The term Pteridospermatophyta (or "seed ferns") refers to several distinct groups of extinct seed plants (spermatophytes). The oldest fossil evidence of plants of this type is of late Devonian age, and they flourished particularly during the Carboniferous and Permian periods. Pteridosperms declined during the Mesozoic Era and had mostly disappeared by the end of the Cretaceous Period, though some fossil pteridosperm-like plants seem to have survived into Eocene times in Tasmania.
The concept of the pteridosperms goes back to the late 19th century when palaeobotanists came to realise that many Carboniferous fossils resembling fern fronds had anatomical features more reminiscent of the modern-day seed plants the cycads. The German palaeobotanist Henry Potonié coined the term "Cycadofilices" for such fossils, suggesting that they were a group of non-seed plants intermediate between the ferns and cycads. Shortly afterwards, the British palaeobotanists Frank Oliver and Dukinfield Henry Scott (with the assistance of Oliver's student at the time Marie Stopes) made the critical discovery that some of these fronds (Lyginopteris) were associated with seeds (Lagenostoma) that had identical and very distinctive glandular hairs, and concluded that the fronds and seeds belonged to the same plants. Soon, additional evidence came to light suggesting that seeds were also attached to the Carboniferous fern-like fronds Dicksonites, Neuropteris and Aneimites. Initially it was still thought that they were intermediate between the ferns and cycads, and especially in the English-speaking world they were referred to as "seed ferns" or "pteridosperms". Today, most palaeobotanists regard them as being only distantly related to ferns and that these names are misleading, but the names have nevertheless "stuck". Nowadays, four orders of Palaeozoic seed plants tend to be referred to as pteridosperms: Lyginopteridales, Medullosales, Callistophytales and Peltaspermales.
Their discovery attracted considerable attention at the time as the pteridosperms were the first extinct group of vascular plants to be identified solely from the fossil record. In the 19th century the Carboniferous Period was often referred to as the "Age of Ferns" but these discoveries during the first decade of the 20th century made it clear that the "Age of Pteridosperms" was perhaps a better description.
Later during the 20th century, the concept of pteridosperms was expanded to include various Mesozoic groups of seed plants with fern-like fronds, such as the Corystospermales and Leptostrobales. Some palaeobotanists also included seed plant groups with entire leaves such as the glossopterids (Arberiales) and Gigantopteridales, which was clearly stretching the concept. In the context of modern phylogenetic models, the groups often referred to as pteridosperms appear to be liberally spread across a range of clades and many palaeobotanists today would regard pteridosperms as little more than a paraphyletic 'grade-group'. One of the few characters that may unify the group is that the ovules were borne in a cupule, but this has not been confirmed for all "pteridosperm" groups.
So, does the concept of pteridosperms have any value today? Many palaeobotanists still use the term in an informal sense for the seed plants that are not angiosperms, coniferoids (conifers or cordaites), ginkgophytes or cycadophytes (cycads or bennettites). This is particularly useful for extinct seed plant groups, whose systematic relationships remain speculative; we can call them pteridosperms with no implications being made as to their systematic affinities. Also, from a purely curatorial or collecting perspective the term pteridosperms is a useful shorthand for describing the fern-like fronds that were probably produced by seed plants, which are commonly found in many Palaeozoic and Mesozoic fossil floras