XXXIX] TRISMERIA j-j 



forations which are so marked a feature in the latter. This may perhaps 

 be related to the large size and sappy nature of G. japonica, for Gwynne- 

 Vaughan found the stelar structure of G. calomelanos not far removed from 

 solenostely {I.e. p. 698). In this Trismeria more closely resembles this Fern 

 than it does G. japonica. The general conclusion is that, given a divided leaf- 

 trace, all these Ferns are essentiallx' solenostelic, but with occasional and 

 variable perforations, and with overlappings of the foliar gaps which are 

 dependent upon the leaf-arrangement. 



The sporangia of Trismeria are mostly inserted on the veins: some, 

 however, are seated on the general leaf-surface. The various ages are 

 intermixed. The sporangial stalk is three-rowed, and the annulus vertical 

 and regular, with a well-defined stomium. The tetrahedral spores give 

 spore-counts between the extremes of 49 and 35. Sorally in fact there is 

 nothing of importance to note beyond the slightly Acrostichoid state. It 

 will be a question for further enquiry how prevalent an Acrostichoid 

 development has been among the Gymnogrammoid Ferns. Evidence of it 

 has been found in Jamesonia and in Trismeria, but there is little sign of it 

 in other genera though it might occur in any of them. Attempts have been 

 made to trace Leptochilns from forms like Gymnogramme or Hemionitis: but 

 this is held in doubt by Frau Eva Schumann {Flora, Bd. 108, p. 249). She 

 finds little suggestion of that affinity in the external habit of Leptoehilus: 

 moreover, while the Gymnogrammoid spores are tetrahedral and naked, 

 those of Leptochilns are, like those of the Dryopteroid Menisciicm, kidney- 

 shaped, and have a perispore. This makes it appear improbable that such 

 developments as are seen in LeptocJiilus are of Gymnogrammoid origin. No 

 other Acrostichoid type suggests this affinity: in fact the Acrostichoid spread 

 of the sporangia seems in these Ferns to be potential rather than widel}- 

 realised. 



The wa)' has been opened by the comparative study of the relatively 

 primitive Gymnogrammoids for that of the central group. It is true that the 

 latter do not exhibit that variability of structure of the sporangia, nor the 

 short thick stalk and pear-shaped head, which are such marked features in 

 the former. But the general soral condition is here so far similar to theirs 

 that the central Gymnogrammoids may reasonably be regarded as advanced 

 derivatives of some similar type. In fact we may hold that they also 

 originated from a non-indusiate ancestry, in which at an early period the 

 production of the sporangia passed from an originally marginal position to 

 the surface of a widening blade : but that in the course of their further 

 development the sporangium had become progressively stabilised after the 

 usual Leptosporangiate type. 



