140 Scientific Proceedings, Royal Dublin Society. 



As Nathorsfc himself states, Goppert's diagnosis is based on a fragment showing 

 three or four pinnae only, and more abundant material would have shown him 

 the aphlehioid pinnules. Plate YIII. is a reproduction of a photograph of the 

 specimen just mentioned. It is labelled " Ci/elopteris hihernica, from the 

 Devonian beds of Grlanmire, Co. Cork," though the counterpart slab is labelled 

 " from the Carboniferous limestone," both being presented by Sir R. Griffith. 

 The specimen is really Arc/iceopteris Roemeriana, Gopp. Its smaller, non- 

 overlappiug, stalked pinnules are its distinguishing feature. Another 

 important character is, it seems to me, the more erect habit of the pinnae. 

 They make a sharper angle (45°) with the raohis than the pinnae in A. 

 hihernica. The difference in habit suggests that A. Roemeriana was, in its 

 habitat, more exposed to rain and sunshine than the shaded A. hibernica, 

 with its broader, more horizontally placed pinnae and pinnules. The former 

 species seems to have provided channels for the passage downwards of the 

 water falling on its leaves. Can one reconcile the difference of statement as 

 to the presence of the rachidial pinnules, apart from the scrappiness of 

 Goppert's material ? If such pinnules are vestigial, they should be less 

 constantly present, and thus less often observable the more remote from its 

 ancestors, i.e. the younger, the species is. If Archceopferis Roemeriana is more 

 recent than A. hibernica, this would help to account for the statements made 

 that in some specimens of A. Roemeriana no aphlebioid pinnules are observable. 

 As both the Bear Island and the Irish specimens show, the aphlebise cannot 

 be excluded from the diagnosis of this species. Further, as Nathorst states, 

 A. hibernica var. tninor must stand as well as A. Roemeriana, since, as his 

 figures and description show, Crepin's variety possesses overlapping, unstalked 

 pinnules, with decurrent base. 



I ought to mention that even in some of our specimens of true Archmopteris 

 hihernica non-overlapping, smaller pinnules do occur on some of the pinnae. 

 True A. Roemeriana must possess no overlapping pinnules, and the pinnules 

 must be attached to the rachis by a distinct stalk, a non-decurrent base. Such 

 is the case in our Glanmire specimen. Hence the material in the Botanical 

 Division of the National Museum, Dublin, shows the presence in the Upper 

 Devonian beds of the South of Ireland of the following species : — 



1. ArchcBopteris hibernica, Forbes, sp., first recorded in 1852. 



(Plate VII., fig. 1, in part.) 



2. A. hibernica, var. minor, Crepiu. 



3. A. Roemeriana, Gopp. sp. (Plate VIII.) 



4. A. TschermaU, Stur. (Plate VII., figs. 1, 2, 3 ; text figs. 1 and 2.) 



