DIV. n ANGIOSPERMAE 607 



Any direct transition from Gymnosperms to Monocotyledons is 

 thus out of the question while a relation of dicotyledonous plants to 

 Gymnosperms is not excluded. The parallels and progressive develop- 

 ments that can be recognised in the male and female organs have been 

 referred to above (p. 544 f .) ; there are also indications of the deriva- 

 tion of the one group from the other in the construction of the 

 flower as a whole. In attacking this problem WETTSTEIN attempts 

 to derive the simplest flowers of the Angiosperms from Gymnosperm 

 inflorescences. 



A male flower with a single perianth and superposed stamens could be derived 

 from a whorl of scale leaves with simple axillary male flowers. Since in male 

 inflorescences of Ephedra single female flowers occasionally appear, it is possible 

 that the female organs might become associated with the stamens. The proba- 

 bility of such a transition is increased by the fact that insect -pollination has 

 been observed in inflorescences of this kind. 



If the systematic arrangement of the Dicotyledons is based on this 

 idea, the most simply constructed flowers would be those with one 

 whorl of perianth segments and borne in catkins. Thus the 

 Casuarinaceae, Juglandaceae, Betulaceae, Ulmaceae, etc., will be placed 

 at the beginning of the system, and to them will be connected the 

 other families with a simple perianth which are grouped together as 

 Monochlamydeae. To these in turn may be connected the Dialypetalae, 

 the flowers of which have both calyx and corolla. The forms with a 

 gamopetalous corolla are separated as the Sympetalae, and the other 

 Monochlamydeae and Dialypetalae contrasted with them as Chori- 

 petalae ; the forms without perianth are grouped with the Choripetalae. 

 Jiince within the Monochlamydeae various lines lead from forms with a 

 simple perianth to those with a pentacyclic structure, any arrange- 

 ment in a simple ascending series is impossible. Various parallel series 

 lead from simple to highly organised floral structure, and similarly 

 numerous parallel series are found in the Dialypetalae. Thus the 

 natural or phylogenetic relationships can only be exhibited in an 

 incomplete fashion in the following arrangement. 



In addition to this line of transition from Gymnosperms to 

 Angiosperms another possibility has to be seriously considered ; this 

 was pointed out a considerable time ago by H. HALLIER. He treated 

 the Polycarpicae, from which the Monocotyledons have been derived 

 above, as the starting-point for the Dicotyledons generally. This view 

 finds support in a biological observation of DIELS ( 17 ), who showed that 

 both some South African species of Encephaknios and some of the 

 Polycarpicae are pollinated by beetles. Since the Coleoptera are the 

 phylogenetically oldest flower-visiting insects and appear as the 

 pollinating agents in the oldest family of Gymnosperms, a similar age 

 may be inferred for the Polycarpicae that are pollinated by beetles. 



The morphological construction of the flower of the Polycarpicae, with its 

 spiral arrangement of all the floral leaves, presents resemblances to the greatly 



