EXFOLIATION. 291 



obliquity, and they therefore attain a higher temperature and might be 

 expected to be modified accordingly. As a matter of fact, the angles of 

 slope of these hills vary greatly and seem to be determined by structural 

 differences, from which it is inferred that structure has more to do with 

 their ultimate topographic forms than has variations of temperature. 



It is not to be inferred, however, that the main features of exfoliation 

 are determined by the banding of gneiss, for the sheets come away for 

 the most part regardless of this banding.* 



The spheroidal weathering of igneous rocks f is supposed to be due to 

 imperceptible lines of fractures caused by the cooling of the mass and 

 developed by weathering. It should be noted here that there is no reason 

 for supposing that either the mountain masses or the exfoliated boulders 

 of Brazil owe their origin to such action, or that they call for any such 

 explanation. 



The exfoliated rocks discussed by Scrope, Mallet and Bonney are igneo- 

 volcanic, while the Brazilian rocks are mostly plutonic. The boulders 

 flake off in the same way, whether the exfoliating masses are broken out 

 naturally or artificially, while there is certainly no reason for supposing 

 that the mountains are cones formed by the cooling of the original mass. 

 The rounding of the hills takes place in precisely the same way as the 

 rounding of any angular fragment of homogeneous rock when its temper- 

 ature is alternately and quickly raised and lowered. 



Professor Shaler has called attention X to the great annual change of 

 temperature suffered by exposed rock surfaces in the northern United 

 States at a variation of about 150 degrees ( — 30° to 125°+) in half a 

 year, and a diurnal variation often amounting to half that. The annual 

 changes suffered by exposed rock surfaces in Brazil is not so great as 

 this, but, as shown above, it is enough to open the rocks for the penetra- 

 tion of disintegrating agents, which are more active there than in tem- 

 perate regions. The fact that the changes in temperate regions are lower 

 in the scale is of no importance so long as they do not fall below the 

 freezing point. 



Professor Shaler points out in the paper just cited that concentric frac- 

 ture is the result of changes of temperature, and that it must produce 

 dome-like topographic forms, and that hitherto covered blocks which 

 did not exfoliate before exposure do so as soon as exposed. 



Finally it may be said of exfoliation that, while changes of tempera- 

 ture tend to cause it, exfoliation is not produced by changes of tempera- 

 ture alone, and decomposition much less so. I quite agree with the 



* See also geology of New England. T. Sterry Hunt. Amer. Jour. Sci., vol. 1, 1870, p. 8. 

 t Columnar, fissile and spheroidal structure. T. G. Bonney. Quar. Jour. Geol. Soc, vol. xxxii, 

 1876, pp. 140-154. Geikie's Text-book of Geology, 3d ed., p. 348. 

 t Proe. Boston Soc. Nat. Hist., vol. xii, 18C9, pp. 289-293. 



