208 ^ D- H. Scott. 



Eur}' also observed in Linopteris discs which he reg-arded as probably 

 of the same nature, but in no case, as 3"et, are the indications per- 

 fectlj' satisfactorj'. 



In the Neuropterideae, as in the Lyginodendreae, we are dealing 

 with plants of which the anatomical structure is known, at least in 

 certain representatives. We are therefore able to point to definite 

 structural characters, quite apart from the habit, w^hich indicate 

 affinity with the Ferns, as w^e have already shown in the case of 

 the Lyginodendreae. In the anatomy of the Neuropterideae with the 

 Medullosa type of stem, the most Fern-like feature is the vascular 

 system, wiiich, in its primary ..polj'stelic" arrangement is essentially 

 Filicinean, and was compared by Weber and Sterzel (1896) with that 

 of Fsaronius. This character, it is true, becomes disguised as secondarj'' 

 growth proceeds, but the original ground-plan of the structure is of 

 unmistakable significance.^) The leaves, apart from the habit and 

 venation, are on the whole of C'ycadean structure and not, in them- 

 selves, much more Fern-like than those of the recent Stangeria. In 

 the new genus SufcUffia the anatomy of the stem, peculiar as it is, 

 can only be compared wdth that of a protostelic Fern, while the 

 concentric foliar bundles accentuate the Filicinean relationship. The 

 discovery of the Sufdifßa type of structure renders it probable that 

 dialystely arose within the family MeduUoseae, and tends to separate 

 the latter further from the Lower Carboniferous Cladoxjieae; in this 

 curious group (Solms-Laubach, 1896) there is the same combination 

 of dialj'stely with secondary growth which we find in MeduUoseae, 

 but the arrangement of the steles, and the nature of the tracheides, 

 not to mention other characters, are quite different, and it seems 

 most probable that the two families represent parallel lines of deve- 

 lopment. The C'ladoxj'leae have been compared with Botryopterideae, 

 especially Asterochlaena , and an affinity in this direction appears 

 highly probable. There is nothing, as yet, to indicate the particular 

 group of primitive Ferns from which the MeduUoseae themselves 

 may have sprung, but on anatomical grounds it seems not unlikely 

 that they and the Lyginodendreae may have had a common origin 

 from simple protostelic forms. 



Other Pteridospermeae. 



There are at least two cases in which seeds have been found in 

 actual connection with Palaeozoic Fern-like fronds, where we have as 

 yet no clue to the internal structure. 



^) See, however, p. 212. 



