part 3] fossil plants from the Falkland islands. 315 



may safely be assumed that the depressions are leaf-scars, and not 

 the scars of roots. The broad ridge seen in several of the depres- 

 sions is probably the cast of a vascular strand supplying a leaf. 

 If the depressions were root-scars, one would expect to find a 

 central vascular scar as in Sfigmaria. 



Dr. Halle figures some stem-fragments from Halfway Cove 

 which he calls ' Lepidodendroid fragments ' : of these the smaller 

 piece represented in his pi. vi, fig. 3, 1 agrees closely with our 

 specimens. Halle's larger specimen differs in the more oval and 

 elongate form of the scars, and in the presence of a small pit on 

 the cast of each leaf-scar. It is, however, probable that despite 

 these slight differences, Halle's specimens and those collected by 

 Dr. Baker ai-e portions of closely allied, or possibly identical, plants. 



Lepidodendroid stems superficially similar to the Falkland speci- 

 mens have been described from several parts of Gondwanaland, both 

 from Permo-Carboniferous and from older rocks. It is noteworthy 

 that such fossils as Bothrodendron Leslii Seward from the Lower 

 Karroo beds at Vereeniging, 2 a specimen from LoAver Gondwana 

 rocks in Brazil referred by White to Lycopodiopsis Derbyi 

 Renault, 3 and Cyclostigma sp. figured by Feistmantel 4 from a 

 supposed Devonian locality in New South Wales, differ from the 

 Falkland specimens in the presence in the depressions of a central 

 pit in place of an obliquely placed cast of a vascular strand. Some 

 of the specimens of Bothrodendron irregulare Schwarz 5 from the 

 Witteherg Beds of South Africa bear, on the whole, the closest 

 resemblance to those reproduced in figs. 1 & 2 (PI. XIX). The 

 possibility of affinity of the Lepidodendroid fragments to some 

 Devonian genera other than Bothrodendron or Cyclostigma should 

 not be overlooked, although it is clearly impossible on the available 

 evidence to make any definite statement. Two pieces of stem from 

 Lower Devonian rocks in Norway assigned by Halle to Arthro- 

 stigma gracile Dawson 6 exhibit features similar to those of the 

 Falkland fossils, but we know nothing of the appendages which 

 were attached to the scars of the latter. On the whole, the Lepido- 

 dendroid stem-fragments suggest comparison with the Witteberg 

 species Bothrodendron irregulare Schwarz, which would be more 

 appnmriately included in the genus Cyclostigma, on the ground 

 that the specimens show no trace of the ligular pit characteristic 

 of Bothrodendron. In the absence of specimens showing well- 

 preserved surface-features, the precise affinities of the Witteberg 

 stems cannot be settled: the fossils described by Schwarz and other 

 authors from the South African beds may not be generically 

 identical with the European stems included in Cyclostigma or 

 Bothrodendron. 



The geological age of the Witteberg Series can hardly be 

 determined with precision on the meagre pakeobotanical evidence 

 available. In this connexion reference may be made to some 



1 Halle (11). 4 Feistmantel (90) pi. ii, fig. 7. 



2 Seward (03) pi. xi. 5 Seward (09) pi. xxviii. 



3 White (08) pi. v, fig. 11. 6 Halle (16) pi. i, fig. 8. 



