A PROBABLE ORIGIN OF EXISTING FLOWERS. 41 



been long known as resembling the fronds of living Ferns ; but an 

 interesting fact is that only a certain number has been proved to be 

 true Ferns ; for many have been discovered to bear naked seeds, and 

 therefore they must be excluded from the Fern family. These "seed- 

 plants " are more nearly allied to certain of the Fern families which at 

 the present day have comparatively few genera. Thus, besides the 

 foreign families Marattiaceae and Schizaeaceae, we have three native 

 Ferns in England which show alliance with the very ancient types. 

 They are the Royal Fern {Osmunda regalis), the Adder 's-tongue {Ophio- 

 glossum vulgatum) , and the Moonwort {Boirychitun Lunaria) . 



Our commonest Ferns are a later product of Evolution. Thus, the 

 sporangium or spore-case of most existing Ferns has a strong, nearly 

 complete ring of large cells, called the annuhis, which contracts when 

 the spores are ripe, and so bursts open the spore-case in order to 

 liberate the asexual spores. This is wanting in the two last mentioned, 

 and is very imperfect in the first. 



One of the best-known members of the Pteridosperms is named 

 Lygtnodendron. The restored illustration as a frontispiece to Dr. 

 Scott's Studies is exactly like a tree-fern, with much-divided 

 fronds ; but it has been discovered with the seeds upon it. The ovules 

 are of a peculiar structure, but very closely resemble those of exist- 

 ing Cycads ; while the stamens have the anther-cells pendulous under 

 a sort of peltate scale, resembling that of the Yew, Araucaria, and 

 Zamia, alHed to Cycas, as well as to the spore-cases of the crypto- 

 gamous Horsetail {Equisetum), 



The ripe seed of Lyginodendron consists of a thick outer coat, and 

 the embryo, presumably, was in the middle, though it appears to have 

 disappeared from the fossil specimens. However, well-preserved dicoty- 

 ledonous embryos are known in later fossils (e.g. Cordaianthits). 



With regard to the existence of a carpel, nothing definite is known 

 in fossil plants. The ovule had one coat corresponding to the ' ' secundine ' ' 

 of living plants, and this was prolonged into a tube above the nucleus 

 forming the "pollen-chamber" at the bottom. The tube sometimes 

 expanded at the top, imitating a stigma, only it was open. I found a 

 curious imitation of this in some self-fertilizing plants of the desert 

 near Cairo, as in Pulicaria arahica ; the stigma was open at the top, 

 leading to the ovules within in a sort of mock pollen-chamber. Some- 

 thing like one, but closed up, is seen in the base of the style in 

 Casuarina. A good example of the trumpet-like ovular coat is seen 

 in Welwitschia and Ephedra ; in Gnetum the ovular membrane does 

 not extend above the nucleus. 



Our own Juniper has three ovules clustered together, each being 

 in front of a scale ; these scales leave an opening at the top, being 

 sHghtly coherent below. When the ovules have been pollinated, the 

 coherent scales close up, and so form the fleshy " Juniper berry." Each 

 scale therefore might be called a " carpellary " scale, but the Juniper is 

 strictly gymnospermous. 



In Ephedra and Welwitschia, of the family Gnetaceae, there is an 



