12 BAHIA BRAZIL. [CHAP. I. 



could hare denuded the granite over so many thousand square 

 leagues ? 



On a point not far from the city, where a rivulet entered the sea, 

 I observed a fact connected with a subject discussed by Humboldt.* 

 At the cataracts of the great rivers Orinoco, Nile, and Congo, the 

 syenitic rocks are coated by a black substance, appearing as if they 

 had been polished with plumbago. The layer is of extreme thin- 

 ness ; and on analysis by Berzelius it was found to consist of the 

 oxides of manganese and iron. In the Orinoco it occurs on the 

 rocks periodically washed by the floods, and in those parts alone 

 where the stream is rapid ; or, as the Indians say, " the rocks are 

 black where the waters are white." Here the coating is of a rich 

 brown instead of a black colour, and seems to be composed of ferru- 

 ginous matter alone. Hand specimens fail to give a just idea of 

 these brown burnished stones which glitter in the sun's rays. 

 They occur only within the limits of the tidal waves ; and as the 

 rivulet slowly trickles down, the surf must supply the polishing 

 power of the cataracts in the great rivers. In like manner, the 

 rise and fall of the tide probably answer to the periodical inunda- 

 tions; and thus the same effects are produced under apparently 

 different but really similar circumstances. The origin, however, 

 of these coatings of metallic oxides, which seem as if cemented to the 

 rocks, is not understood ; and no reason, I believe, can be assigned 

 for their thickness remaining the same. 



One day I was amused by watching the habits of the Diodou 

 autennatus, which was caught swimming near the shore. This 

 fish, with its flabby skin, is well known to possess the singular 

 power of distending itself into a nearly spherical form. After 

 having been taken out of water for a short time, and then again 

 immersed in it, a considerable quantity both of water and air is 

 absorbed by the mouth, and perhaps likewise by the branchial 

 orifices. This process is effected by two methods: the air is 

 swallowed, and is then forced into the cavity of the body, its 

 return being prevented by a muscular contraction which is exter- 

 nally visible : but the water enters in a gentle stream through 

 the mouth, which is kept wide open and motionless; this latter 

 action must, therefore, depend on suction. The skin about the 

 abdomen is much looser than that on the back; hence, during 

 the inflation, the lower surface becomes far more distended than 

 * Pers. Narr., vol. v. pt. i. p. 18. 



