THE FOSSIL OSMUNDACE^E. 769 



formed numbers of small, oval, dark-coloured spores. The stele of Osmundites Dowkeri 

 in all essentials closely resembles that of 0. Gibbiana. There are about thirty separate 

 strands in the xylem ring, and many of the tracheides have more than one vertical series 

 of pits on their broader walls. The phloem is too badly preserved to show whether 

 there was a porose layer or not. The pith is still intact and fairly well preserved (PI. IV., 

 fig. 21, P.). It consisted of rather thick- walled parenchymatous cells with conspicuous 

 coarse, irregular reticulate markings on their walls. The elements at the periphery of 

 the pith are smaller and more sclerotic than the rest, and some of the central cells 

 appear to be thin-walled, but this may be an accident of preservation. A marked 

 feature of Osmundites Dowkeri is the strong curve taken up by the xylem of the leaf- 

 trace almost immediately it has left the stele (PI. IV., fig. 21). In 0. Gibbiana the 

 curve is only very slight until the leaf-trace has left the cortex of the stem. The 

 coating of leaf-bases is well preserved in Osmundites Dowkeri ; even the parenchymatous 

 tissues of the stipules are still present. Towards the periphery of the section the limits 

 between the separate stipules are clearly marked out by brown lines, although in 

 close proximity to the axis they are all concrescent. The diagram text fig. 5 represents 

 a restoration of a section of the leaf-base of Osmundites Dowkeri, showing the distri- 

 bution of the sclerenchyma, and it is altogether distinct from that of 0. Gibbiana 

 (PI. VI., fig. 4). The stipule as a whole is much thicker in proportion to its length, and, 

 in particular, the single series of oblong sclerotic strands in the wing of Osmundites 

 Gibbiana are here replaced by a number of more or less rounded strands scattered 

 irregularly and at different levels in the thickness of the stipule. In both species 

 isolated strands of sclerenchyma occur within the sclerotic ring, but the stout band 

 of sclerenchyma that lies in contact with the concave surface of the leaf- trace in 

 Osmundites Doivkeri was not to be found in O. Gibbiana. 



The Root. 



The roots arise one at a time, or possibly sometimes in pairs, at the angles of the 

 leaf-gap, before the leaf- trace is yet free from the stele of the stem. The xylem strand 

 is always diarch, and the root obtains a cortex of its own while passing through the 

 outer cortex of the stem. The outer zone of this cortex is at first heavily sclerotic, but 

 as the root passes outward it becomes all more or less thin-walled. The number of 

 roots that occur among the leaf-bases is very large, and they differ considerably in size 

 and structure. The smallest roots are exceedingly delicate, with a very small stele. 

 The endodermis consists of some six to nine rather flattened cells, and these are sur- 

 rounded by the same number of especially large cortical cells filled with some finely 

 granular matter. There is no reason, however, to believe that these smaller roots belong 

 to some other plant. They are very probably branches borne by the larger main roots ; 

 indeed, one case of branching was actually observed. Moreover, excellent examples of 

 branching roots within the coating of leaf-bases are provided by Osmunda Claytoniana 



