284 J. C. BRANNER — DECOMPOSITION OP ROCKS IN BRAZIL. 



expansion, and even the same mineral expands differently along its dif- 

 ferent axes (excepting in the isometric system, to which none of these 

 minerals belong.) f In the case of quartz, for example, it is known that 

 the anojles of its faces change at different temperatures. All this causes 

 a slipping back and forth of the minerals upon each other and the open- 

 ing of minute fractures that mark the beginning of decay. 



The mineral constituents of the gneiss from Rio de Janeiro are quartz, 

 orthoclase, plagioclase, biotite (little), microcline, apatite (very little), 

 sillimanite, and rutile needles (very few). 



The breaking up of the rock is also aided to a certain extent by the 

 expansion of minerals upon decomposition and hydration. Of the min- 

 erals in the list it has been shown that biotite becomes enlarged upon 

 h3alration ; J for the other minerals it has not yet been demonstrated, so 

 far as I have been able to learn. 



Comte de la Hure has stated that the Brazilian diorites increase their 

 volume by at least one tenth, and he has expressed the opinion that the 

 hills of Brazil have increased their height by sucli liydration.§ Merrill 

 has also expressed the opinion || that hydration causes granite to expand. 

 In neither instance, however, can it be said that the case is demonstrated. 



Range of temperature. — L^dng mostly within the tropics, the climate of 

 Brazil is necessarily a hot one. The temperature seldom falls below the 

 freezing })oint even in the high valleys of the Minas watershed. The 

 high angle of the noon sun, it being almost directly overhead the year 

 round, affects the rocks directly. 



It is to be regretted that so few systematic meteorologic observations 

 have been n:iade in Brazil ; but even among the few recorded there are 

 fewer still upon temperatures in the direct rays of the sun, and it is with 

 these temperatures we have to deal — the maximum and minimum to 

 which rocks are subjected. I am of the opinion that while many changes 

 of temperature will do more than a few, the chief work is accomplished 

 by any change or by any number of changes capable of oi3ening cracks 

 in the rock. We find the rocks decomposing always along joints or frac- 



3d ser., vol. xxi, p. 428), by Hawes in the mica-schist of mount Washington (New Hampshire 

 Snrv., vol. iii, pt. 3, p. 2l<)), and by G. H. Williams as a contact mineral in the Cortland series (Amer. 

 Jour. Soi., 3d ser., vol. xxxvi, p. 251), in which relations it is also known abroad. It is also recorded 

 by Zirkel from the Humboldt range, Nevada, in mica-schist (40th Parallel Survey, vol. vi, p. 10). 

 Kalkowsky mentions it in the Saxon granulites, in which occurrence and in the Manhattan island 

 locality it probably more nearly resembles the specimens in hand than in the others mentioned. 



I am indebted to Professor J. F. Kemp, of Columbia College, New York, for the above. There is 

 also a general description of Rio gneiss by T. Sterry Hunt in Hartt's Geology and Physical Geog- 

 raphy of Brazil, page 550. 



f The coefficients of expansion for these minerals do not appear to have been determined. 

 JQuar. Jour. Geol. Soc, Feb. 20. 1889; Geol. Mag., April, 1889, vol. vi, p. 187. 

 § Revist do Inst. Hist, do Brazil, 1866, xxix, pt. 2, pp. 422-429. 



II Bui. Geol. Soc. Am., vol. 6, p. 332. 



