ISLE OF MAN. GEOLOGY. GRANITE. 537 



the decomposed rock, the former is convreted into a 

 yellow clay, and the latter into a greyish one. In addition 

 to this, small strings and fragments of a fine dark clay are 

 found in it. These appear to result from the decomposi- 

 tion of the entangled fragments of schist already described, 

 and have been sought after with some anxiety for the 

 purpose of a polishing powder : they form the Dun 

 earth of the islanders. The mode in which the granite 

 decomposes, is not very common : at the same time it is 

 well known to geologists. In trap, this mode of decom- 

 position is frequent, and is not the only point of resem- 

 blance which occurs between the older and the most recent 

 unstratified rocks. Although the general mass of the rock 

 is at first uniform, and apparently homogeneous, the 

 commencement of decomposition is marked by circular 

 hues which respect a central nucleus, or rather different 

 centres. A section exposes the whole process, which 

 takes place without the contact of air. While the parts 

 most distant from the centre of each nucleus, are com- 

 pletely detached in successive crusts and converted 

 into clay, lines, gradually becoming more feeble, are 

 found marked in succession towards the centre, where 

 the solid rock, ultimately destined to undergo the same 

 change, is still in a state of integrity.* The intervals 

 between these irregularly globular bodies are occupied 

 by curved and by straight lines, the granite in certain 

 parts decomposing in schistose laminse. In a general 

 view there is nothing to be remarked on this subject 

 which has not frequently been discussed already, and 

 the appearance in this case, as in that of trap, has been 

 by all geologists referred to the concretionary structure. 

 But a circumstance, which has recently occurred in my 

 observations, proves that in certain cases at least, the 

 exfoliation of granite is independent of any internal struc- 

 ture. It was already alluded to on a former occasion. 



* Plate XXVIII. fig. 3. 



