JIK. 0. A. DKUDY ON NEniELINE-ROCKS IN BRAZIL. 257 



gnoiss*. This is the more rcmnrkiible as the foyaite resists decay much 

 better tliaii the otlier rocks of the reirion, so that its al)seiice cannot 

 he explained by decomposition t. The centi'al eruptive mass seems 

 to be destitute of dyke-like apophyses, or these are represented by 

 phonolite or basic rocks, and not by foyaite. The nearest approach 

 to a dyke that has been seen is in an immeni^e block below the lower 

 tSao-Pedro dam, that stands between two blocks of gneiss, each of 

 the three being the size of a small cottage, in sucli a way as to 

 suggest a dyke decomposed along the contact. This exposure, and 

 the patch of foyaite and phonolite blocks to which it belongs, are 

 interesting from the lack of surface-connexion with the peak. 8uch 

 a connexion, if it exists, should appear along the road up the iSao- 

 Pedro valley or the high-level ditch, which, however, shows nothing 

 but a few insignificant dykes of phonolite. This patch apparently 

 represents a small iudcpendent (as regards the surface) centre of 

 eruption. 



In following np the stream-beds the same fact is noticed. These 

 mountain-torrents, so choked as to bo almost impassable with 

 immense loose blocks, among which foj'aite predominates over all 

 other rock-species, nowhere, so far as examined, show this rock in 

 situ. Their beds, even high up on the flanks of the mountain, are 

 in gneiss or, as in the case of the upper section of the liio do Ouro 

 at the point visited, in effusive rocks, either phonolite or consoli- 

 dated tufaceous material. The impression that one gets is that the 

 foyaite blocks come from a bedded mass, broken down by under- 

 mining, rather than from a dyke or stock. 



The long narrow ridge between the Eio do Ouro and the Santo 

 Antonio confirms this impression. The surface everywhere bristles 

 with projecting masses of foyaite. The rivers on either side flow 

 over gneiss beds, and are choked with loose masses of foyaite. Here, 

 if anywhere, an immense dyke, or stock, of the same rock should 

 be met with in a cross-section of the ridge. Two such sections are 

 ])resented. One is by a road passing, with a considerable cutting, 

 through a gap near the end of the ridge. This shows gneiss 



* On the main tram-line these blocks are only met with on the last two or 

 tliree kilometres near the Rio do Ouro reservoir. None occur on the Concei9ao 

 branch, nor ou that of the Sao Pedro, except at the end near the dam. The 

 r.)ad from the lower to the upper J?a,o-Pedro dam is free from tbem below the 

 mouth of the Barra. From here along the forest-ruad up the Sao-Pedro 

 valley, they occur at frequent intervals in large patches. The paved road over 

 tiie ridge from Concei^ao has none from the summit down to the crossing of 

 the Erava. The high-level Sao Pedro ditch has them above the mouth of the 

 Earra and in the last two kilometres near the Santo Antonio, but for the rest 

 of the distance (12 kilometres of continuous rock-exposure) not so much as a 

 pebble was seen. 



t The blocks in general have a decomposition crust of only 1-2 centimetres 

 at the most. Special search was made for totally decomposed masses, for thd 

 ])ur])Ose of obtaining rare accessory minerals by washing. Only a few such 

 blocks, and these comparatively suuill, could be found. These were so tough 

 and spongy as to retain their form, and require crushing with a hammer, 

 whereas gneiss and granite in the immediate vicinity were so completely trans- 

 formed into an earthy mass as to be readily reduced to mud and sand by the 

 pressure of the lingers under water. 



