2.38 MR. 0. A. DERBY ON XEPHELINE-ROCKS IN BRAZTL. 



decomposed in situ under a soil-cap with loose blocks of foyaite. 

 The other section ia a tunnel, 300 to 400 metres long, for the passage 

 of water to the storage reservoir on the liio do Ouro. It is cut 

 exclusively in gneiss, with a few insignificant dykes of basic rocks. 

 A footpath over the top shows gneiss in situ for a few metres on 

 the slopes above each tunnel-opening, and blocks of foyaite rising 

 abundantly' from the soil for the rest of the distance. It is clear 

 that the foyaite, instead of piercing the gneiss, here rests as a sheet 

 upon its surface. As the level of the gneiss rises rapidly in the 

 stream-beds and lateral spurs of the mountain, the sheet of foyaite 

 must have occupied an inclined position, sloping from the higher 

 parts of the mountain to the lowlands at the base. 



Outside the stream-beds and artificial cuttings, which only give 

 access to the marginal zone of the eruptive mass, it is difficult, 

 owing to the almost unbroken soil-cap covered with dense forest, 

 to determine the true character of the exposures. The fragmentar}' 

 condition of the foyaite is general, even over the surface of the 

 summit of the peak, where the underlying rock is unquestionably 

 ill sltH, though whether as a boss, dyke, or sheet could not be deter- 

 mined. On the flanks of the mountain over the elevated portions 

 outside the stream-beds, the general surface-character of the ex- 

 posures is that of boulder-trains, which might result from the 

 broken-up outcrop of either a dyke or a sheet. The hypothesis of 

 a continuous boss is excluded by the lack of continuity of the ex- 

 posures in a horizontal plane. Something like half of the outer 

 face of the San to- Antonio peak has been traversed at about a 

 quarter of its elevation. Occasional boulder-patches were met with, 

 but for the most part the trail was over a reddish argillaceous soil, 

 which must have come from the decomposition of some rock other 

 than gneiss or foyaite. The former is excluded by the absence of 

 quartz grains; the latter, by its mode of decomposition and its 

 appearance at intervals as boulders. The greater part of the mass 

 of the peak is here evidently formed of some easily-decomposed 

 eruptive rock traversed by dykes or sheets of foyaite. The appear- 

 ance, in the same section, of the effusive type of phonolite and tuff, 

 so subject to decay that even in the bed of a mountain-torrent it is 

 not well preserved, affords a clue to the probable nature of this 

 rock, and at the same time suggests the hypothesis that, contrary 

 to appearances, the foyaite plays a subordinate part in the com- 

 position of the Tingua eruptive mass, considered as a whole. 



The evidence above presented of the occurrence of a large portion 

 of the Tingua foyaites in sheet-like masses warrants a comparison 

 with the PoQos de Caldas locality, where this mode of occurrence is 

 seen in great perfection in the foyaite mass cut by the tunnel, as 

 described in a former paper *. This mass forms an irregular sheet, 

 some 10 to 20 metres in thickness, inclined at an angle of 15°- 

 20^, and cutting through a rock which is in part a volcanic 

 conglomerate, in part apparently an imperfectly-individualized 

 phonolite. It may be considered as lava, which, instead of being 



* Quart. Jouru. Geol. Soc. vol. xhii. (1887) p. 466. 



