184 CRETACEOUS KOCKS OF S. rSDIA. [PAR'T II. § 2. 



The thickness of this soil is very variable. On much of' the gneiss 



Thickness and stiw country it does not exceed 1 or 2 feet, but 



"^*^* I have noticed elsewhere sections of 10 or 12 



feet, as on the watershed of the Vellaur and Cauvery Rivers, and it 



probably attains a much greater thickness in the valley of the former 



river. 



• The regur is usually tolerably homogeneous, and exhibits little 

 structure. In the ochreous clay of the Ootatoor beds, where, as above 

 mentioned, the lower part of the regur is tinted by the admixture of 

 the local rock, this passes gradually upward into the usual black mud. 

 Elsewhere I have noticed irregular alternating beds of different tints. 



As regards the relations of the regur to the underlying rocks, it occurs 

 indiscriminately on the Cretaceous rocks and gneiss, and though I 

 have never seen it in section resting on the Cuddalore sandstones, I 

 believe that it does so on the left bank of the Coleroon, and there can be 

 little doubt that it does in the lower part of the valley of the Vellaur. 

 The mineral character of the lower part, as I have mentioned, is occasion- 

 ally affected by that of the rock on which it rests, but this is rarely the 

 case for more than 2 or 3 feet, and the character of the soil generally 

 may be said to be independent of that of the local underlying rocks. 



The only condition which I have observed generally to determine 



the occurrence of regur is that of relative low 

 Elevation the detennin- 

 ing condition of its occur- elevation, that is to say it affects de[)ressions in the 

 rencc. 



country irrespective of absolute height above the 



sea, while the sandy soils and lal occupy the higher ground. Before how- 

 ever entering into this, which is a most important point, as bearing on 

 the theory of its origin, I will pass on to the description of the sandy 

 soils, the relations of which to regur are such, that it is impossible to 

 avoid the conclusion, that the two soils are due to concurrent causes. 



4. Sandy soz'/s.— Under this head I include the ferruginous sand 

 which covers the Cuddalore sandstones in Tanjore, Trichinopoly, 



