EVOLUTION OF SEED-PLANTS 131 



in essentials from some of the seeds of Pterido- 

 sperms. The form of the seed is flattened, which 

 is also the case in some members of the latter 

 class. It is not possible at present to distinguish 

 with certainty in all cases between the seeds of 

 Pteridosperms and those of Cordaitese. This 

 is an imnortant point, for in most respects these 

 great trees bore little enough resemblance to 

 the Seed-ferns, A fairly complete series of Pal- 

 aeozoic stems, however, is known, connecting 

 the stem-structure of Cordaites with that of such 

 a Pteridosperm as Lyginodendron, while the struc- 

 ture of the leaves, as already mentioned, is like 

 that of a Cycad, though they were simple in- 

 stead of compound in form. On the whole, it is 

 evident that the Cordaitese were related to the 

 Pteridosperm-Cycadophyte line of descent; the 

 enormous antiquity of such highly organised 

 Seed-plants is a most striking fact; their connec- 

 tion with the Seed-fern stock must lie very far 

 back, in Lower Devonian times, if not earlier. 



It thus appear that during Palaeozoic ages 

 the Seed-plants were represented by at least 

 two great classes, both of which go back as far 

 as our records of a Land-flora extend. On the 

 one hand, there were the Seed-ferns, probably 

 the more extensive group of the two; plants which, 

 while retaining many of the characters of the 

 Fern stock, had already evolved a high type of 



