The very numerous impressions of their fernlike foliage from the Palæozoic rocks indicate that the plants which bore such leaves must have existed at that time in great quantity. They must have been, in fact, one of the dominant types of the vegetation of the period. The recent discovery that so large a proportion of them were not ferns, but were seed-bearing plants, alters the long-established belief that the ferns reached their high-water mark of prosperity in the Coal Measure period. Indeed, the fossils of this age which remain undoubtedly true ferns are far from numerous. It is the seed-bearing Pteridosperms which had their day in Palæozoic times. Whether they led directly on to the Cycads is as yet uncertain, the probability being rather that they and the Cycads sprang from a common stock which had in some measure the tendencies of both groups. That the Pteridosperms in themselves combined many of the most important features of both Ferns and Gymnosperms is illustrated in the account of them given above, which may be summarized as follows:— Salient Characters of the Pteridosperms _G_=_Gymnospermic_ _F_=_Fernlike_ _F_ Primary structure of root. _G_ Secondary thickening of root. _F_ In _Heterangium_ and _Medullosa_ the _F_ solid centripetal primary wood of stele. _G_ Pits on tracheæ of primary wood. _G_ Secondary thickening of stem. _G_ Double leaf trace. _F_ Fernlike stele in petiole. _F_ Fernlike leaves. _F_ Sporangia pollen-sac-like. _F_ Reproductive organs borne directly on ordinary foliage leaves. _G_ General organization of the seed. Thus it can be seen at a glance, without entering into minutiæ, that the characters are divided between the two groups with approximate equality. The connection with Ferns is clear, and the connection with Gymnosperms is clear. The point which is not yet determined, and about which discussion will probably long rage, is the position of this group in the whole scheme of the plant world. Do they stand as a connecting link between the ferns on one hand and the whole train of higher plants on the other, or do they lead so far as the Cycads and there stop? CHAPTER XIII PAST HISTORIES OF PLANT FAMILIES VI. The Ferns Unfortunately the records in the rocks do not go back so far as to touch what must have been the most interesting period in the history of the ferns, namely, the point where they diverged from some simple ancestral type, or at least were sufficiently primitive to give indications of their origin from some lower group. Before the Devonian period all plant impressions are of little value, and by that early pre-Carboniferous time there are preserved complex leaves, which are to all appearance highly organized ferns.