the Anatomy of the Genus Selaginella , Spr . 153 
folia ; the former shows a distinct tendency to that splitting 
of the single stele into dorsal, median, and ventral steles, which 
is carried to its extreme in the latter type. All the species 
which have the Martensii- type of leaf belong to the mono- 
stelic and tristelic series as defined in that paper. A few 
species, also monostelic so far as their stem-structure is 
concerned, vary somewhat from the normal type of leaf- 
structure. 5. molliceps , S. apus , and S'. albonitens present no 
difficulty, seeing that their leaves differ from the Martensii- 
type only in having the reticulate mesophyll reduced to 
a minimum. S. Bakeriana , S'. Braunii, and 5. concinna , on 
the other hand, differ from the Martensii- type in two respects ; 
viz. in having a more or less distinct palisade-layer, and in 
that the epidermis of the ligular and aligular surfaces is more 
or less similar. As I have, however, tried to point out above, 
the length of the epidermal cells is by no means constant 
even in the same species ; and several species belonging to 
the Martensii- type, e.g. S. Helvetica and S'. plumosa , form 
pseudo-palisade-layers, so that these three species are perhaps 
not so aberrant as they at first sight appear. In the series 
with quite equivalent epidermis on the ligular and aligular 
faces, we find all the species with two laterally placed steles, 
e.g. S. Kraussiana , as well as a few monostelic species. Of 
these latter, three (S'. pilifera, S. lepidophylla , and S'. involvens ) 
have a peculiar in-rolling habit, which may have something 
to do with the equivalence of the epidermis on both leaf- 
surfaces, whilst the remaining three (S'. rnpestris , S. spinosa , 
and 5. oregano) are homophyllous. S'. Lyallii is peculiar not 
only in regard to the anatomy of its stem, but also in its 
leaf-structure. Looking at the genus as a whole (at all events 
the fifty-two species here discussed), one may say that there 
are three chief types of stem-structure : the dorsiventral 
monostelic, leading to the tristelic, and the lateral bistelic. 
The leaf, of course, is much more likely to be liable to the 
influence of the environment than the internal anatomy of the 
stem, and therefore it is less to be depended upon for classi- 
ficatory purposes. Still, it is interesting to note that, with the 
