CHAPTER XIV, 



THE SYSTEMATIC ARRANGEMENT OF THE ANGIOSPERMS. 



547. Many attempts haye been made to arrange the vast 

 number of species of Angiosperms in a logical system, but 

 none of them have proved to be quite satisfactory. For a 

 long time the Candollean system, and later its modification 

 by Bentham and Hooker, have been foUovred in most 

 botanical publications, but within a few years the system 

 of Engler and Prantl has been favorably received by many 

 botanists. 



548. The sequence adopted in this chapter differs some- 

 what from either system mentioned, and is based upon the 

 proposition that in the primitive flower all the parts were 

 separate. The first flowers on the earth, in the Permian 

 or Triassic period, must have been apocarpous, that is, 

 with their pistils simple and separate. Their stamens 

 must, likewise, have been separate from one another and 

 from other oi'gans. So too their floral leaves (perianth) 

 must have been of separate phyllomes. This is the struc- 

 ture of the typical Apocarpse, the lower Thalamiflorse, and 

 the lower Calyciflorffi, which are accordingly placed at the 

 beginning of the system. 



549. The earliest modification of this primitive structure 

 was probably the union of the carpels into a compound 

 pistil, as in the Ooronariess and many families of the 



330 



