29 



ble ; constant in its mineral characters, as in its 

 position. 



Every principal division, or species, should 

 appropriate to itself a distinct set of subdivisions. 



As yet, however, no prevailing general ar- 

 rangement among rocks has been ascertained, 

 although it has been imagined to exist : and it is 

 indeed now certain, that no order can be assumed 

 which is not subject to numerous exceptions. In 

 the same way, it is demonstrated that the order of 

 succession is inconstant in all the minor details, 

 and moreover, subject to frequent variations. 



So far also from being definite and invariable 

 in their mineral characters, rocks are subject to 

 frequent changes, both of structure and of compo- 

 sition ; passing, in many cases, into each other 

 by imperceptible transitions. This objection, 

 however, operates equally against a mineralogical 

 arrangement. 



Lastly, rocks resembling each other in com- 

 position, are found under divisions often far re- 

 mote in geological character and position. 



These are, unquestionably, serious defects; 



