the Anatomy of the Genus Selaginella . 503 
walls with root-hairs followed by another thin-walled layer and then by 
a three-layered hypodermis composed of somewhat thick walls, before one 
observes the cortex proper, this being the regular succession of tissues 
found in practically all roots with which the writer has worked. Near the 
boundary where the rhizophore partly touches the wet soil and partly the 
free atmosphere, one is able to find any modification between the aerial and 
terrestrial root, due to the direct action of environment. Both types can be 
observed in Fig. V, 5-7. 
The writer suggested above that the thin-walled layers of a terrestrial 
root which underlie the epidermis have been formed from the latter, whereas 
the somewhat thick-walled layers belong to the hypodermis. Such can 
clearly be observed on N. viticulosa , the aerial roots of which have a red- 
coloured epidermis and hypodermis, the colour pigment being encrusted in 
the secondary walls. The same colour can be observed in the thick-walled 
layers of the terrestrial root, although its colour is not present in the thin 
walls of the epidermis cells, on account of the absence of secondary layers. 
Moreover, a longitudinal section of an aerial with a terrestrial root shows the 
connexion of the hypodermis in both. The author is unable to understand 
why these cells of hypodermis are thick-walled in a terrestrial root. 
In twenty-two instances roots of N. rubella were grown in the soil 
and, still being connected with the plant, were carefully taken out of the 
soil, washed off, and kept so that they were unable to touch the soil ; some 
of these rapidly died ; in two instances the roots continued their growth in 
the air, their root-caps gradually consisting of less cells and finally dis- 
appearing ; at last this part of the root became in construction again 
entirely an aerial root. 
The above anatomical and physiological investigations are not in favour 
of the stem habit of the rhizophores, which are on the contrary entirely 
root-like in construction and behaviour. In one of the above-mentioned 
experiments it could not be stated that the half of the rhizophore which 
was open to the air was a stem, while the half which was surrounded by 
soil was a root. 
All roots possess a monarch vascular bundle, the endodermis of which 
is in some species difficult to distinguish ; it is very clear in N. Wildenowi 
and N. rubella ) but does not show very well in N. serpens , N. Bakeriana , and 
some other species. No root, whether grown in the air or in the soil, ever 
develops any lacunae or trabeculae. The pericycle is composed of one 
layer. The elements of the phloem, although not as abundant as in the 
vascular bundle of a stem, are arranged in the same manner as in the stem, 
and can to a certain extent be compared with the studies of Gibson on the 
anatomy of the stem. The sieve-tubes are consequently present, although 
apparently not much developed near the protoxylem. The phloem sur- 
rounds the xylem entirely. This xylem possesses but one group of proto- 
