182 
Lady Isabel Browne. 
in the phylogeny, become very slow, and there is some evidence that 
should the plant’s descendants revert to a more rapid growth in 
length the protoxylem may reappear in a different position. There 
is no positive evidence that this has occurred in the creeping stem 
of S, spinosa and the analogy suggested is with a very remote group 
of plants. The creeping axis of S. spinosa might be compared to 
such of the Stigmarian axes as possess only centrifugal xylem ; but 
the cases are not strictly comparable, for in the different species of 
Stigmaria we can trace the disappearance of the centripetal xylem 
and its replacement by pith, whereas in the creeping axis of Selagi- 
nella spinosa the centrifugal xylem appears to have assumed the 
position usually occupied by the centripetal, or, in other words the 
protoxylem is absolutely central. Another objection to the view 
that the exarch condition is primitive lies in the fact that according 
to latest observations the hypocotyl of Selaginella always contains 
an endarch protostele (8). This is a serious objection to considering 
exarchy as primitive, for, in a general way, subject to many 
exceptions, the phylogeny of the stele is normally repeated in its 
ontogeny. But on comparative grounds there are very strong 
reasons for regarding exarchy as primitive. Not only was such a 
condition clearly the primitive one of the older dendroid Lycopods, 
but it is surely a very significant fact that Miadesmia, the only 
herbaceous Palaeozoic Lycopod whose anatomy is known, possessed 
an exarch protostele (1). 
Besides the exarch and endarch monostele there are various 
other types of stele. S. Galeottei, for example is bistelic (11); such 
bistely may well have arisen by the division of a single stele such 
as that of 5. spinosa. S. Braunii may show an intermediate stage, 
for its creeping axis is at first monostelic, but becomes bistelic, while 
the erect stem remains monostelic, the two steles of the creeping 
axis uniting to form its stele. The majority of species of the genus 
are, however, monostelic, and their steles dorsiventral; this dorsi- 
ventrality of the stele is nearly always associated with heterophylly.. 
The single stele of these forms is ribbon-shaped, and has one dorsal 
and two marginal strands of protoxylem ; it is easy to see how this 
stele might be evolved from the exarch protostele pari passu with 
the assumption of a dorsiventral habit. In some tristelic species 
the dorsal protoxylem of the ribbon-shaped stele becomes a dorsal 
cord in the young plant and a dorsal stele in the mature plant, while 
the two marginal protoxylems also become separate. In other 
species, such as 5. incequalifolia, there are, besides these three 
