in Natural Hislori/. 491 



least liable to vary. His classes and orders are avowedly so 

 many assumptions, which practice has shown to be convenient ; 

 but when we come to genera, the artificial system falls in with 

 the natural, as Linneeus framed their characters upon resem- 

 blances founded in nature. 



Now in the natural system this machinery of terms cannot be 

 employed in the same manner. It is an ascending series from 

 the less to the greater predicate. From genera we proceed up- 

 wards to orders, and orders we combine into classes. We be- 

 come more and more general in our characters, instead of more 

 and more definite. Here indeed we ought not to sacrifice, as in 

 the artificial scheme, to convenience; and break up well-defined 

 genera and orders because they contain a large number of spe- 

 cies. If we find a large genus, for instance, as Erica, agreeing 

 in some well-marked characters of structure, form, station, and 

 properties, it appears contrary to the end proposed by the na- 

 tural system, to divide and subdivide the species into small 

 groups, and to give each of these the same value as is now pos- 

 sessed by the whole. This is frittering away characters which 

 are essential to the use of a genus, and destroying our power 

 over it when we proceed to generalise. The value of generic 

 terms consists essentially in the distinct conceptions we have of 

 them ; but if we go on to multiply them, as is at present the 

 fashion, we render it as impossible to circumscribe them, as it is 

 to parcel out the colours of the rainbow ; and instead of making 

 Natural History familiar and popular, it will requii'e the com- 

 pass of a man's life to master the terms we employ. If indeed 

 the object be to analyse, division may be very convenient, be- 

 cause the inquirer may be otherwise bewildered in the multitude 

 of particulars. It does not follow from hence that the student 

 of the natural system may not avail himself of subordinate 

 groups by whatever characters they may furnish; only the 

 giving them equivalent names, and making them co-ordinate, 



is 



