The fossil record of ancient gymnosperms is surprisingly complete and provides good data from which to construct phylogenies.
Many botanists believe the progymnosperms are the most likely ancestral group from which the seed plants evolved. They have many features of the seed plants, but are still spore producers and are not themselves seed plants. They were important members of the vegetation from Middle Devonian through the Lower Carboniferous. Many were large trees with fern‐like leaves and probably formed forests in these early landscapes.
One feature of evolutionary significance that advances and separates the group from the ferns and trimerophytes is the bifacial cambium present in the progymnosperms; the group had a vascular cambium that produced secondary xylem and secondary phloem. A bifacial cambiumis is a characteristic of the seed plants and it appears for the first time in this group. The fossil progymnosperm wood resembles that of more modern conifers with tracheids and bordered pits.
Another Late Devonian group is the seed ferns, with leaves so like ferns that if no seeds are attached the fossils often are cataloged as ferns. This is an unnatural mixed group (like the protista) with no taxonomic ranking and not enough specimens to determine phylogenies well. One hypothesis derives the seed ferns from the progymnosperms in a lineage with no modern descendants. Another places one of the seed fern lines, the Medullosans, in a clade with the cycads—with good arguments for and against the arrangement. The group remains a puzzle.
This phylum derives its name from an extinct genus of trees that formed upland and swamp forests in late Carboniferous and Permian time. Current understanding of the phylogenies of the group place Cordaites near the cycads, but their lineage is still a matter of much discussion. They had long strap‐like leaves and produced two kinds of cones, pollen and seed, on separate branches. They had a well‐developed root system with secondary xylem. The stems also had secondary xylem surrounding a large central pith.
The bennettitaleans resemble modern cycads and for many years were placed in the seed fern melange. Some recent phylogenetic analyses—aimed at finding the origin of the angiosperms—place them in or near the angiosperm clade with ancestory still unknown. One bennettitalean, William‐soniella, from the Jurassic, had a bisexual strobilus—a rare feature in gymnosperms that mostly have separate organs for pollen and ovules. The microsporophylls were in whorls surrounding a center of several ovulate whorls and the whole was enclosed in bracts. In many ways, this is the structure of a flower before there were angiosperms on the scene.