GENERAL PHYLOGENY 155 



relation to the Cycadean line of descent but also offer many 

 suggestions of that course of development which is realized in 

 the higher Coniferales. They therefore constitute the real start- 

 ing point for two lines of descent, the first of which embraces 

 the Cycadales. At the present moment we have little or nothing 

 to do with this beyond establishing its probable relation to the 

 other gymnosperms. The second line emerges in a type of plants 

 having characteristics distinctly allied to those of the Coniferae, 

 and it is this line of descent with which we are now chiefly con- 

 cerned. It is now possible to define the origin of this phylum 

 somewhat more exactly than Coulter has done (11, 12), since 

 there is good reason to believe that it emerges from the Cycado- 

 filices through Poroxylon. Scott (81, 398) has already pointed 

 out the relations of this genus to the Cycadofilices and the 

 Cycadaceae on the one hand, and to Cordaites on the other, so 

 clearly as to remove the necessity for detailed discussion at this 

 time, beyond giving emphasis to one or two important structural 

 relations. In Calamopitys saturni it has been noted that the most 

 primitive distribution of the bordered pits upon both the radial 

 and tangential walls is represented in the protoxylem structure. 

 Such distribution, however, undergoes rapid modification where- 

 by it is wholly limited to the radial walls in the secondary wood. 

 A similar limitation appears in other somewhat closely related 

 genera, and it is fully expressed in Poroxylon, where the multi- 

 seriate disposition and hexagonal form are typically preserved, 

 though there is at the same time a tendency to segregation to 

 such an extent that the pits sometimes become round. In this 

 it is possible to notice the first indication of a character which, 

 while infrequent, is nevertheless occasionally expressed among 

 the Cordaitales, though it is generally characteristic of the related 

 phyla, Gingkoales and Coniferales. 



Among the Cordaitales there is but one genus (Cordaites) which 

 we have heretofore been accustomed to associate with that phylum, 

 and, so far as our present knowledge goes, it undoubtedly stands 

 in the closest relations to Poroxylon. It is, however, improbable 

 that the two were in any sense conterminous, and it is altogether 



