MARATTIACEAE 529 



The shoot of the Marattiaceae, && a whole, being of a usual Fern-type, 

 it will be considered in relation to the theory of the strobilus at the con- 

 clusion of the Filicales. Meanwhile a comparison of the characters of the 

 known representatives, modern and fossil, may be made, and this not 

 only of the mature structure, but also of the details seen in the seedlings 

 of the living forms. Such a comparison gives some foundation for an 

 opinion as to the nature of the stock from which the family may have 

 sprung. In the first place it had an upright, radially constructed shoot, as 

 indicated by the fact that no dorsiventral fossil stem of this affinity has 

 been described, while those which exist among the living genera are probably 

 derivative : moreover the embryos are all upright, and radially constructed. 

 Presumably it had a protostelic structure of the axis, as indicated by the 

 simple anatomy of the young seedlings of the living genera. From this pro- 

 tostele sprang the leaf-traces, at first without leaf-gaps, as is still seen in some 

 living seedlings : the formation of leaf-gaps probably followed early as the stele 

 dilated and became medullated. Each leaf-trace itself was a single strand, 

 as is seen in the fossils even in the mature shoot, but only in the seedlings 

 of the living genera : this would suggest some simple form of leaf in the 

 ultimate parentage. The arrangement of the leaves was on a radial plan, 

 but was probably simpler than in the living forms : this is indicated by the 

 occurrence of early fossils with regular orthostichies of leaves. The root- 

 system was endogenous, and after traversing the cortex with a more or 

 less lengthy course, the roots emerged at the surface, forming sometimes 

 a supporting external felt. The arrangement of the sori on the leaf was 

 on the plan of a simple row on either side of the midrib, and the form of 

 the sorus circular. The relation of the sporangia was probably synangial, 

 and their number in each sorus small, or liable to be reduced to a solitary 

 one. The individual sporangia were large, the spore-output numerous, and 

 the mechanism of dehiscence simple, or altogether absent. 



These characters, which comparison would indicate as those of the 

 Marattiaceous ancestry, show convergence in many points of form and 

 structure towards the apparently distinct series of the Botryopterideae. As 

 regards the reproductive organs also, it is to be remembered that certain 

 Ferns referred to a nearer relation with the Botryopterideae have a more 

 or less pronounced groupirig of the sporangia into sori : this is suggested 

 in Zygopteris itself (Fig. 272), and it is a marked feature in Corynepteris 

 (Fig. 273). Thus the two families are not absolutely distinct even in this 

 respect. It is probable that they represent divergent branches from some 

 common primitive stock. 1 



1 As regards relation to Pteridosperms, see Kidston, Phil. Trans., 1906. 



2 L 



