Bower.—On Medullation in the Pteridophyia . 569 
which appears as a ‘ mixed pith ’ in O. Kolbei . Finally, in the more modern 
Osmundaceae the pith is fully formed, while the leaf-pockets are never seen 
to be so large as in other Filicales. It is thus seen that where, as in the 
Osmundaceae, the axis is upright and relatively bulky, and the leaf insertion 
relatively narrow, the intrastelar pith is dominant, while the leaf-pockets 
remain relatively small. But in the Ophioglossaceae the proportion of the 
upright shoot is habitually different. They are commonly monophyllous, 
producing only one leaf in each season, but that leaf is larger proportionately 
to the axis than are the numerous leaves of a season in the Osmundaceae. 
We should then expect that the intrastelar medullation in them would be 
proportionally smaller, and the foliar pockets be relatively more prominent. 
This result is clearly indicated by the observations on the seedlings of 
Botrychium above detailed. 
But if the balance of size of the parts really does influence the internal 
structure, as suggested, an interesting condition should be seen in the 
seedling of Botrychium Lunaria, where the first leaves are all minute scale- 
leaves. If the size of the leaf-pocket is in any way proportional to that of 
the leaf itself, the pockets of the earliest leaves of the seedling of this species 
should be small. That is exactly what is shown by the series of sections 
of the young plant of B. Limaria described above. There is indeed no 
intrusive pocket at all in relation to the first minute scale-leaves of this 
seedling, and the barrier of endodermis is unbroken, so that there is not any 
communication at all between the cortex and the pith. The latter is 
present, however, though of relatively small dimensions, and is thus intra¬ 
stelar in its origin. But in the larger leaves the endodermis curves convexly 
inwards, forming a pocket that encroaches on the pith. There is thus 
indication that in the Ophioglossaceae there is fluctuation of the balance 
between the two factors, and that it may even vary in the individual plant. 
The facts strongly suggest that the pith is here of two distinct origins, 
partly intrastelar and partly extrastelar, a condition which was contemplated 
as possible at the opening of this essay. But the original endodermal limit 
between the two constituent parts is often obliterated by the loss of its 
endodermal characters, partially or altogether. This is found in the upper 
regions of B. Lunaria , and it appears to be the constant condition in 
B. ternatum and virginianum , as it is also for most of the species of 
Ophioglossum. In this genus, however, the obliteration of the endodermis 
has gone still further, and involved commonly the external endodermis also. 
In the family as a whole the irregularity of occurrence of the endodermis is 
probably due to there being no physiological need for the maintenance of 
the barrier between the various parenchymatous masses, which are used for 
storage purposes. That being their prime function, the less the barriers to 
physiological transit the better. Thus the imperfection of the endodermis 
finds a physiological explanation. Nevertheless, the two components of the 
