106 TAXONOMY. 



as in the Scitamineae, Orchidea% and Coronariae, which must 

 often lead us into error in arranging these vegetable tribes. 



159. 



To this must be added, that there are a multitude of sepa- 

 rate genera, and even of entire families, which cannot at pre- 

 sent be arranged with respect to others. The genera Bego- 

 nia^ CynomGrium^ Datisca, and Nepenthes^ for instance, are 

 completely unknown in their affinities ; and it thus remains 

 doubtful, whether the Caryophylleae are not more properly 

 connected with the Chenopodea?, than with the Liliaceas and 

 Myrteae. 



160. 



We here already perceive one of the difficulties which the 

 Natural Order presents in instruction. For as several gene- 

 ra stand quite insulated in this order, — as others have some 

 properties of one family, and others have the marks of a dif- 

 ferent family, — as the situation of families with respect to one 

 another is not pointed out by nature, but is the work of hu- 

 man genius, — we must confess that such a method, from its 

 uncertainty and difficulty, is by no means fitted for begin- 

 ners, however elevating to the human mind the study of those 

 alhances may be, — however much it sets all the powers of the 

 mind in activity, — and however much it is to be wished that 

 this species of knowledge should constantly be making pro- 

 gress. 



161. 



The difficulty of the natural method is still greater, when 

 w^e look for some bond, or, what is the same thing, for 

 some common form, which may unite the natural families 

 with each other, and lead us to their arrangement. If we 

 take this leading principle from one or a few ever so essen- 

 tial parts, we in fact do nothing else but connect an artificial 

 system with the natural arrangement, and this arrangement it- 

 self can no longer be called a natural one. We may employ, 

 as our principle, the form and situation of the embryon, 



