262 



PEINCIPLES Ob 1 PL ANT-TEE ATOLOGY. 



seems quite as reasonable to suppose that the ancestral 

 condition has been preserved all along as that it has 

 once been submerged and again resurrected. Further, 

 the obvious relationship of the PapaveraceaB to the 

 Poly car picas subverts this idea. 



Abnormal changes in the direction of simplification 

 and reduction, i. e. in the direction of evolution, are 

 quite frequent. The very common 2-merous flowers 

 of orchids afford a good instance of a progressive 

 abnormality. 



Velenovsky regards the abnormal 2-merous corolla 

 of Forsythia, mentioned on a former page, as a rever- 

 sion to the immediate ancestry which exhibited 2-mery 

 in all floral whorls. But in the closely allied genus 

 Fraxinus there is a manifest process of reduction 

 obtaining in the corolla, for this is completely absent 

 in the common ash (F. excelsior). The 2-merous 

 corolla of F. dipetala, Velenovsky marshals, on the 

 flimsiest of grounds (that of the common occurrence 

 of 2-mery in the order), in favour of 2-mery being the 

 primitive condition ; it is, on the contrary, best regarded 

 as a half-way stage between the 4-merous corolla of 

 the manna ash {F. Ornas) and the completely suppressed 

 one of F. excelsior. This is strongly supported by the 

 fact that in the majority of Oleacege the 4-merous 

 corolla of F. Ornus obtains. In any case the completely 

 suppressed corolla of F. excelsior can be regarded in no 

 other light than as a very recently-acquired feature. 

 If, then, reduction has occurred in this species, it is 

 most probable that this is the cause of the condition in 

 F. dipetala. There are thus two valid reasons for 

 regarding the 2-merous condition of the corolla in 

 this order, whether normal or abnormal, as a pro- 

 gressive and not a reversionary phenomenon. If this 

 be true of the corolla, it is also almost certainly true 

 of the androecium, and the 4-merous andrcecium of 

 T essarandra (belonging to the same order) must be 

 regarded as more primitive than the much commoner 

 2-merous one. If the 2-merous condition is a pro- 



