— 
AND AFFINITIES OF STEPHANOSPERMUM, 395 
Very few instances of early seed-like struetures without pollen-chambers are at present 
known. Lepidocarpon alone has been at all fully described *; but even here nothing 
has been ascertained in regard to either pollination or fertilisation. In view of this 
paucity of data, a comparison of the two seed-types is hardly possible. It may he 
observed, however, with some confidence, that the Lepidocarpon-type has left no seed- 
bearing progeny, whilst there is at least equal probability for the assertion that the 
whole of existing Gymnosperms have originated from an ancestral type characterised 
by the possession of a pollen-chamber. 
One reason for the predominance of the pollen-chamber may reside in the water- 
relations. The Pteridophytic alternation of generations, though.a step in advance 
of the Algal condition, was not the ideal state for a dry-land flora. The early seed- 
habit, in its initiation, perhaps, largely a xerophilous adaptation, contained the germ of 
much else beside. "The water-relations in question are very evident in the seeds of the 
Stephanospermwum-type, where the ample sheath of transfusion-tracheides would provide 
such moisture as may have been required. But as yet the nutritive possibilities had 
not been realised in the way in which they are realised in recent plants. Already, before 
fertilisation, the seed seems to have reached maturity, and it is diffieult to conceive how 
such a structure could have enlarged and accommodated rich supplies of food-material 
for the embryo. ‘The real ovular stage was intercalated at a later date. 
In point of fact our information as to the early developmental stages of Palwozoic 
seeds is exceedingly meagre. Renault has described young Cordaitean seeds, and these 
were already pollinated; but we know nothing as to the time of pollination in 
Stephanospermum. The mature seeds contain pollen; two slightly younger stages, one 
figured in the present paper (Pl. 42. fig. 13), are without pollen; but these preparations 
are neither numerous nor favourable enough to justify any confident iuference as to the 
time of pollination. Nor, on the other hand, can we assume that these seeds were 
pollinated, like Cordaites, whilst quite small; for the latter type is immensely old, and its 
seeds are not necessarily on quite the same plane of evolution as Stephanospermum. 
If it be recognised that the nearest affinities of recent Cycads lie with the 
Lyginodendron-branch of the Cycadofilices rather than with the Stephanospermer, 
it almost follows that the latter have died out and are unrepresented in the existing 
flora. The apparent perfection of the vascular mantle in Stephanospermum may have 
proved an obstacle to further development. And if this type of seed should be shown 
to have belonged to a Medullosa-type of stem, have we not in their conjunction an 
over-elaborated organisation marked out for subsequent decline ? 
The phylogenetic history of so specialised a seed-type as Stephanospermum is extremely 
obscure. It may be conjectured that the pollen-chamber corresponds to some modi- 
fieation of sporangial dehiscence, whilst the tracheal mantle finds its analogue in an 
unassigned fossil sporangium with tracheal lining from the same horizon}. The 
morphological nature of the integument is no less obscure. 
* D. H. Scott, “ The Seed-like Fructification of Lepidocarpon &c.," Phil. Trans. series B, vol. 194, p. 291. 
+ "A Vascular Sporangium,” New Phytologist, vol. i. 1902, p. 60.—A type of sporangium that may be termed 
conveniently Tracheotheca. 
912 
