6o2 MOSSES AND FERNS chap. 



strong evidence of the geological record that the Eusporangiatse 

 were the prevailing types in the earlier formations, and have 

 been supplanted by the more specialised Leptosporangiatse in 

 more recent times, is reasonably conclusive. 



Owing to the very small number of living Eusporangiatse. 

 the relationships of these among themselves and to the Lepto- 

 sporangiatse are difficult to determine. From the frequent oc- 

 currence of dimorphic leaves among the older fossil types of 

 Ferns, as well as on grounds of comparative morphology, the 

 type of leaf in the Ophioglossacese is probably to be considered a 

 more primitive one than that of the living Marattiacese. Of 

 the existing genera of Marattiacese, Dancea is the only one in 

 which the sporophylls differ in form from the sterile leaves, 

 and this dimorphism probably indicates that on the whole it is 

 the most primitive of the living genera. Whether the extreme 

 type of synangium found in Dancca is older than the nearly free 

 sporangia such as those of Angiopteris, has been questioned, as 

 both types are found among the Paleozoic Marattiacese; but 

 the greater specialisation shown in the latter type indicates that 

 it is of more recent origin. There is a possibility that the two 

 types represent two lines of development originating from dif- 

 ferent stocks comparable to Ophioglossum and Hehnintho- 

 stachys among the Ophioglossaceae. The occurrence of Ferns 

 of unmistakable Marattiaceous affinity, but with fertile leaf 

 segments completely covered with free sporangia like those of 

 Botrychunn or Osnmnda supports this view. 



While in such species of Botrychhtm as B. Viginianiim, 

 there is a strong resemblance in the tissues to the lower lepto- 

 sporangiate Ferns, it is not so marked, on the whole, as those 

 in the Marattiaceae, which probably are nearer the Leptosporan- 

 giatse, and probably have given rise directly to them. 



The homosporous Leptosporangiatse or Filices constitute a 

 very natural order. The Osmundacege are without much ques- 

 tion the most primitive members of the order, this being indi- 

 cated both in the gametophyte and sporophyte. While they 

 show certain points of resemblance to Hclminthostachys and 

 Botrychunn, their affinities seem to be rather with the Marat- 

 tiacese, and presumably they have arisen from some Palaeozoic 

 Marattiacese with free sporangia borne upon special leaf seg- 

 ments. It is not impossible that two others of the lower fami- 

 lies the Schizseacese and Gleicheniacese, may have originated 



