in reality very distinct. Both male and female inflorescences
have frequently been found in connexion with leaf-bearing branches
(see restoration, fig. 30). The inflorescence is usually a spike
bearing lateral cones or catkins, arranged sometimes distichously,
sometimes in a spiral order. The investigation of silicified
specimens has, in the hands of Renault, yielded striking results.
A longitudinal section of a male Cordaianthus (the name applied to
isolated fructifications) is shown in fig. 31, A, Plate. The organ
figured is one of the catkins (about a centimetre in length) which
were borne laterally on the spike. Some of the stamens are inserted
between the bracts, in an apparently axillary position, while others
are grouped about the apex of the axis. Each stamen consists
of a long filament, bearing several erect, cylindrical pollen-sacs at
its summit (cf. fig. 31, B, Plate). Some of the pollen-sacs had
dehisced, while others still retained their pollen. The stamens are
probably best compared with those of Ginkgo, but they have also
been interpreted as corresponding to the male “flowers” of the
Gnetaceae. In any case the morphology of the male Cordaitean
fructification is clearly very remote from that of any of the Cycads or
true Coniferae, though some resemblance to the stamens of
Araucarieae may be traced. The female inflorescences vary considerably
in organization; in some species the axis of the spike bears solitary
ovules, each accompanied by a few bracts, while in others the lateral
appendages are catkins, each containing from two to several ovules.
In the catkin shown in longitudinal section in fig. 32, A, it appears
that each ovule was borne terminally, on an extremely short axillary
shoot, as in Taxus among recent Gymnosperms. The ovule consists
of an integument (regarded by some writers as double) enclosing
the nucellus. In the upper part of the nucellus is a cavity
or pollen-chamber, with a narrow canal leading into it, precisely
as in the ovules of Stangeria or other Cycads at the present day
(fig. 32, B). Within the pollen-chamber, and in the canal, pollen-grains
are found, agreeing with those in the anthers, but usually
of larger size (fig. 32, C). It was in this case that Renault first
made the exceedingly interesting discovery that each pollen-grain
contains a group of cells, presumably representing an antheridium
(fig. 32, C). Recent observations have completely confirmed
Renault’s interpretation of the facts, on which some doubt had been
cast. In the isolated seeds of Cordaitales and Pteridosperms,
pollen-grains are often found within the pollen-chamber, and the
pluricellular structure of these pollen-grains has been repeatedly
demonstrated. In the light of our present knowledge of Ginkgo
and the Cycads, there can scarcely be a doubt that spermatozoids
were formed in the cells of the antheridium of the Cordaitean
pollen-grain and that of other Palaeozoic Spermophyta; the antheridium
is much more developed than in any recent Gymnosperm,
and it may be doubted whether any pollen-tube was formed. The
morphology of the female inflorescence of Cordaiteae has not yet
been cleared up, but Taxus and Ginkgo among recent plants appear
to offer the nearest analogies. Much further investigation will
be needed before the homologies between Cordaitean cones and
the fructifications of the higher Cryptogams can be established.
Anatomically the connexion of the family with the Pteridosperms
(and through them, presumably, with some primitive group of
Ferns) seems clear, but we have as yet no indications of the stages in
the evolution of their reproductive organs. The class Cordaitales
extends back to the Devonian, and it must be borne in mind that
our knowledge of their fructifications is practically limited to
representatives from the latest Palaeozoic horizons.
Isolated fossil seeds are common in the Carboniferous and Permian strata; in all cases they are of the orthotropous type, and resemble the seeds of Cycads or Ginkgo more nearly than those of any other living plants. Their internal structure is sometimes admirably preserved, so that the endosperm with its archegonia is clearly shown (fig. 32, D). It is a curious fact that in no case has an embryo been found in any of these seeds; probably fertilization took place after they were shed, and was followed immediately by germination. There is good evidence that many of the seeds belonged to Cordaitales, especially those seeds which had a flattened form, such as Cardiocarpus, Cycadinocarpus, Samaropsis, &c. Seeds of this kind have been found in connexion with the Cordaianthus inflorescences; the winged seeds of Samaropsis, borne on long pedicels, are attributed by Grand’ Eury to the genus Dorycordaites. Many other forms of seed, and especially those which show radial symmetry, as for example Trigonocarpus, Stephanospermum and Lagenostoma belonged, as we have seen, to some of the plants grouped under Pteridospermeae, though other Pteridosperms had flattened seeds not as yet distinguishable from those of Cordaitales. The abundance and variety of Palaeozoic seeds, still so often of undetermined nature, indicate the vast extent of the spermophytic flora of that period.
The modern Gymnospermous orders have but few authentic representatives in Palaeozoic rocks. The history of the Ginkgoales will be found in the Mesozoic section of this article (see also Gymnosperms); their nearest Palaeozoic representatives “were probably members of the Cordaitales, an extinct stock with which the Ginkgoaceae are closely connected” (Seward). Remains referable to Cycadophyta, so extraordinarily abundant in the succeeding period, are scanty. The curious genus Dolerophyllum (Saporta) may be mentioned in this connexion. This genus, from the Permo-Carboniferous of Autun, is represented by large, fleshy, reniform leaves or leaflets, with radiating dichotomous venation; the vascular bundles have in all respects the structure of those in the leaves of Cycads or Cordaiteae. The male sporophylls are similar in form to the vegetative leaves, but smaller; sunk in their parenchyma are numerous tubular loculi, containing large pollen-grains, which are pluricellular like those of Cordaites; the female fructification had not yet been identified with certainty. The curious male sporophylls may perhaps be remotely comparable to those recently discovered in Mesozoic Cycadophyta, of the group Bennettiteae. Some leaves of Cycadean habit (e.g. Pterophyllum, Sphenozamites) occur in the Coal Measures and Permian, and it is possible that the obscure Coal Measure genus Noeggerathia may have Cycadean affinities. A fructification from the Permian of Autun, named Cycadospadix milleryensis by Renault, appears to belong to this family.
Now that the numerous specimens of wood formerly referred to Coniferae are known to have belonged to distinct orders, but few true Palaeozoic Conifers remain to be considered. The most important are the upper Coal Measure or Permian genera Walchia, Ullmannia and Pagiophyllum, all of which resembled certain Araucarieae in habit. In the case of Walchia there is some evidence as to the fructifications, which in one species (W. filiciformis) appear to be comparable to female Araucarian cones. There are also some anatomical points of agreement with that family. It is probable, however, that under the same generic name very heterogeneous plants have been confounded. In the case of Ullmannia the anatomical structure of the leaf, investigated by Solms-Laubach, proves at any rate that the tree was Coniferous.
There is no proof of the existence of Gnetaceae in Palaeozoic times. The very remarkable plumose seeds described by Renault under the name Gnetopsis are of uncertain affinity, but have much in common with Lagenostoma, the seed of Lyginodendron.