DICKSONIKAE 591 



runs round two sides, but at the point (*) it appears to stop, but it does not 

 do so actually : it merely curves round upon the peripheral face, and dis- 

 appears behind the body of the sporangium. That the annulus is really 

 a continuous one is seen from Fig. 329 E, which represents the peripheral 

 face : it is very irregular, especially at the base of the sporangium, and 

 consists of a large but not definite number of rather narrow cells ; together 

 they form a twisted hoop, so disposed that the distal end of the hoop is 

 curved in the direction of the apex of the receptacle, and this curvature 

 makes the sporangium a peculiarly difficult object to represent in a drawing. 

 The thickening of the walls is greater at the distal part of the annulus than 

 at the proximal, while at one side or the other is a part not strictly defined, 

 where rupture will take place. There is no definite stomium, and though 

 rupture usually occurs about the region to the right in Fig. 329 E, the 

 actual point of dehiscence may vary. 



The features thus noted mark off the annulus of Thyrsopteris as one of 

 the least specialised among Leptosporangiate Ferns : the inequality of its 

 thickening suggests a comparison with Loxsoma. It seems probable that, 

 while showing clear points for comparison with Dicksonia, it has been derived 

 from a type with a completely indurated annulus and median rupture : that 

 this became modified in consequence of the close packing of the sporangia 

 in the sorus, which would interfere with a median dehiscence : that a lateral 

 rupture had been adopted, but the sporangium not definitely specialised for it. 



The anatomy of Thyrsopteris is very imperfectly known. The leaf shows 

 a vascular system, with a few separate bands arranged in the usual horse- 

 shoe outline ; in fact, the structure suggests similarity with Dicksonia. There 

 are no data as to the internal structure of the axis. 



From the known characters it would appear that the relationship of 

 Thyrsopteris is primarily with Dickso?iia, but in some remote degree also 

 with the Hymenophyllaceae : there appears little reason to relate it to the 

 Cyatheaceae, as has been suggested by various writers. It is, however, 

 best placed as a separate family, on account of the peculiar characters of 

 the sporangium and annulus : these show less perfect differentiation than 

 of the Ferns named, while an archaic character is indicated by the numerous 

 cells of the annulus, and / the imperfect localisation of the point of 

 rupture. 



DICKSONIKAE. 



The family of the Dicksonieae included, as arranged in Hooker's Sy?wpsis 

 Filicu m, six genera ; but of these the affinity of Onoclea appears to be rather 

 with the Cyatheae, while Hypoderris, Woodsia, Sphaeropteris, and Deparia 

 differ from Dicksonia itself not only in the position of the sori, which are 

 superficial (except in Deparia 1 ), but also in the fact that the various ages 



1 'Deparia Aloorei, a fern in which the sori are mostly marginal, shows occasional 

 superficial sori also : but these are upon the upper surface, and the case is perhaps com- 

 parable with that of Aspidiuin anomaium. Hk. and Arn : see p. 117. 



