190 ANATOMICAL EVIDENCE 
It is indicated by the presence in the transverse section of two or more 
vascular masses, each being constructed and delimited in the same way 
as the single stele of normal axes. At first it was thought by Van Tieghem 
that to produce this condition the stele, originally simple, underwent a 
branching, notwithstanding that the axis in which this would take place 
remained simple. This suggestion seemed inherently improbable, and _ it 
has since been shown by direct examination of specific cases that the 
real origin of the polystelic state as it occurs in Ferns is by the formation 
of large leaf-gaps below the bases of insertion of the successive leaves : 
the steps of increasing complexity of stelar structure in Ferns have been 
tentatively outlined by Gwynne-Vaughan as follows: The most primitive 
type of vascular construction was probably the single protostele, with 
uninterrupted central xylem, and this is met with in some of the early 
Fern types in the mature stem, but it is also seen in polystelic types of 
Ferns at the very base of the young plant. Internal parenchyma then 
makes its appearance about the peri- 
phery of the protostele at points just 
\ \\ above the departure of the leaf-traces: 
Oe this advances gradually inwards from 
™~SOSsithese «points until the most central 
if \ & Se \ region of the stele is affected. A struc- 
\¢) | ture resembling a cylinder or, as it is 
a : erste styled, a ‘“solenostele,” may then be 
Fic. 95. attained by the gradual differentiation 
Dicksonia punctiloba, Diagram of vascular Of Phloem and endodermis through the 
pytem, of rhsome, inleging ¢ nous andthe Jeafgaps and all round inside of the 
aR Ns the observer. (After xylem-ring. Below the insertion of 
each leaf a large leaf-gap occurs in the 
solenostele (Fig. 95): in transverse section at such a point the stele will 
appear as an incomplete ring. If, then, the arrangement of the leaves be 
a close one, two or more of these gaps would occur in a single transverse 
section, and the result would be an appearance as of several steles arranged 
in a ring. These originate, however, not by branching, as Van Tieghem 
thought, but by resolution of the stele, first into a cylinder and then into 
a cylindrical network. The result of this mode of amplification would, 
therefore, be more correctly styled ‘a dictyostele” rather than a polystelic 
state, and the parts “meristeles” rather than a plurality of distinct steles, 
since the whole is a result of amplification, not of branching, of the 
original monostele. 
But polystely is also found in the genus Se/aginella. There is little 
doubt that the monostelic type is the original one for this genus also, 
since it exists in many species. The origin of the more complex state 
is, however, connected rather with the branching of the axis than with 
the insertion of the minute leaves. It has certainly taken place within 
the genus, but the comparative study of the illustrative species from the 
