554 PROCEEDINGS OF THE GEOLOGICAL SOCIETY. [June 13, 



(c) The above-mentioned two kinds of nuclei, more or less inter- 

 mixed, but with smaller interstices, occupied with red earth. 



The above varieties of condition of the ferruginous hornblendic 

 rocks sufficiently indicate that they are more or less composed of a 

 paste or matrix, relatively more easily decomposable than the con- 

 tained concretionary nuclei. This paste, decomposed, has furnished 

 the Red Soil. When in situ, it has not the fidl red hue of that 

 soil, on account of its containing all the insoluble ingredients of the 

 original rock ; amongst these is alumina, the presence of which, from 

 its white colour, diminishes the depth of red due to the peroxide of 

 iron so largely contained in the red soil. The alumina is in a highly 

 pulverulent state, and, in consequence, is longer held in suspension 

 in water than the other ingredients of the rock, and thus is generally 

 separated by being transported to a greater distance than the other 

 materials, when the debris is subjected to the transporting influence 

 of water in motion. 



This separation can be easily performed artificially, and the natural 

 operation is to be seen going on in brooks running amongst or over 

 the decomposed rock : the bright red soil is deposited in the less 

 rapid parts of the stream, while the more buoyant alumina is carried 

 further away. 



The deposition of the Red Soil. — Very frequently the red soil of 

 Southern India rests on granitic gneiss. That this granitic floor 

 has been the bed of a sea or lake is amply demonstrated by the 

 water- worn stones of granite, quartz, &c. which sections through 

 the red soil down to this granitic floor brought to light. I had 

 often seen these water-worn stones cropping out at the bottom of 

 the nearly perpendicular red-soil banks of brooks, but concluded 

 they were due to the action of the water of those brooks ; but, 

 when excavating through red soil at distances from water-courses, I 

 found that, on reaching the granitic bed, it was rare that between 

 it and the red soil a layer of water-worn stones was not found : ob- 

 servations made over a considerable tract of country, and miles apart, 

 fully proved that even the present elevated table-land of the Mysore, 

 Southern India, has been at one time under water, of which these 

 widely-spread water- worn stones are indubitable proof. 



Thus, while the stratum on which the red soil is generally found 

 to rest is shown to have been under water, the movement of which 



a nearly perpendicular bank, about 8 feet high and some 300 or 400 feet long, 

 there crop out spheroidal nuclei, varying from a few inches to about 3 feet in 

 diameter, and of concentric structure : they are distributed at distances from one 

 to several feet apart, in a brownish earth, which is the debris in situ of the more 

 easily decomposable portion of the trap. If a stream of water were to run over 

 this partially decomposed trap, the loose earth in which the concretionary sphe- 

 roidal nuclei are at present imbedded would be washed away to form a deposit in 

 the Loch, if the stream run into it, while the nuclei would present a confused 

 assemblage of rounded stones, and the interstices of those at the top of the bank 

 would be empty. Thus on a small scale we should have realized that which, on 

 the wide-spread surface of Peninsular India, has in bygone ages taken place on a 

 gigantic scale. 



