45 



as some of them contain no subordinate members, 

 while others comprise an extensive range of sub- 

 stances, often very discordant, united by one dis- 

 tinct geological bond. 



It is in these subdivisions, or families, if the 

 term may be admitted, that the defects of the 

 geological method of arrangement appear most 

 striking. They may hereafter perhaps be dimi- 

 nished ; but they cannot be wholly removed, as 

 they belong to the very essence of the arrange- 

 ment. Nature has, in one place, given us, per- 

 haps, twenty rocks, united by a common bond 

 of mutual transition, and bearing one general re- 

 lation to the surrounding substances ; in another 

 place, we find only one rock, or perhaps two, in- 

 dependent of all the others, but still bearing to 

 them, in a similar manner, one general and fixed 

 relation . 



It is obvious that, under such circumstances, 

 no family, or no rock which contains more than 

 one member, or variety, can admit of that which 

 is properly called a definition. A notion more 

 or less distinct may sometimes be conveyed, by a 



