AREAS STUDIED AND THEIR CHARACTERISTICS. 257 



to the southernmost part of Rio Grande do Sul. Neither is decay con- 

 fined to the immediate surface, but it penetrates the solid rocks as far as 

 they are affected by varying temperatures or by crevices, however obscure, 

 along which water can enter. The rocks necessarily vary more or less 

 in their resisting powers, but they are all more or less affected. 

 Rocks are attacked in three ways : 



First, by surface disintegration ; 



Second, by exfoliation ; 



Third, by profound decay in place. 



Areas studied and their characteristics.- — The most striking instances of 

 rock decay which have fallen under my own observations have been in 

 the vicinity of Rio de Janeiro and in the states of Minas Geraes, Per- 

 nambuco and Para. 



About Rio de Janeiro this decay is to be seen on 6very hand, and has 

 been recorded by almost every geologist who has visited that region. 

 The road from the city to Tijuca has many cuts in the soft decayed 

 gneiss ; the road which ascends toward the Tijuca peak and that leading 

 from near Tijuca toward Pedra Bonita and the Chinese view reveal the 

 decayed rock in almost every cut. West of the Botanical gardens the 

 road leading toward the Gavea expose in several places cuts 20 and 30 

 feet deep in the decomposed rock. 



In the Larangeiras suburb many deep cuts have been made in this 

 material, especially close to the foot of the hill on either side of the valley 

 toward its upper end. At one place in the upper part of the Larangeiras 

 a tunnel more than 100 feet long driven into the foot of the mountain 

 for the purpose of making a cooling chamber for domestic purposes pene- 

 trates only softened gneiss. A tunnel cut through the hill in 1887 to 

 connect Larangeiras and Rio Comprido passed through more than 100 

 feet of decayed rock on the Rio Comprido side. 



In 1879-'80 a large reservoir was built on the Morro do Pedregulho 

 near the Ponta do Cajti in the northwestern outskirts of the city of Rio. 

 The hill was originally 225 feet high, and the site for the reservoir was 

 prepared by cutting off the top of the hill to a depth of 65 feet.* This 

 thickness of the rock was decayed gneiss, and undecomposed rock was 

 not found at this depth of excavation. 



In one of the plates accompanying Pissis' paper he represents the de- 

 composed gneiss at the base of the Corcovado at Rio as having a depth 

 of 120 meters.f I know that the gneiss at this point is profoundly de- 

 composed, but I never saw it exposed to such a depth as he there repre- 



* Relatorio sobre o reservatorio D. Pedro II. W. Milnor Roberts. Revista de Engenharia, ii, no. 

 7, Rio de Janeiro, July 15, 1880, pp. 106, 111, 112. 

 t La position geologique des terrains de la partie australe du Bresil. M. A. Pissis, 1842, p. 358. 



