36 T. G. HALLB, LOWEK DEVONIAN PLANTS FROM RÖRAGEN IN NORWAY. 



well be expected to be gained from the fossils, provided the record goes sufficiently far 

 back. It is a general opinion that the Devonian floras are already too far advanced to 

 throw any light on this question. In the well developed floras of Kiltorkan, Bear Island, 

 etc., to which the attention has usually been confined, both megaphyllous forms, such 

 as Archaeopteris, and microphyllous forms, such as Cyclostigma, occur as dominant 

 elements. These floras, however, belong to the Upper Devonian. The Lower Devonian 

 floras, from reasons mentioned in the introduction, have mostly been neglected, although 

 it would appear that a critical review of the available material would lead to the recog- 

 nition of some noteworthy facts. 



In the following pages the evidence of the occurrence of respectively the Lycopsida, 

 the Pteropsida and the Bryophyta will be shortly discussed. 



Occurrence of microphyllous Pteridophyta. 



The most reliable evidence of the occurrence of microphyllous Pteridophytes in 

 the Röragen flora is offered by Arthrostigma gracile. It can hardly be doubted, on grounds 

 set forth in the descriptive part (p. 13), that this plant consists of leaf-bearing stems. 

 The forms with lateral appendages placed in spiral and provided with single veins would 

 seem to leave no room for doubt, especially if compared with the branching stems. The 

 central strand of tracheids proves that these fossils are remains of vascular plants, and 

 it would be absurd to interpret them as fern-rachises. 



In Psilophyton princeps the case is not so clear, if the species is regarded by itself. 

 The spines are not very leaf-like, though they sometimes show indications of a regular 

 spiral arrangement, and they cannot be shown to receive any vascular supply. On the 

 other hand, the mode of branching is suggestive of stems rather than of fern-rachises, 

 and, if the latter was their nature, where are the pinnules? The chief reason, however, 

 for regarding Psilophyton princeps as being of stem-nature, and the spines as rudimentary 

 leaves comparable to those of Arthrostigma, is the great resemblance which it shows to 

 the latter species. It is indeed often difficult to decide whether a specimen should be 

 referred to one genus or the other. One might compare, for instance, figs. 1 and 6 or figs. 

 3 and 10 in pl. 4. 



There seems to be a fairly continuous series of forms leading from the problematic 

 Psilophyton princeps through Arthrostigma gracile to Cyclostigma kiltorkense, species of 

 Bothrodendron, and other undoubtedly lycopodiaceous forms of the Upper Devonian. 



On the other hand, both Arthrostigma gracile and Psilophyton princeps present 

 some peculiarities which characterize them in comparison with the typical Palaezoic 

 Lycopodiales. The most important of these features, which their occurrence in the 

 oldest known forms of the series gives a strong presumption to regard as primitive, are 

 the following. 



1. The slenderness of the central vascular strand in Arthro- 

 stigma as compared with the bulky ste m. The vascular strand cannot 

 have been very effective as a mechanical tissue. There is no trace of any peripheral 



