336 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1918. 



as large as the main stem on which they were arranged in a two-fifth 

 spiral. The roots (Kaloxylon) were partly adventitious, the stem 

 cortex was fibrous and reticulate and was clothed with spines and 

 capitate glandular hairs. The seeds (Lagenostoma) were borne in 

 lobate cupules while the pollen was on the other and more reduced 

 fronds in rosettes of six or seven fusiform bilocular sporangia (Cros- 

 sotheca). The accompanying restoration gives an idea of* the habit 

 of Lyginopteris, but fails to show the forking of the frond stipes. 



The absence of an embryo has led to questioning the use of the term 

 seed for Lagenostoma and other Paleozoic forms. They are indubi- 

 table seeds, however, and the absence of an embryo may be explained 

 by the resting period having occurred after pollination, while embryo 

 formation was postponed until after the seeds had been shed, and 

 immediately preceded germination. 



Lyginopteris was monostelic with a large pith containing sclerotic 

 tissue. Primary wood consisted of from five to nine collateral 

 strands. In all but the most immature stems there is a broad zone of 

 secondary wood of pitted tracheids and medullary rays. The cam- 

 bium was persistent and is sometimes petrified, as is the phloem. The 

 cortex comprises a thin periderm and an inner, poorly preserved soft 

 cortex and an outer cortex characterized by radial bands of fibrous 

 tissue. Leaf traces are mesarch and double. Anomalous features 

 are the occasional formation of inverted secondary wood by the in- 

 trusion of the cambium through a foliar gap. 



The seeds, 5 or 6 minims in length, we're borne in a lobed cupule 

 and were orthotropous and radially symmetrical, with a single in- 

 tegument confluent with the nucellus except distad. The free part 

 forms a plug and the pollen chamber was hence reduced to a conical 

 slit. The integument was supplied by nine vascular strands which 

 ran to the apex, which formed a fluted dome or radially septate canopy 

 at the apex of the barrel-shaped seed. 



The microsporangia were found in connection with vegetative 

 fronds by Kidston. They are of a type known as Crossotheca and 

 consisted of a rosette of six or eight bilocular fusiform sporangia. 

 Other types of microsporangia may well have been present in dif- 

 ferent species of Lyginopteris, as, for example, those called Telan- 

 gium, in which the sporangia are concrescent proximad. A second 

 genus of seed ferns which is fairly well known is Heterangium, 

 which had long angular slender stems with large and graceful forked 

 (Sphenopteris elegans Brongn.) fronds arranged in a three-eighths 

 spiral. While the evidence is less conclusive than in Lyginopteris 

 there are good reasons for considering the seeds known as Sphaeros- 

 toma as those of Heterangium. Heterangium had a monostelic stem 

 like Lyginopteris, but the pith was replaced by mesarch primary 

 wood, as in some recent ferns. The secondary zone was thin. The 



