224 MR R. KIDSTON AND MR D. T. GWYNNE-VATJGHAN ON 



lignified, as is actually the case in the corresponding cells of Osmunda cinnamomea. 

 The phloem ring is a broad one, and consists for the most part of large and conspicuous 

 sieve-tubes (figs. 9 and 10), amongst which a few elements are interspersed which may- 

 be regarded as phloem parenchyma (fig. 10, ph. par.). These are much smaller than 

 the sieve-tubes in size, and have somewhat firmer walls. The sieve-tubes were very 

 elongated and tubular (fig. 4, ph.), with stout but somewhat soft walls, for in most places 

 they are much distorted and often pressed flat, so that the lumen is almost obliterated. 

 Where they are well preserved, a thin brown outline is visible inside the sieve-tube, 

 suggesting a pellicle of some substance that has separated off from the wall and become 

 more or less contracted and folded (fig. 10, pel.). It may have been the inner layer of 

 the wall, or possibly represents the contents. No suggestion of a " porose layer" or a 

 protophloem could be detected in any part of our sections, transverse or longitudinal. 

 On the contrary, towards the outside, the phloem is directly continuous with a zone 

 of four or five layers of parenchymatous cells, conspicuous in the fossil owing to the 

 dense dark brown colour of their contents (figs. 4 and 9, per.). These contents are not 

 homogeneous, but exhibit a curious vesicular structure varying in texture from a 

 coarse foam to one so fine that it might almost be described as granular. The cells of 

 this layer are angular in outline, and fit into one another without leaving intercellular 

 spaces. Those of the innermost layers are comparatively small and vertically elongated ; 

 towards without they increase in size, becoming at the same time shorter, and thus 

 they pass over gradually into the parenchyma of the inner cortex. This innermost 

 zone must be regarded as a pericycle, although it is admittedly not marked off from the 

 cortex by any definite single layer comparable to an endodermis. At the same time, 

 the stele as a whole has a fairly sharp external contour, for the cells of the inner cortex 

 distinguish themselves from those of the pericycle by the general absence of the dense 

 brown contents and by the presence of well-marked intercellular spaces at the corners 

 (fig. 3). The broad inner cortex consists of thin- walled parenchymatous ovoid or rounded 

 cells, with intercellular spaces at the angles. In most of the cells the contents have been 

 completely disintegrated, but a considerable number are filled with a brown vesicular 

 substance similar to that described in the cells of the pericycle. Judging from the 

 contents of the cells in a corresponding position in the recent Osmundacese, and also in 

 some of the fossil forms (Osmundites skidegatensis, 0. Doivkeri), the vesicular masses 

 probably represent aggregates of starch grains in various grades of decomposition. 

 This is the more probable because in those parts of the cortex where the cells are best pre- 

 served the vesicular mass resolves itself into a number of separate contiguous spherical 

 outlines. 



At the periphery of the widest part of the cortex the cells gradually become 

 narrower and somewhat more vertically elongated. At the same time their walls 

 become much thicker and dark brown in colour (fig. 2, Sc. C). This region doubtless 

 represents the inner margin of a sclerotic outer cortex, the greater part of which is 

 wanting in our specimen. 



