the Anatomy of the Genus Selaginella . 455 
point of view, therefore, of general external morphology there 
does not seem any stronger reason for distinguishing the 
aerial part of the root by a special name and ascribing to it 
stem-characteristics than there exists for differentiating the 
aerial part of the root of a Pandgnus or a Rhizophora from 
that which is subterranean or subaquatic. 
As a rule the root is intimately related to the branching of 
the stem, and normally one root arises from the stem at each 
fork. There is frequently an abortive root on the upper side 
of the stem, which scarcely appears above the surface, or is 
represented by a minute papilla, as in 5. Martensii , 5. Will- 
denowii , &c. Occasionally both roots are developed, e. g. 
vS\ caulescens. In 5. inaequalifolia, as Pfeffer points out, the 
upper, usually abortive, root develops into a leafy shoot, and 
on that ground an argument is based for the cauline value of 
the aerial portion of the root. On the other hand, roots, as is 
well known, not infrequently bear both leaves and leaf-buds ; 
and further, as Jones has pointed out in the allied genus 
Lycopodium (11), the differentiation between root and stem is 
far less marked in the vascular cryptogams than in higher 
plants, and consequently the organ in question may, without 
undue violence to morphological conceptions, be conceived as 
having a preponderance of cauline or radical characteristics 
according to circumstances. Pfeffer also draws attention to 
the neutral nature of the aerial part of the root. Another ex- 
planation is, however, possible, suggested by the condition 
of the roots in 5. Lobbii. Normally, in this species, there is 
one aerial root developed at each forking of the stem, and 
that the lower one. The upper root is represented as usual 
by a papilla. Frequently, at all events in the material 
I possess, the upper root-papilla is replaced by a leafy branch ; 
but at a short distance from its point of origin there is given 
off an aerial root which behaves exactly in the same manner 
as the lower aerial root. The question thus arises whether 
the leafy shoot of .S. inaequalifolia and other species which 
possess this abnormality is not a genuine shoot, proliferating 
at the normal bifurcation, which is not robust enough to bear 
