644 STUDIES IN FOSSIL BOTANY 



seed-structure, and may prove to be a valuable indica- 

 tion. At present, however, the case is too isolated for 

 any phylogenetic conclusions to be drawn. 



There can be little doubt that the Pteridosperms 

 have some affinity with the Marattiaceae in the widest 

 sense, though the connection between them may be 

 indirect and impossible to trace accurately in the 

 present state of our knowledge. We have already 

 considered in some detail the evidence for the presence 

 of Marattiaceae in the Carboniferous Flora (Chapter 

 VIII.). The probability is decidedly in favour of 

 their existence as a group of some importance at 

 that time, though the question is by no means without 

 difficulty (see p. 468). In any case the structure of 

 the Pteridospermous male fructification Crossotheca 

 (p. 396), and still more that of the stamens in the 

 Mesozoic Bennettiteae (p. 586), indicates affinity between 

 the older Seed-plants and Ferns of the Marattiaceous 

 type, and heightens the probability that the latter are 

 an ancient race. There are thus good grounds for the 

 view expressed by Mr. Kidston that the Pteridosperms 

 and Marattiaceae had a common origin. This hypothesis 

 is not inconsistent with the idea of a relation between 

 the Pteridosperms and the Botryopterideae, for, as we 

 have seen, possible links between the latter family and 

 Marattiaceae are to be found in Corynepteris and other 

 genera (p. 344). 



In the absence of any evidence of heterospory 

 among Palaeozoic Ferns, it appears that the gap 

 between any known group of Ferns and the related 

 Seed-plants must be a wide one. The general conclu- 

 sion, however, that the Pteridosperms belonged to the 



