1 1 1 
an archaic type of Seed from the Palaeozoic Rocks. 
8. The pollen-grains show a general agreement, and internal cells have 
been detected in all four species. 
In Physostoma and L. ovoides , bodies have been found both within the 
pollen-grains and also associated with them, which there are grounds for 
regarding as spermatozoids. The form of the supposed spermatozoids is 
identical in the two cases, and recalls that of Cycas , except that the sperms 
in the fossils are much smaller, and no trace of the band of cilia has been 
detected. 
The difference in size between the sperms of Physostoma and L. ovoides 
accords with the difference in the dimensions of their pollen-grains. 
These several points of agreement, comprehending all the regions of the 
seed, appear to be too numerous and striking to admit of interpretation 
except as the outcome of close affinity, i. e., community of descent. 
Whilst it seems evident that in this series of seeds of the Lagenostoma- 
group, the existence is disclosed of a number of closely-related species 
of Pteridosperms all occupying the same habitat, a difficulty is encountered 
when we try to point with any confidence to vegetative remains in our coal- 
nodules that could have belonged to the plants that bore these seeds. Up 
to the present time Lagenostorna Lomaxii , alone of the petrified seeds, has 
been definitely correlated with Lyginodendron oldhamiuml Among those 
preserved as impressions, several cases have come to light. Mr. Arber has 
referred Lagenostorna Sinclair ii and L . Kidstonii to fronds of the Spheno- 
pteris-type ; 2 and Monsieur Grand’ Eury has found strong grounds for 
referring other seeds of the same general character to Sphenopteris 
Dubuissonis? 
The position, therefore, is as follows: One petrified seed (Z. Lomaxii ), 
and at least three impressions, superficially in agreement with the seeds of 
the Lagenostorna-^ roup, have been referred to the frond-type Sphenopteris. 
We are still left with Physostoma, L. ovoides , and the Conostomas, all petri- 
factions from the same group, and — excepting for Heterangium — there are 
no species of Sphenopteris yet separated from Lyginodendron by anatomical 
characters to which they could be assigned. 
The close structural resemblance shown by Z. ovoides and Z. Lomaxii 
clearly points to Lyginodendron as their common source — indeed it has long 
been recognized that a vigorous analysis of the anatomical material that 
goes by the name of Lyginodendron oldhamium should lead to the dis- 
crimination of a group of related forms. The Burntisland species of 
Conostoma may be dismissed for the moment as probably the seed of 
Heterangium , in view of its common association in the nodules with the 
vegetative organs of this plant — to the total or almost total exclusion of 
1 Oliver and Scott, loc. cit. 
2 E. A. N. Arber, ‘ On some New Species of Lagenosto?na \ Proc. Roy. Soc., B., vol. lxxvi,p. 245. 
3 Grand’Eury, loc. cit. 
