INTRODUCTION 53 



methods as a means of arrangement and a mode of 

 of subdivision, was no improvement upon earlier 

 systems from the natural standpoint. All these 

 systems were artificial. They hit upon some one 

 factor which was applied to all groups of plants, such 

 as the tree habit. 



A truly natural classification is necessarily based 

 upon more than one character and takes into account 

 the essential differences. 



The vegetative structure is more variable than the 

 floral structure. Hence the latter is most valuable 

 as a basis for classification. The natural arrangement 

 was more or less anticipated by Ray, who recognised 

 the divisions of flowering plants into Dicotyledons 

 and Monocotyledons. 



De Candolle, improving Linnaeus's later attempts 

 to found a Natural System, and adopting Jussieu's 

 divisions Apetalae, Monopetalae, Polypetalae, laid the 

 foundations of Bentham and Hooker's system. 

 Before the latter was elaborated Brown discovered 

 the distinction between Gymnosperms and Angio- 

 sperms. The later classifications have been largely 

 based upon the modifications of the whole floral 

 structure in relation to the visits of insects and 

 pollination, and though not formed primarily as a 

 result of a study of the latter, researches carried out 

 by Sprengel and others, and later by Miiller, have 

 shown that these divisions are related to the adapta- 

 tion of the flower to entomophily. 



Botanical description, as distinct from nomen- 



