THE ANATOMY OF TODEA. 255 
occasionally forms an unbroken cylinder. The medullary parenchyma is widest in 
Osmunda Claytoniana and most reduced in Todea superba, The position of the 
protoxylem also varies: in Osmunda the spiral elements are usually on the inner edge 
of the metaxylem, but in Todea, especially in the filmy species, the xylem is mesarch, 
while in T. hymenophylloides some of the strands are almost exarch. Another feature 
of importance is the presence of an internal eadodermis in O. cinnamomea and 
T. hymenophylloides; and in the former species internal phloem has been found in 
several specimens examined by Jeffrey and Faull. 
Irregular and ineonstant groups of sclerous cells occur in the centre of the medullary 
region in the stems of Osmunda regalis, O. cinnamomea, T. barbara, and T. hymeno- 
phylloides. The iuternal endodermis, like the sclerous cells, is not always present in 
the same stem, and both kinds of tissues are characterized by a sporadie manner of 
occurrence. If we were familiar only with Osmunda regalis and Todea barbara we 
should naturally describe the Osmundaceous stem as possessing a medullated monostele, 
but in the internal endodermis of Osmunda cinnamomea, Jetfrey and Faull recognize 
evidence which leads them to abandon this simple interpretation. These authors 
describe the Osmundaceous stele as a reduced form of an amphiphloie siphonostele *. 
This brings before us a wide question, the discussion of whieh depends primarily upon 
the value to be attached to the endodermis as a layer of supreme morphologieal 
importance. The well-known cases of Jquisetum t and Ranunculus} illustrate the 
difficulty of adopting the endodermis as the determining factor in morphological 
decisions; it was this difficulty that led Strasburger to depart from the definition of 
Van Tieghem and others, who recognize in the endodermis the innermost layer of the 
cortex, and, as such, a layer which may be accepted as a safe guide in defining 
the boundaries of stelar and extrastelar regions.  Strasburger substituted the term 
phloeoterma, to be used in a strictly morphological sense, retaining the term endodermis 
as a designation of histological but not necessarily of any morphological significance. 
An Equisetum stem with a special endodermis round each vascular bundle may be 
regarded as monostelic, the layers of cells with suberized walls being merely portions 
of stelar conjunctive tissue modified for a special physiological purpose. 
The view taken of the stelar structure of the Osmundacez depends to a large extent 
on the importance to be attached to the internal endodermis of Osmunda cinnamomea 
and Todea hymenophylloides. Faull states that an internal endodermis is present in the 
medullary rezion of both young and old stems; he notes the close structural agreement 
between the inner and outer endodermis of Osmunda cinnamomea, and draws attention 
to the fact that both endodermal layers become continuous through the ramular, and 
occasionally through the foliar, gaps. The occurrence of sclerous cells in the centre 
of the stem is adduced as another argument against including the axial tissues within 
the stelar area, and in support of this view Faull adds that sclerenchymatous elements 
are not found in the stele of other ferns. 
We are disposed to take the view that the medullary tissue is not homologous with 
* Jeffrey (1899) ; Faull (19o1). + Pfitzer (1867) ; Strasburger (1891), p. 434. t Marié (1885). 
202 
