XXVI INTRODUCTION. 



some evidence (see pp. xxviii, xxix), would also account for the 

 presence of this genus in Natal. 



It is significant that, although Sphenophyllum has heen recorded 

 from the Lower Carboniferous rocks of New South Wales, no trace 

 of it has yet been recognised in association with the Glossopteris 

 flora in Australia. 



Filicales and ?Cycadofilices. Fern-like plants form a very 

 characteristic feature of the Glossopteris flora, but in no case is the 

 fructification sufficiently well known to decide whether the plant 

 in question is a true Fern or a seed-bearing, fern-like plant 

 (Pteridosperm). Also, in the absence of petrified material, the 

 existence of the Cycadofilices has yet to be demonstrated among 

 this flora. The only comparison possible with the Northern 

 Pernio- Carboniferous vegetation is one based upon the vegetative 

 habit, which is notorious as an unreliable guide to affinities. 

 Still, it must be admitted that the fern-like plants of these two 

 floras offer a somewhat sharp contrast, at least in habit, although 

 types common to both occur. We may first consider the chief 

 differences between them. 



In Glossopteris and Gangamopteris we have fronds pre-eminently 

 characteristic of the southern type of flora, and quite distinct iu 

 form and nervation from any known representative of the northern 

 province. Both are not only diversified genera, especially Glossop- 

 teris, but flourished in great abundance numerically. Glossopteris is 

 found almost everywhere within the boundaries of Goudwanaland, 

 the most noteworthy exceptions being the colony of Victoria in 

 Australia, and Brazil. It is interesting to find that where Glossop- 

 teris is absent, Gangamopteris, a plant probably very closely allied 

 to the former genus, occurs, as in Victoria and Brazil. A large 

 number of species of Glossopteris are known, at least thirteen, 

 which in the details of their habit varied considerably. Gangamop- 

 teris was probably a smaller genus; four species and one variety 

 being recognised here. It also appears to have been a somewhat 

 earlier type, as it is of common occurrence in the oldest Glossopteris- 

 bearing beds in India, in which Glossopteris itself is much less 

 abundant. In New South Wales, however, it is unknown in the 

 Muree Series. The genus apparently became extinct before Glosxop- 

 teris, for it is not known to persist into the Triasso-Rhsetic period, 

 whereas Glossopteris is found in beds of that age iu India, Tonquin, 



