28 



modern experimentalists have long since ex- 

 ploded. 



A geological arrangement of rocks is in fact a 

 history of their natural affinities : it is a branch of 

 geological science ; and, to this main object, all 

 minor considerations ought to be rendered sub- 

 servient, even at the risk of some trifling incon- 

 sistencies of order. That order which is the most 

 useful ought first to be selected : if it were possi- 

 ble to add to utility, an unexceptionable regula- 

 rity, such a system would be perfect : but a pre- 

 cision which serves no purpose, is a shadow de- 

 ceiving with the appearance of a substance. 



In examining the conditions requisite for a 

 perfect geological arrangement, it will be more 

 easy to judge of the value and amount of the pre- 

 sent defects, and to conjecture what chance there 

 is of future improvement. 



The order of arrangement of every rock in na- 

 ture ought to be known : to ensure facility, as 

 well as perfection, it ought indeed to be con- 

 stant- 



Every rock ought to be definite and invaria- 



