Vol. 62.] CATARACTS OE THE RIVER MADEIRA, ETC. 97 



which are probably filled with liquid. Quartz occurs in small blebs, 

 and shows rounded idiomorphic boundaries. There is abundance 

 of brown prismatic hornblende, with a maximum extinction- 

 angle between the direction of vibration of the slower wave-surface 

 and the trace of the cleavage of about 24°, and strong pleochroism. 

 It contains inclusions of quartz. 1 Crystals of pyroxene are also 

 present ; and much of it is distinctly pleochroic, ranging between 

 very pale shades of green and yellowish-brown. This pyroxene 

 appears to be partly augite and partly enstatite. Some of the 

 augite shows signs of alteration into diallage. Irregular masses 

 of magnetite also occur. The rock, which has a specific gravity 

 of 3*012, is apparently on the border-line between a micro - 

 diorite and a fine-grained dolerite. 



In the neighbourhood of the cataract, pebbles again appear in 

 the river-beaches. Some are formed of crystalline rocks ; but the 

 majority are derived from sedimentary beds, and include con- 

 solidated grits or quartzites and silicified limestone or chert, 

 showing traces of stratification. The latter contain organic remains 

 with concretions and rhombohedra of a ferruginous carbonate, 

 perhaps ankerite, which has been oxidized to a yellow-brown tint ; 

 and a groundmass of colloid and chalcedonic silica. 



I submitted the microscope-sections, that I had had made, to 

 Dr. George J. Hinde, F.R.S. He kindly examined them, and wrote 

 to me as follows : — 



' The circular and elliptical sections in slide (M 1 e) are sponge-spicules. 

 Probably they are the anchoring spicules of siliceous Hexactinellids. The 

 inner tube is the axial canal of the spicule. They seem to be fairly common in 

 this chert. There are traces of rods in slide (M 1 c) ; these are probably also 

 spicules, but the chert is so much altered in this slide, that one cannot be 

 positive. 



'There are other bodies in slide (M 1 e) the nature of which is unknown to 

 me. One is a thin shell, almond-shaped in section, which vas^y be the carapace 

 of some crustacean ; a smaller form may be of similar origin. 



'There is also an imperfect flask-shaped body with a reticulate structure, of 

 which I cannot guess the character. 



: I do not see anything that can be considered radiolarian ; but the chert 

 has been very much altered and, judging from the changes which have affected 

 the sponge-spicules, the radiolaria, if originally present, would have been 

 obliterated beyond recognition.' 



In a subsequent letter Dr. Hinde states that he considers the 

 chert to be marine, and that it may well be Palaeozoic, although the 

 evidence is insufficient to prove it. Rocks containing Hexactinellid 

 sponge-spicules occur to the north of the Amazon on the rivers 

 Trombetas and Maecuru. The accompanying fossils indicate a 

 horizon at the base of the Silurian (using that expression so as to 

 exclude the Ordovician). 2 



The sedimentary rocks cannot have come from the Andes, the 



1 See p. 121, footnote 1. 



2 F. Katzer, ' Grundzuge der Geologie des Unteren Amazonasgebietes ' 1903. 

 pp. 218, 222. 



Q.J.G. S. No. 2±o. k 



