13'2 TAXONOMY. 



\vhorl-sha|X}d position belongs to the higher, or to the albu- 

 minous plants. But there are many exceptions to this. 



197. 



It is very difficult to establish any fixed principles respect- 

 ing the value of numerical proportion. On the one hand, 

 Nature shews a regularity which cannot be mistaken in the 

 numerical proportions of many parts, as, for instance, in the 

 impregnating organs of the Orchidese and Liliaceae. On the 

 other hand, she seems sometimes to sport in such a manner 

 with numbers, that in many genera we scarcely find the same 

 number of filaments and pistils in ail the species. Astonish- 

 ing changes often take place, in this res}:>ect, from abortion 

 and union of parts. We shall endeavour to state some 

 rules on this subject. 



198(7. 



Numerical proportion appears to be more steadfast, and 

 consequently more important, the more scanty the number 

 is. It is on this account that the numerical proportion 

 of the anthers remains so steadfast in the Scitaminea?, Or- 

 chideae, and Grasses, and also in the Labiatae, because they 

 have only one, three, or four. Dodecandria Plants observe 

 much seldomer tlie same numerical proportion than those of 

 the Class Hexandria. But there are exceptions also to this 

 rule, of which Valerinna and Bocrhavia are well known in- 

 stances. 



In the organs of impregnation, unity seems chiefly to be- 

 long, as a character, to the pistil. In most of the other parts 

 of fructification, unity appears only as a consequence of abor- 

 tion or union. This has been already remarked in the 

 Orchidese; it is equally evident respecting the Scitamineae, 

 because on both sides of the principal filaments, these com- 

 monly have two filiform processes, which, on account of the 

 resemblance of the plant in other respects to the Irideae, 

 lead us to suspect that they are properly two abortive fila- 

 ments. 



