HELATION OF COLOR AND TEXTURE TO DECAY. 293 



Mr Beudant has proven that the rounded forms in crystalUzation may be caused 

 by a sudden jar or shock, as may be beaatifully shown in a saturated salt sohition, 

 which forms genuine globes when the vessel is slightly jarred." 



Color as related to decay. — The colors of the rocks of Brazil are a matter 

 of some importance in connection with this subject. It is noteworthy 

 that the natural surfaces of the gneisses and granites are not white, but 

 are of a dark chocolate brown or dark gray color,* colors which absorb 

 and radiate the heat more promptly than would the colors of fresh sur- 

 faces, which are all light. 



So far as I am aware, scarcely a single dike of basic rock is known in 

 Brazil which is not profoundly decomposed. The diorites and diabases 

 are often found as boulders of decomposition, but the only rocks of the 

 kind which have been reported not so decomposed are the quartz porphy- 

 ries of the island of Santo Alexio, and these appear fresh only because 

 the decayed rock is removed by the waves as fast as decay takes place. 



This more rapid decay is to be attributed in part, I believe, to the dark 

 color of the rock, which causes it to absorb and radiate its temperatures 

 more rapidly than do the gneisses.f 



In addition to the temperature effects, it has already been pointed out 

 that the basic rocks are more soluble in acidulated waters than are the 

 acid granites. j; 



Texture as related to decay. — The boulders of decomposition are mostly 

 of rather fine grained rock. Coarsely crystalline rocks I seldom saw as 

 boulders, and when they were so found they were comparatively new. 

 These two facts led me to believe that coarsely crystalline rocks were 

 more readily attacked by the agents of decomposition than were the 

 more compact ones. This is also in keeping with my experience of crys- 

 talline rocks elsewhere. In Arkansas, for example, the more compact 

 blue syenites seem to be unaffected below the thin coat of decayed rock, 

 while the light gray, coarse grained ones are partially decomposed to a 

 depth of several feet — as far, in fact, as they have been quarried. This is 



* Burmeister notes that all the undecayed granite walls of Brazil are blackened and ^streaked 

 by water. (Reise nach Brasilien, p. 518.) 



fThe difficulties surrounding the absorption and radiation of heat and the want of uniformity 

 in methods of observation have prevented my giving specific data on this subject. Actinometrie 

 observations were made at the observatory at Rio for a few years, but it is not clear that they are 

 available in the present study, as will be seen from Dr Cruls' description of the apparatus : " The 

 actinometer employed is that of Dall-Eco of Florence, and is composed of two conjugated ther- 

 mometers of Salleron. . . . One has the bulb blackened with lamp-black, the other is un- 

 blackened. Each thermometer is enclosed 7'?z vacuo in a glass tube." (Bulletin Astronomique et 

 Meteorologique de I'Observatoire Imperial de Rio de Janeiro, 1882, pp. 23, 24.) In March, 188:>, the 

 difference in reading between these thermometers ranged from 3.4 to 16.6 degrees centigrade at 

 9 a. m. ; from 4 to 16.3 degrees at noon, and from 0.4 to 16 degrees at 3 p. m. The mean differences 

 during that month were 13.26 degrees for 9 a. m., 13.5 degrees for noon, and 12.17 degrees for 3 p. m. 

 (Annaes do Imperial Observatorio do Rio de Janeiro, iv, 1889.) 



J Annales des Mines, 7me ser., viii, p. 698. 



