204 PROF. F. W. OLIVER ON THE STRUCTURE 
seeds as definitely settled. The resemblances which they present with the seeds of 
recent Cycads have been pointed out by various writers and are quite evident. 
Nevertheless, in view of the close Cycadean affinity offered by the Lagenostoma- 
Lyginodendron Cycadofilices in respect of reproductive and vegetative organs alike, it 
seems quite probable that the Cordaiteze and Cycadaceve are less nearly related than has 
often been supposed. The former are an ancient group and retain relatively little of 
their Filicinean ancestry. ‘The likeness in seed-structure may perhaps be regarded as 
a parallel development in related, but not very closely allied, lines of descent. 
In the same way the agreement in details of seed-structure in Cycads on the one 
hand, and in Stephanospermum and similar forms on the other, should not be too 
closely pressed. For, in view of the position of Lagenostoma, the relation of Stephano- 
spermum to Cycas is more remote than may have appeared to be the case before the 
sorting-out process began. For, if the Cycadean seed be interpreted in terms of 
Lagenostoma, as it needs must be, it is difficult to resist the conclusion that the 
inner system of vascular bundles in the former is, in reality, integumental in character 
rather than nucellar. Consequently, the homology of this system of strands with 
the vascular mantle of a Stephanospermum, an homology implied by the present writer 
on a former occasion *, stands in need of revision. 
SUMMARY AND CONCLUSION. 
In whatever respects the seeds of Palmozoie age may have differed among themselves, 
as in the structure of the integument and in the position of the vascular system, they all 
show marked agreement, as originally pointed out by Renault +, in the possession of a 
pollen-chamber. This applies to both provisional groups of the Radiospermic series 
(Lyginodendrez and Stephanospermez) as also to the Platysperms (Cordaitex). And 
in this respect they contrast in marked degree with recent plants, always excepting such 
archaic Gymnosperms as are fertilised by spermatozoids. The pollen-chamber was 
evidently an adaptation of the distal portion of the macrosporangium, connected with 
the reception and storing of the microspores and the liberation of spermatozoids. 
Indeed, it is difficult to conceive any essentially different method by which a seed could 
have been evolved, assuming always, of course, that fertilisation by spermatozoids was 
retained. And it would be improbable in the highest degree that the particular 
conjunction of factors that led to the production of a pollen-tube should have operated 
contemporaneously with those others which produced the pollen-chamber. 
The old type of seed was of necessity, then, a complicated structure, as the micro- 
spores had to be collected and nursed, and provision made for the swimming of the 
spermatozoids. The pollen-chamber seems to have reached its zenith in Permo- 
Carboniferous times, and its decline was doubtless correlated with the evolution of 
pollen-tubes. The result we see in recent Conifers, with their simple reduced nucelli 
and siphonogamie methods. 
* Cf.* Annals of Botany,’ vol. xvii. p. 466. f ‘Flore carbonifère; p. 310, 
