374 Sahni—On the Branching of the Zygopteridean Leaf . 
true conception of the embedded trace of Metaclepsydropsis , however, throws 
a weight of argument in favour of the conjecture, which was not available to 
Dr. Bertrand. 
III. Distribution and Affinities. 
The discovery in the Australian region (see below, p. 375) of a member 
of the Zygopterideae, a family until recently known only from Europe and 
West Siberia, is of considerable interest. It is one more piece of evidence 
in support of a fact which has repeatedly been demonstrated in the past, 
namely, the essential similarity of the ancient floras (and faunas) of regions 
of the earth now widely separated by the sea. To take the most recent 
case, we know that the world-wide distribution of the modern Osmundaceae 
has been known, within the last few decades, to have been almost paralleled 
by that of the fossil representatives of the family. The recent work of 
Dr. Kidston and the late Professor Gwynne-Vaughan has established a close 
relation between the fossil Osmundaceae and the Zygopterideae, and it 
would not be surprising if members of the latter family are discovered in 
other parts of the world in rocks underlying those containing Osmundaceous 
remains. 
From anatomical evidence a common ancestry has been advocated for 
the Coenopterideae and the Osmundaceae (8, p. 778 ; 14 , p. 472). It is 
a striking fact, however, that whereas the Coenopterideae are known to 
extend down to Upper Devonian times, the earliest record of the Osmun¬ 
daceae is only in the Upper Permian ( 9 , Table, p. 465). The Coenopterideae 
were indeed just becoming extinct when, as far as present knowledge 
goes, the Osmundaceae began to make their appearance. A time relation 
of this kind between two undoubtedly closely allied groups would seem to 
suggest, although somewhat vaguely, that the older group may have stood 
towards the younger in the relation of an ancestor. It is, however, too early 
to speculate on the latter view : we have not yet arrived at the limit of 
primitiveness which the Osmundaceae probably attained, and in the earlier 
rocks forms may yet be discovered which it would be difficult to assign to 
one or the other of these two families. 
Indeed, it may be that we already have one such form in Grammato- 
pteris Rigolloti , B. Ren. (12), although it has not yet been recorded from 
rocks older than the Permo-Carboniferous. As Kidston and Gwynne- 
Vaughan (8, p. 778) have remarked, this species may be a near approach to 
the primitive Osmundaceous type ; at the same time the leaf-trace bears 
undoubted resemblance to that of Dineuron. 
A few words are necessary in connexion with the affinities of some of 
the Zygopterideae. In classifying the different Zygopterid petioles Kidston 
and Gwynne-Vaughan ( 9 , p. 470) have expressed the opinion that the most 
