viii MARATTIALES 303 



suppression of the wall in the antheridium of an Anthoceros 

 where only one antheridium is formed, there would be produced 

 at once an antheridium of the type found in Botrychium, and 

 by a further reduction of the division of the cover cell, by which 

 it remains but one cell thick, the type found in Marattia would 

 result. 



Such an origin of the antheridium of the Filicinese is, at 

 any rate, not inconceivable, while not so obvious perhaps as the 

 resemblances in the archegonium, and is simply suggested as a 

 possible solution of a very puzzling problem. 



The Marattiacese agree closely among themselves, and the 

 structure of the gametophyte is like that of the Ophioglossacese, 

 so far as the latter is known, and also offers most striking 

 resemblances to the Hepaticse. The long duration of the pro- 

 thallium, and its persistence after the sporophyte is independent, 

 as well as the long dependence of the latter upon the game- 

 tophyte, are all indications of the low rank of this order. The 

 sporophyte, while showing many points of resemblance to the 

 Ophioglossacese, still differs very much also, and in general 

 habit as well as the position of the sporangia comes nearer the 

 leptosporangiate Ferns. Of the Ophioglossacese, Helmintho- 

 stachys on the whole approaches nearest to the Marattiacese, so 

 far as the general character of the sporophyte is concerned. 

 The venation of the leaves and dehiscence of the sporangia are 

 very similar to Angiopteris, and the green sterile tips to the 

 sporangial branches hint at a possible beginning of the lamina 

 of the sporophylls in the Marattiacese. 



The synangia of Dancea show a certain analogy, at least, 

 with the sporangial spike of Ophioglossum, and it is possible 

 that a comparison might be made between the leaf of O. 

 palmatum, with its numerous sporangial spikes, and a 

 sporophyll of Dancea (see Campbell (26) ) . Both archegonium 

 and antheridium of Ophioglossum pendulum are strikingly 

 similar to those of the Marattiacese. 



While any relationship between these orders is necessarily a 

 remote one, nevertheless there are too many agreements in struc- 

 ture to make it at all probable that the Ophioglossacese and 

 Marattiacese have had an entirely independent origin. 



In seeking a connection with the leptosporangiate Ferns 

 there are two points where this is possible. The higher species 

 of Botrychium show an unmistakable approach to the leptospo- 



