472 R. KIDSTON AND D. T. GWYNNE-VAUGHAN ON 



at last it consists entirely of sclerenchyma (figs. 13 and 14). The detailed structure 

 of the sclerenchyma is shown in the more highly magnified petiole (fig. 15). Many 

 of the smaller rachides are more or less winged (fig. 16). 



Some of our specimens consisted almost entirely of isolated pinnules lying parallel 

 to one another so as to form a laminated peaty mass, and these yielded good trans- 

 verse sections of the midribs, veins, and mesophyll. In the midribs and larger veins 

 the vascular bundles are surrounded by a stout zone of sclerenchyma that extends 

 out to the epidermis on both surfaces of the leaf but thins down at the sides (figs. 17 

 and 18). The smaller veins are surrounded by one or two layers of sclerenchyma 

 only. In some cases the mesophyll of the lamina is extraordinarily well preserved. 

 On the under side of the leaf there is a considerable depth of spongy tissue with 

 rather large intercellular spaces, while on the upper side there are one or two layers 

 of well-marked but rather short and square palisade cells (figs. 19 and 20). The 

 chloroplasts which crowd the palisade cells are still perfectly distinct in the fossil 

 (fig. 21). The upper epidermis is well shown in fig. 20 ; the lower is less distinct, 

 but it can be seen in fig. 19, in which two stomata cut in vertical section can be made 

 out at the points a, a. The deeper tissues of the mesophyll are shown in the obliquely 

 tangential section (fig. 22). At the point a the palisade tissue is cut through, and 

 again the chloroplasts are perfectly distinct. At b the section passes through the 

 spongy tissue with its irregular-shaped cells and large intercellular spaces. 



In one of our sections we were fortunate to discover a portion which was cut in 

 such a manner that the cells and stomata of the lower epidermis are seen in surface 

 view (figs. 23 and 24). The vertical walls of the epidermal cells are markedly sinuous 

 and the stomata do not appear to differ in any essential point from those of the 

 modern Osmundas. 



The Branching of the Eachis. 



Too few cases of branching in the stouter parts of the rachis were found to admit 

 of a detailed description of the process. As shown by figs. 25 and 26, the xylem 

 strand of the rachis curves outwards on the side of the approaching branch, and this 

 projection is subsequently nipped off, probably as a closed ring, in a manner similar to 

 that which occurs in Osmunda regalis. It is clear from the inspection of fig. 27 that 

 no gap is formed in the leaf-trace as is the case in Todea swperba and T. hymeno- 

 phylloides. Cases of branching were found rather more frequently in the midribs 

 of the pinnules, and here again the same process is followed as in the pinnules of 

 several living Osmundacese* In some cases the endarch protoxylem on the side of 

 the approaching branch elongates and divides. The extremity of the leaf-trace along 

 with the outer protoxylem is then simply nipped off as the branch trace (figs. 28, 29, 

 and 30, l>r.). In other cases the extremity of the xylem strand first of all becomes 



* The vascular phenomena of the branching of the petiole in the Osmundacese is in course of investigation by 

 one of the authors. 



