902 PROF. F. W. OLIVER ON THE STRUCTURE 
the smallest of the seeds described by Brongniart *. It is three-ridged, but appears to 
resemble Tripterospermum (with which, no doubt, it was closely allied) in the possession 
of a sarcotesta, of which traces are present. 
Of the remaining genera of Radiosperms, the seed Polylophospermum, Brongn. t, may 
be cited as possessing, in essentials, a Stephanospermum-organisation, 7. e. a vascular 
nucellus and a large broad pollen-chamber in conjunction with radial symmetry. The 
testa is complicated by chalazal and mieropylar expansions, in addition to longitudinal 
ridges, whilst an outer fibrous or succulent **sarcotesta ” enveloped the stone of the 
seed. 
Beside this seed, Codonospermum, Brongn. t, naturally takes its place. As Renault 
has shown $, this seed also possessed a sarcotesta. 
In addition to the above, the ridged seeds Colpospermum, B. Renault , Ptychotesta, 
` Brongn. Y, Polypterospermum, Brongn.**, and Hexapter ospermum, Sg TT, and 
perhaps others, may be included as possessing the same general organisation as 
Stephanospermum, and as being marked out for retention in the same group. 
To summarize the remarks, it may therefore be suggested that the following genera 
of Paleozoic seeds, having a like organisation, possess a bond of real affinity :— 
Stephanospermum, Aitheotesta, Trigonocai Hos Tripterospermum, and the 'Trigonocarpons 
of Hooker & Binney and Williamson $1, Polylophospermum, Codonospermum, Colpo- 
spermum, Ptychospermun, Polypterospermum, and Hexaspermum, It is these forms, at 
present largely recruited from the French Permo-Carboniferous, that constitute the 
Stephanospermum-group of the Radiosperms. 
The main features of this group are: radial symmetry; a nucellus traversed 
peripherally by a mantle, network, or distinct strands of tracheal elements; a relatively 
broad pollen-chamber ; a stony testa with a tendency to be produced into transverse or 
longitudinal ridges, &c.; an outer fleshy or fibrous sarcotesta. 
In the case of the type-genus, Stephanospermum, apart from the apex of the seed, 
evidence for the existence of the sarcotesta is lacking; but when the structure of this 
seed is considered as a whole (cf. p. 363), we can hardly avoid the conclusion that it 
resembled its contemporaries in this feature also. 
During the passage of the present paper through the press, strong evidence has been 
adduced for referring the English Lower Coal-Measure seed, Lagenostoma Lomasi, to 
the well-known fossil plant Lyginodendron oldhamium $9. Now this seed, and the 
* Brongniart, * Graines fossiles silicifióes,! p. 94. t Brongniart, loc. cit. pl. C. figs. 6, 7, 8. 
€ Brongniart, loc. cit. pl. C. figs. 10, 11, 12 ; also Renault, * Flore fossile d'Autun et d'Épinae, p. 393, and pl. 87. 
$ Renault, loc. cit. p. 394. || Renault, loc. cit. p. 400. 
{| Brongniart, loe. cit. pl. B. figs. 10, 11 ; also Renault, loc. cit. p. 401. 
+: Brongniart, loc. cit. pl. C. figs. 1, 2, 3. tt Brongniart, loc. cit. pl. B. figs. 12, 13, 14. 
Xt When the time arrives for a revision of these genera of seeds, the name 7'rigonocarpus may be appropriately 
Bechet with the Hooker and Binney specimens on the ground that the description of these peti ‘ified specimens 
holds unquestioned priority in the field of investigation of seeds with structure preserved. 
$8 Oliver, F. W., and D. H. Scott, * On Lagenostoma d the Seed of Lyginodendron," Proc. Roy. Boc. 
vol. lxxi. (1904) p. 477. 
