Vol. 62.] CATAEACIS OF THE RIVER MADEIRA, ETC. 107 



at a period long anterior to the crystallization of the groundmass. 



But, during each period of crystallization, the quartz seems to have 



been one of the earliest minerals to crystallize out. 



This rock was analysed by Mr. George S. Blake, of the Scientific 



Department of the Imperial Institute, with the following result. 



The second column shows the proportions of the molecules of the 



different constituents, obtained by dividing the percentage by the 



molecular weight. 



Pe rce n tage Molecula r 



composition . proportions. 



Si0 69-58 1-1597 



A1 2 3 1372 -1345 



Fe„0 3 1-95 -0122 



FeO.. 2-30 '0319 



MnO 0-09 -0013 



MgO 0-34 -0085 



BaO 0-12 -0008 



CaO 1-54 -0275 



K 2 519 -0552 



Na 2 357 "0576 



S .: 019 -0059 



F 0-07 -0026 



H.,0 above 100° C. ... 078 -0433 



H.;Oatl00°C 0-52 



Less O replaced by F. . . — 003 



99-93 



From a consideration of the porphyritic minerals alone the 

 rock might be classed as a quartz-hornblende-porphyrite (diorite- 

 porphyry); but the orthoclase in the matrix and the alkalies shown 

 by the analysis prove it to be more nearly allied to quartz-felsite 

 (granite-porphyry). It probably forms part of a large 

 intrusion. Rocks of similar composition and granitic structure have 

 been described as granite, segirine-granite, and granitite. 



The rock from the Salto do Girao is rather poorer in alumina 

 than most rocks with the same percentage of silica. It falls within 

 the group (subrang) toscanose of the American quantitative 

 classification. 1 



Some 3 miles below the Salto do Girao is another cataract, known 

 as the Caldeirao do Inferno or Cauldron of Hell. Here the 

 river, instead of narrowing, spreads out into numerous branches 

 which flow in rocky channels. The prevailing rock is similar at 

 first sight to that which is exposed at the Salto do Girao. It is 

 mainly composed of bright-red felspar, with areas of grey rounded 

 quartz, and here and there specks of a dark-green mineral. 



Under the microscope (M 8) this rock appears to be fairly coarse, 

 crystals of 1*5 millimetres in diameter being frequent. There is 



1 Whitman Cross, J. P. ladings, L. V. Pirsson, & H. S. Washington, 

 'Quantitative Classification of Igneous Eocks' Chicago, 1903. This classi- 

 fication is so widely used in America that it cannot be ignored, but there 

 appear to be grave reasons against its general adoption. 



