I 5 12 CORMOPHYTA. 



characterised by the bases of the leaf-stalks remaining attached to 

 the scars, as in many recent Tree-ferns. Other forms are Cyatheo- 

 pteris of the Bunter, and Thamnopteris ranging from the Permian 

 to the Keuper; the latter having persistent leaf-stalks. Finally, 

 Rhizomopteris of the Carboniferous, and Sphallopteris of the Bunter, 

 are based on specimens generally regarded as rhizomes of large 

 creeping ferns. The genus Palceopteris, from the Carboniferous, 

 which was long considered to be a fern, is named from a specimen 

 which appears to be the stem of one of the Cordaitece. 



Order 3. Rhizocarpe^e. — The Rhizocarps are distinguished 

 from the Ferns by the development of two kinds of spores termed 

 macrospores and microspores. Their young shoots may be either 

 straight, or circinate as in Ferns and Psilophyton (fig. 1378). 



Although the macrospores and microspores are true spores, as de- 

 veloping plants without fertilisation, yet they may be regarded as 

 incipient sexual elements, and thus throwing back the sexual differen- 

 tiation to an early stage. Thus the microspores, or male elements, 

 develop only male prothallia, which produce antheridea ; while the 

 macrospores, or female elements, develop female prothallia, which pro- 

 duce only archegonia. 



The four existing genera of this order are aquatic plants, which 

 may be simply floating, or may have a creeping rhizome. Of the 

 Salviniacece the rootless genus Salvinia is represented in the Upper 

 and Lower Miocene of the Continent, and also in the Laramie and 

 higher beds of America. Of the two existing genera Pilularia (Pill- 

 worts) and Marsiiia, constituting the family Marsiliacece, it is prob- 

 able that a species of the former occurs in the Upper Miocene of 

 CEningen, while Marsiiia is recorded from the Miocene of Oregon 

 in the United States, and also from the Lower Miocene of Ronzon 

 near Puy-en-Velay. It has been suggested that Sagenopteris, ranging 

 in Europe from the Rhaetic to the Lower Jurassic, and also oc- 

 curring, together with the allied Dactylopteris, in the Damudas of 

 India, may be more or less closely allied to the Marsiliacece. They 

 are plants of considerable size, with long-stalked leaves terminating 

 in a palmate expansion of four or more members. Marsilidium, of 

 the Wealden, has also been referred to the same family. 



Sir J. W. Dawson considers that in the early Palaeozoic the char- 

 acters afterwards separated in the Club-mosses, Horse-tails, and 

 Ferns were united in the Rhizocarps, and it will accordingly be 

 convenient in this place to notice certain Palaeozoic plants ap- 

 parently more or less closely allied to the Rhizocarps, some of 

 which should probably be included in the same order, while others 

 may be intermediate types connecting that order with the Equi- 

 setaceae and Lycopodiaceae. In the first place, as previously noted 

 (p. 1484), certain spherical bodies known as Sporangites occurring 



