Stigmarian Rootlets . 567 
than the lighter external parenchyma. It is a noticeable 
feature of these rootlets, too, that the thin-walled cells of the 
outer cortex are considerably larger where they adjoin the 
tracheidal patch than in any other part of the circumference, 
and present therefore a specialized appearance, and one 
suggestive of very active absorption at this point. This 
larger-celled portion of the outer cortex does not run con- 
tinuously along the rootlet, but, as can be seen from longi- 
tudinal sections, is confined to those regions where a vascular 
branch terminates in the cortex in wide spiral tracheids. 
We are in the case of fossil plants at a great disadvantage 
in endeavouring to explain the function of unknown organs, 
as we cannot adduce any experimental evidence, and must 
base our conclusions on inferences drawn from anatomical 
facts only. In the present case the specialized cells of the 
cortex just referred to may be taken to give an important 
clue to the function of these peculiar vascular strands of the 
Stigmarian rootlets. It would appear from Figs. 3 and 4 that 
the rootlets were more particularly active as absorbent organs 
at those points at which the large-celled parenchyma occurred 
beneath the epidermis. Not only are these cells marked out 
as likely to maintain active osmosis, but they were in the most 
advantageous position to pass on the water they had absorbed 
to the central cylinder across the parenchymatous bridge which 
joins the outer cortex at this same point. The passage of water 
might be considerably accelerated by the presence of the vas- 
cular strands described above. As in the leaves, so in the root- 
lets, there was no direct connexion between the turgid paren- 
chyma and the narrow spiral tracheids, but an intermediate 
tissue of wide spiral tracheids was interpolated. 
In the case of the leaves of Monocotyledons and Dicotyledons, 
these tracheids, through which the outward diffusion of water 
takes place, are generally spirally- marked and of a very 
angular appearance 2 — due, as De Bary states, to the end 
surfaces bordering on the parenchyma being usually ‘ cut off 
sharply either transversely or obliquely.’ The same description 
1 Cp. Sachs, J. (’ 87 ), Fig. 196. 
