TORSA. GEOLOGY. 149 



so that we have seldom an opportunity of seeing rocks 

 that have undergone this operation and are still in a 

 state of integrity ; the decomposed parts being commonly 

 removed as soon as that process is completed. Yet in 

 certain situations where granite or gneiss have been pro- 

 tected from mechanical violence by a covering of soil 

 or peat, they are sometimes found decomposed to great 

 depths; retaining nevertheless their natural structure and 

 much of their ordinary aspect; but ready to fall to pieces 

 on the application of any violence. I have mentioned 

 an example of this in the gneiss of Sky, and it is also 

 very conspicuous in Guernsey, in the same rock, as well 

 as in the very small portion of granite which occurs in 

 the Isle of Man. As a case more in point, it is also 

 to be seen in the trap rocks in many parts of Sky, 

 as was remarked when treating of that island. The 

 weathering of the vein in Torsa is an analogous case ; 

 although the rock is not so far affected by the progress 

 of the decomposition as to be susceptible of actual disin- 

 tegration. It is not at present easy to explain the causes 

 of these deep-seated changes in rocks, as it is not easy 

 to conceive how air and water can find access to their 

 interior parts; but it is possible that they may result 

 from agencies of an internal nature independent of the 

 action of the weather. Under such circumstances it is 

 also possible that many rocks of the trap family, whether 

 belonging to the traps strictly so called, or to the syenite, 

 porphyry, and claystone divisions, may have been changed 

 from the state in which they were originally formed, by 

 corresponding causes. In such a case the changes would 

 neither be limited in extent nor position ; they would not 

 be confined to the surface of one mass of rock, nor to 

 the outer of many consecutive beds ; but might perhaps 

 occupy the whole of a single mass, or merely affect an 

 inner one of which the composition admits such an altera- 

 tion, when the outer, differently constituted, might escape. 

 However speculative this view may be deemed, it is not 



