539 
Origin of Medullation in the Ophioglossaceae . 
depends upon the nourishment of the young plant. If it be well nourished 
the non-medullated phase is omitted, and pith is present from the first. 
But it is in ill-nourished or depauperate plants that the start is made with 
a protostele, which subsequently becomes medullated. 
It will be important to observe the exact method of origination of the 
pith as it first appears in the protostele, in order to form an opinion on its 
morphology. The first appearance of the pith in a plant with a solid pro¬ 
tostele at the base has been followed in O. reticulatum , and it is illustrated 
in the Figs. 2-5. At first the solid xylem-core is without parenchyma, 
being composed only of tracheides. But a little below the origin of the 
first leaf-trace, single cells, or small nests of them, retaining their thin walls 
and protoplasmic contents, appear among the mature tracheides (Fig. 2). 
Passing upwards, these merge into a central parenchymatous pith, which, 
however, at first has no clear limit from the xylem. Tracheides and 
parenchyma are intermixed (Figs. 3, 5). This continues only for a short 
distance, the pith consisting subsequently of pure parenchyma, which is 
then continued without interruption upwards. The surrounding ring of 
tracheides very shortly opens, throwing the pith into communication with 
the outer conjunctive tissue of the stele. This is the first step in the 
segregation of the leaf-trace (Fig. 4). 
There can be no doubt that the parenchyma nests which precede the 
first establishment of the pith, and merge with it, are not only intrastelar, 
but also intraxylic, in origin. A proportion of the pith may, however, owe 
its origin to an intrusion of the outer conjunctive tissue: but this source 
also is intrastelar. The similar origin of an intrastelar pith to be 
described below for Botrychinm Lunaria , where the endodermis is clearly 
defined, confirms this conclusion. 
The well-known imperfection of the endodermis in Ophioglossnm is 
certainly a difficulty in the interpretation of the development of the young 
plant in most species of the genus. But on the other hand, the small 
species, O. Bergianum , has been quoted by Poirault (loc. cit., p. 170, 
Fig. 12), as having not only an external, but also an internal endodermis, 
the latter in the pith. I have cut the stock of my only specimen into 
transverse sections, but with negative results. No endodermis of the axis 
was found to take the safranin stain, though that of the roots, wherever 
present, showed up well in the same sections, with the same stain, and on 
the same slides. The structure was, however, conformable otherwise to 
that shown by Poirault, and I have no reason to doubt his observations, 
which were probably made on more favourable material. But it appears 
that the development of the endodermis is variable in the species. The 
above plant was sent to me by Mr. H. Bolus from the Cape in 1894. With 
it was sent a distinct plant, which I recognize as one of the small Cape 
forms of O. lusitanicum. This was also cut into sections, and showed the 
