﻿Manchester Memoirs, Vol. xlix. (1905), No. 12. 17 



there is no longer any difficulty in accounting for the 

 advent of the Cycadophytic vegetation which so rapidly 

 established itself in the succeeding epoch. Its source is 

 doubtless to be sought among the Pteridosperms, and 

 through them may be traced back to ancestors with 

 Filicinean affinities. 



At this point, however, a word of caution is not out of 

 place. Those who take the view which I am endeavouring 

 to commend to you this afternoon, are accustomed to speak, 

 for shortness' sake, of the origin of the Gymnosperms, or at 

 least of the Cycadophyta, from ferns. It must, however, be 

 remembered that seed-bearing plants, including both 

 Pteridospermeae and members of the more advanced 

 family Cordaiteae, can be traced as far back as the ferns 

 themselves. They appear to have been well represented 

 in the Devonian, and so far as I am aware, records of 

 land-plants earlier than the Devonian are still too dubious 

 to take into account. Further, the Pteridospermeae, 

 though so clearly related to the ferns by many of their 

 characters, possessed, as we have seen, not merely seeds, 

 but highly organized seeds. Their most striking pecu- 

 liarity — and this extends to all known Palaeozoic seed- 

 plants — is the constant absence of an embryo, though the 

 preservation is often so good that we should expect to 

 be able to detect it if it had been there, as is the case 

 in the Mesozoic Bennettites, where the embryo is well 

 preserved. It is probable that the postponement of the 

 embryonic development till germination began (a condition 

 of which reminiscences persist in living Cycads) may ready 

 have been a characteristic of ancient seeds. But granting 

 them this primitive character, the fact remains that the 

 earliest known seeds were already so far advanced in 

 their evolution as to leave little trace of their derivation 

 from cryptogamic sporangia. Thus, properly speaking, 



