PALEOBOTANY BERRY. 335 



last decade. It is needless to dwell on the immense advantages which 

 seed bearing confers on the plants which have acquired this habit. 

 The mere fact that seed plants are the dominant existing plants is 

 sufficient proof of this. 



Over a generation ago Stur suggested that the fronds of Neurop- 

 teris, Alethopteris, and other form genera of fern-like fronds exceed- 

 ingly common in the Paleozoic were probably related to the cycads 

 since they were never found in a fruiting condition like normal 

 ferns. Subsequently the anatomy of certain petrified stems showing 

 a combination of fern and cycad characters led to the proposal of a 

 group, the Cycadofilicales, for these intermediate types. Meanwhile, 

 the structure of a considerable number of Paleozoic seeds suggestive 

 of cycads, ginkgos, and gnetales had been described, but beyond corre- 

 lating some of them with Cordaites little was known of the plants 

 which bore the majority. In 1903 Oliver and Scott succeeded in 

 proving that certain seeds (Lagenostoma) were borne on fronds of 

 the Sphenopteris type and these in turn were attached to stems known 

 as Lyginodendron. This discovery stimulated an interest in the sub- 

 ject and a succession of discoveries followed, so that enough is now 

 known to warrant considering a large number of the supposed Paleo- 

 zoic ferns as Pteridosperms or seed ferns. The manifestly primitive 

 characters show in one feature or another, such as the more or less 

 free nucellus, the complex vascular supply of the seeds, their total 

 lack of an embryo, the fact that both megasporangia and microspo- 

 rangia were borne upon but slightly modified foliage leaves of de- 

 compound fernlike fronds, and various recondite histological charac- 

 ter-, justify regarding the seed ferns as representing a distinct phy- 

 lum — the Pteridospermophyta. They were gymnospermous in habit 

 and some students regard them as a subordinate class of gymno- 

 sperms — a taxonomic term that has outlived its usefulness for other 

 than descriptive purposes. 



'j 'h 3 Pteridospermophyta may be tentatively characterized as plants 

 wit i the habit and, to a large extent, with the anatomical features of 

 ferns, but differing in producing integumented megasporangia or 

 seeds borne on the usually but slightly modified fernlike foliage and 

 never aggregated into true strobili ; having secondary thickening in 

 both stems and roots. 



From the rapidly increasing contributions to the knowledge of the 

 Pteridosperms it will suffice to describe a few of the better known 

 forms. Among these the best know T n is Lyginodendron, or Lyginop- 

 teris as it is more properly called. Lyginopteris represents a group 

 of species with slender, scrambling, mostly unbranched stems of con- 

 siderable length bearing large forked decompound fronds upward of 

 a meter in length and long known under the name of Sphenopteris 

 Hoeninghausi, whose persistent petioles (Rachiopteris) were almost 



