LIJfNEAN SOCIETY OF LONDON. 53 



it has evolved ouly once, and from it on the one hand diverged 

 the Beunettitales and on the other the direct ancestors of tlie 

 Angiosperins. 



Let me now consider shortly the manner in which cones may 

 have arisen from the loosely-arranged sporophylls met with in the 

 J:*terido,-iperms. Without speculating as to how originally the two 

 kinds of sporangia were borne relative to one another on the 

 Pteridospermous frond, it is safe to assume that eventually two 

 kinds of spore-bearing fronds were evolved, viz., the mega- and 

 miorosporophyll. The massing together of such sporophylls into 

 cones (strobili) can be conceived of as taking place in two ways. 

 Either both kinds of sporo]ihylls were aggregated into one and 

 the same cone, producing the bisexual condition; or each kind 

 was segregated apart, forming liistinct male and female strobili. 

 The Beunettitales evidently took the former course, and the 

 Cycadales probabi}'- the latter. The ari-angement of the aiega- 

 sporophylls in Cycas is dillicult to reconcile with the view that the. 

 diclinism of the Cycads arose from an earlier hermaphrodite state 

 through reduction. 



Adopting the view of a Pteridospermous origin of all Gymno- 

 sperms, one may hazard the 0[)inion that the ancestors of the 

 Conifers were evolved from the Pteridosperms at a very early 

 period by segregating their sporophylls into unisexual strobili. 

 Ou this supposition the uuisexual nature of their cone has not 

 been due to reduction from a previous bisexual condition. There 

 is no evidence to show tliat any of the Coniferales, including their 

 forerunners, tlie Cordaitales, were ever other than unisexual. The 

 (xinkgoales were probably another evolution from the same plexus 

 which counnenced on unisexual lines. Consequently, of all 

 known Gymnosperms, extinct and extant, tlie Bennettitales and 

 the Gnetales would a[)pear to be the only groups to ha\ e possessed 

 primitively bisexual cones. 



In our joint paper we advanced the view that the rise of the 

 Angiospermous type of anthostrobilus — the one with the closed 

 carpel — was bound up with the substitution of entomophily for 

 anemophily (5. p. 73). The writer is now inclined to extend this 

 idea and to suggest that the antlioslrobilus may have owed its 

 origin to insect-visitation. Adopting the view that cross-fertili- 

 sation is of paramount importance in evolution, it follows that in 

 the case of wind-pollinated plants it is an advantage in their 

 passage to the stri)l)ilate condition to segi'egate the two kinds of 

 s[)orophvlls. The chances fur cross-pollination will thereby be 

 increased. Possibly this may have been the general trend in 

 cone-formation among seed-plants in Palteozoic times before 

 pollen-seeking insects appeared. Such a nutritive i)abuluiii as 

 pollen, one can imagine, would early attract primitive insects. In 

 seeking it from the primitively unisexual Gymnosperms no advan- 

 tage in the way of cross-fertilisation would accrue, as the female 

 cones would not be visited. The Pteridosperms would on this 

 supposition also be visited for pollen. They would potentially be 



