766 MR R. KIDSTON AND MR D. T. GWYNNE-VAUGHAN ON 



concrescent in close proximity to the stem, as in the modern OsmundacesB. As in Osmun- 

 dites Dunlopi, the xylem strands of the small leaf-bases supposed to belong to scale-leaves 

 were very badly preserved, and it seems probable that those elements only of the meta- 

 xylem which were in immediate contact with the protoxylem groups were ever fully 

 differentiated (PI. IV., fig. 20, Sc.L.). 



The Roots. 



The root steles in this species are very small and delicate, with a diarch xylem 

 strand of very few elements. They arise from the margins of the leaf-traces as they 

 pass through the inner cortex of the stem ; apparently not in pairs, but one from each 

 leaf-trace. The root does not obtain a cortex of its own in its passage through the stem, 

 and those that occur in the coating of leaf-bases are so badly preserved that nothing 

 can be made out in them except that they seem to have had no cortical sclerenchyma 

 at all. 



Affinities. 



The two fossil stems just described agree so exactly in all essential characters with 

 the stems of the modern Osmundaceie that we have no hesitation in including them in 

 this order. Again, they differ from each other so much in detail, they undoubtedly 

 represent two distinct species. On the other hand, it is a very difficult matter to decide 

 whether they belong to the same genus or not, and, in any case, whether they may be 

 identified either with the modern genus Osmunda or with Todea. In the first place, a 

 careful examination of the stocks of the living representatives of the order only pro- 

 vided a single morphological criterion whereby any given Osmundaceous stock may be 

 definitely assigned to one or the other of the two genera. This distinction is the 

 presence of a transverse commissure formed by the fusion of the upper part of the 

 stipular wings across the adaxial side of the petiole in the Todeas (T. barbara, T. 

 superba, T. hymenophylloides, and T. Frazeri), which appears to be entirely absent m 

 the Osmundas (0. regalis, O. Javanica, 0. cinnamomea, 0. Claytoniana). Unfortun- 

 ately, this does not help us at all with the fossils, because the preservation is not good 

 enough to determine whether a transverse commissure was present or not in either 

 of them. 



The only thing to be done, therefore, is to obtain some convenient and reliable 

 anatomical characters which will serve to distinguish between the several living species 

 of the whole order, considered separately and without respect to the genera to which 

 they belong, and then to apply these to the fossils. Such distinctive characters proved, 

 however, to be very difficult to discover ; in fact, we have been reduced to the selection 

 of the sclerenchyma in the base of the petiole. The arrangement of the sclerotic tissues 

 in most Ferns is admittedly a variable and insecure feature for such a purpose, but it 

 was found that in the Osmundaceie the various sclerotic strands that occur in the base 



