Milro sperm it m compressum (Will.). 505 
Kidston, 1 for instance, points out that Cardiocavpon Lindleyi is almost 
invariably found associated with Cordaites principalis. We are confronted, 
however, with the curious fact that there seem to be many more species of 
Cardiocarpon than of Cordaites. A few years ago Grand’Eury 2 wrote, 
‘ On a institue au moins cinq fois plus d’especes de graines que de feuilles 
de Cordaites / As an example he mentions that, associated with leaves iden¬ 
tified with, or closely similar to, Cordaitespalmaeformis, Gopp., five species 
of seeds are found—in the Franco-Belgian basin, Cardiocarpus Lindleyi , 
Carr., C. Pitcairniae , Lind., C. cornutus , Daw., and in the Loire basin, 
Samaropsis fiuitans , Daw., and .S. forensis , Gr. Dawson 3 had previously 
noticed a similar disproportion between the small number of species of 
Cordaitean leaves, and the numerous species of seeds, occurring in the 
Canadian rocks with which he was concerned. Oliver 4 has drawn attention 
to a precisely analogous difficulty in the case of the Lagenostoma group of 
seeds. He writes, ‘ One petrified seed (L. Lomaxii), and at least three 
impressions, superficially in agreement with the seeds of the Lagenostoma 
group, have been referred to the frond-type Sphenopteris. We are still left 
with Physostoma, L. oroides , and the Conostomas, all petrifactions from the 
same group, and—excepting for Hetcrangium —there are no species of 
Sphenopteris yet separated from Lyginodendron by anatomical characters 
to which they could be assigned/ In the case of Cordaites , Grand’Eury’s 
explanation is that evolution in this group has proceeded principally in the 
direction of changes in the reproductive organs, the vegetative structure 
remaining unaltered owing to uniformity of climate and conditions. It is 
easy, however, to take an exaggerated view of the uniformity of anatomical 
structure within these groups. It is likely that in the future a rigorous 
analysis of the anatomical characters of the Lyginodendreae, and of the 
Cordaitales, will reveal the existence of many more species, distinguishable 
on vegetative grounds, than those we know at present. In the case of the 
Cordaitales this analysis has already begun, and the work of Scott and 
Maslen 5 on the subject is bringing to light many new anatomical types. 
On the other hand, it is possible that there may actually be some truth in 
the idea that there is a disproportion between the number of so-called 
Cordaitean seeds, and the number of species recognizable by their vegetative 
characters. Some of these seeds may perhaps prove to belong to other 
groups. Until 1905 no case was known in which seeds of the Platysperm 
type could be definitely attributed to any other plant than Cordaites. In 
that year, however, seeds with bilateral symmetry were discovered attached 
to two Pteridosperms, Aneimites 6 and Pecopteris Pluckeneti? The notion 
that every member of the Platyspermeae was necessarily a Cordaitean seed, 
was thus discredited. 
1 Kidston (’86). 2 Grand’Eury (’05 2 ). s Dawson (’ 91 ). 
4 Oliver (’ 09 ), p. in. 5 Scott and Maslen (’10). 6 White (’ 05 ). 7 Grand’Eury (’05 ! ). 
