282 J. C. BRANNER — DECOMPOSITION OF ROCKS IN BRAZIL. 



weakness along which the agents of disintegration and decomposition 

 penetrate. If the massive rock be jointed, boulders of decomposition 

 may still be formed, but the proximity of the joints determine the sizes 

 of the boulders produced by exfoliation. 



The process of mechanical breaking up of rocks in Brazil is different 

 from that in operation in cold climates. 



In Brazil frost, of course, has no part in the disintegration of the rocks. 

 It is generally supposed that the action of frost greatly facilitates rock 

 disintegration. While this cannot be questioned with regard to steep- 

 faced declivities and hills with high slopes and with regard to the break- 

 ing up of the upper three or four feet of the ground, the action of frost 

 through our long northern winters is protective rather than otherwise, 

 preventing the penetration of water and of the dissolving agents it usually 

 carries. Snow is especially protective on account of its preventing radi- 

 ation, or, as Rozet says, it " acts as a screen between the soil and space."* 

 Frost, moreover, is active only where there is freezing and thawing, not 

 where there is one freeze in the year, as there is in high latitudes. On 

 the other hand, the change of temperature in the tropics is, in the main, 

 a diurnal variation, while in the temperate regions the marked change is 

 an annual one. The diurnal changes of the tropics necessarily take place 

 within narrow limits, while the annual changes in the temperate zones 

 are between extremes that are wide apart. 



The tendency of the tropical temperature changes is to affect the rocks 

 to shallow depths and very rapidly, while the slow annual changes affect 

 them more profoundly but much less rapidly ; but in the tropics the sun 

 at noon never sinks low on the horizon and the range of temperature is 

 necessarily smaller than that of the temperate regions. 



The total annual range of temperature for rocks in the northern part 

 of the United States is from 150 to 170 degrees Fahrenheit, while the 

 daily range in Brazil is only about 100 degrees. This com})aratively 

 small range of temperature in Brazil at first glance suggests inefficiency, 

 but the rapid radiation and humidit}'' of that climate a^re important 

 elements in its effectiveness. 



Changes of temperature attack the rocks in two wa^^s : first, by the dis- 

 integration of the mass ; and, second, by causing exfoliation. 



Temperature and disintegration. — Theoretically, incipient disintegration, 

 in so far as it is caused by changes of temperature alone, penetrates as 

 far as do these changes. So long as the rock is homogeneous disinte- 

 gration caused in this way must penetrate the rocks in parallel planes 

 corresponding to isothermal planes. These isothermal planes are only 

 rudely parallel to the rock surfaces, both in the case of large mountain 



* Comptes Rendus, 40, 1855, p. 299. 



