Gondwanaland towards the close of the Palaeozoic era may be compared with Alaska and Greenland at the present day, where glaciers and ice-sheets are bordered by an Arctic vegetation. One of the commonest Permo-Carboniferous plants on the southern continent was the genus Glossopteris (fig. 1), so named from the tongue-like form of the leaves, the larger of which reached a length of more than a foot. From a well-defined mid-rib are given off arching veins which by repeated unions form a fairly regular network. Glossopteris was formerly regarded as a fern similar in its fronds to the existing hart's tongue (Scolopendrium) but dif fering from it in the architecture of the venation ; it is now be lieved to be a member of the Pteridosperms (see above.) Seeds and leaves have not been found in organic union, but their fre quent association in the rocks and the discovery of certain other pieces of evidence favour the conclusion that the stems which bore Glossopteris leaves bore also seeds. Because of the abundance of Glossopteris fronds and pieces of the stems (V ertebraria) the flora of Gondwanaland is usually spoken of as the Glossopteris flora. Another common genus is Gangamopteris with leaves on the whole larger than those of Glossopteris and distinguished by the feeble development or absence of a mid-rib : the two forms of leaf are not always easy to separate. Both genera are almost certainly Pteridosperms. Schizoneura is also a characteristic south ern genus : more robust than Equisetum but similar to the horsetail in its jointed stems, it is distinguished from the northern hemis phere Calamites by its longer and broader leaves coalescent in varying degrees into a sheath which enveloped the foliage-shoots at each node. The genus Neuropteridium (Gondwanidium) (fig. 2) is represented by simple pinnate fronds superficially resembling those of some modern ferns and characterized by the large lobed leaflets ; from the lack of fertile leaves one suspects that it may be a Pteridosperm. Two sets of fronds have been called Neur opteridium, Permo-Carboniferous and Triassic fronds. It has re cently been suggested that, as there is no satisfactory evidence of generic identity, the older forms should be renamed Gondwan idium, Neuropteridium being reserved for Triassic fronds (fig. ii), which were probably borne on ferns and not on Pteridosperms. These four genera do not occur in the later Carboniferous or in the Lower Permian floras on the continents north of the Tethys sea. The Tethys was a broad sea, stretching across the world. The Mediterranean is its diminutive modern representative. The discovery in 1912 of Glossopteris and thin beds of coal by the heroic members of the second Scott expedition on the Beard more glacier, 3oom. from the South Pole, points to the existence of this plant far within the Antarctic circle. It also suggests the possibility that Glossopteris and perhaps some of its associates had their origin in the far South.
With the more typical and more abundant members of the Glossopteris flora are associated in some regions, plants generically indistinguishable from northern forms. The long strap-like leaves originally described as Noeggerathiopsis, similar in size and shape to the foliage of Yucca, appear to be the foliage of trees closely allied to the common northern genus Cordaites. Similarly a few species of Sigillaria and Lepidodendron are recorded from South Africa and South America in company with Glossopteris. Other genera which afford points of contact between the two great botan ical provinces are Psaronius, recorded from Brazil, a tree-fern allied to the tropical Marattiaceae in modern floras; the genus Sphenophyllum discovered in India, South Africa and Australia, Psygmophyllum with wedge-shaped, lobed leaves similar in form to those of Ginkgo biloba (the maidenhair tree) but not neces sarily a member of the same class ; also a few fern-like fronds which are probably Pteridosperms.