Evolution of Plants 359 



flowers (Ephedra, Gnehim) a staminal and ovular structure 

 that is greatly more specialized than in Cordaitales, and even 

 than in other gymnosperms. For each stamen is now divided 

 into filament and anther, while the stamens surround the 

 pistil and mature TN-ith it, in sub-hermaphrodite flowers of 

 Welwitschia, after an exactly angiospermic manner. 



While too much stress need not be placed on the condition, 

 Pearson's demonstration of insect pollination in flowers of 

 Welwitschia is indicative that this genus is probably the oldest 

 now surviving in which such a biological connection had de- 

 veloped. 



The formation of secondary floral leaves round the ovules 

 as in Gnetum, and the great prolongation of the ovular integu- 

 ments beyond the ovular mass, so as physiologically to simu- 

 late a style, in all three genera, are adumbrations of higher 

 angiospermic characters. The varying modes of development 

 of prothallial tissue, of egg-cells, and of embryo, which have 

 been elucidated during the past two decades, cannot be treated 

 in detail here, though some consideration will be given to 

 them later (p. 365). All however emphasize the transition 

 value of the three genera. 



But in all of these, as in the higher Gymnospennia and 

 in all of the Angiospermia, the production of large motile 

 spermatozoid nuclei — a hereditary derivative by descent from 

 plants wdth free vegetative prothallia — has entirely ceased, 

 and instead only motionless sperm-nuclei, that are guided 

 -to the egg-cells by chemotactic affinity, are produced. 



At this point now, instead of advancing to immediate con- 

 sideration of the angiosperms, we may turn back to trace 

 two important pathways of advance up to the gymnospermic 

 stage, both of which evidently have been derived from a more 

 ancient cordaital ancestry. These in the diagram we have 

 designated the Ginkgoales and the Protopinales, or in common 

 language the maidenhair trees and the primitive pines. 



Though by no means abundant, and only imperfectly known 

 to us as yet in the fossil state, not a few genera of plants have 

 been described that strongly indicate a gradual splitting up 



