

 section of rock in situ (in great part decomposed) for a distance of 

 ](3 kilometres. Unfortunately, however, it barely touches the 

 eruptive mass of the peak proper, at two points of no special 

 interest. A paved road from Conceigao across the ridge to the east 

 of the ])eak also affords excellent sections. Outside these lines every- 

 thing is covered with heavy forest and dense jungle, and explora- 

 tion is difficult and unsatisfactory. Trails have been cut from the 

 summit-level of the paved road to the highest point of the peak, 

 and from the llio do Ouro reservoir around the western flank of tlie 

 peak to the upper (provisional) dam of the Sao Pedro, and along the 

 side of the llio do Ouro valley to a point well within the central 

 crater-like depression. The stream-beds have also been followed 

 for a certain distance ; but it was found that in the most interest- 

 ing portions they are so obstructed by falls and loose boulders th;it 

 one is obliged to take to the woods, and thus lose all opportunity 

 for geological observation. 



The fundamental rock of the region is a biotite-gneiss, generally 

 coarse-grained and porphyritic, like the characteristic variety of the 

 mountains about Eio de Janeiro. This is cut by numerous dykes of 

 biotite-granite (granitite) and diabase, such as are common in all 

 the Brazilian gneiss-regions that have been examined, and which, 

 being clearly anterior to the eruptions that produced the peak, need 

 not be more fully considered here. Very small dykes of muscovite- 

 granite, decomposed to kaolin, also occur rarely. 



In the peak proper, the predominant type is the orthoclase-nephe- 

 line combination, either holocrystalline, as foyaite*. or porphyritic, 



limited number of aneroid observations. The heavy black of the pipe-lines and 

 ditches, and the light-dotted trace of the roads and trail.-^. show the lines that 

 have been examined ; while the heavy-dotted trace shows tJie approximate oit- 

 line of the area occupied by the eruptive rocks and of the peak proper. Outside 

 this area, the contoured portions are gneiss, and the light parts mainly alluvial 

 flats, with detached outliers of gneiss and granite. 



* To avoid the cumbersome and misleading designations derived from 

 syenite, it seems convenient to employ some one of the numerous simple and 

 non-committal names that have been applied to this grou]:) of rocks as a 

 general name for the whole. Of these, yb//«?Yc, as defined by Prof. Rosenbusch 

 (the hornblende-, augite-, or regerine-bearing members of tlie group), applies 

 best to the Brazilian types, and seems best adapted as a general denomination. 

 Aside from this, there is a certain propriety in retaining a name of Portuguese 

 origin for a group best known through occurrences in Portuguese-speaking 

 countries — Portugal, Cape-Yerde Islands, and Brazil. In his recent paper in 

 Tschermak's Mittheihmgen (vol. xi. p. l(iO), Prof. Rosenbusch appears to 

 sanction this usage in the term Foyaitmagma. The term Thu/uaite, proposed 

 by Prof, Rosenbusch for the phonolitic types of the Brazilian orthoclase-nephe- 

 line rocks, is not here adopted, as it seems to be based on a misapprehension, 

 for vhich the writer is responsible, since, in the collection sent to lleidelbeig, 

 it happened that only dyke-phonolites were represented, thus creating the 

 impres.sion that the typical effusive forms did not occur in the Brazilian 

 localities. 



