648 GENERAL COMPARISON OF THE FILICALES 
former, though its soral structure has remained virtually unaltered, shows 
the highest condition known of the solenostelic development. 
Parallel with such structural progression of the axial system goes an 
elaboration of the leaf-trace, In protostelic, and usually in solenostelic 
forms, it is represented by a single strand, which may, however, be widened 
into a broad strap, and curved in transverse section into a horse-shoe outline: 
and this may again be contracted into a pseudo-stelic condition (compare 
Fig. 98, p. 194). But with dictyostely comes usually a division of the 
single strand into many. It is interesting in Ducksonia to see a middle 
condition illustrated; for in D. Culcita and D. Barometz the leaf-trace at 
its base is still a single strand, but at a point above the base, varying in 
different leaves, it breaks up into many separate strands: it thus represents 
the various -stages of the probable phyletic sequence (Fig. 97). The 
complete subdivision is seen in the larger species of Dzcksonza and in 
Cyathea, as well as in most of the Mixtae, and it is held to be an 
advanced and derivative state. 
The seedling structure gives a strong support to the view of elaboration 
here put forward: in all observed cases the stele of the axis is either 
protostelic or very closely related to that structure, and the leaf-trace is a 
single vascular strand. In the primitive forms this may remain permanently 
so; but in others there may be a quick transition to the more complex 
and presumably derivative state. The example of Alsophila excelsa 
(p. 608) shows that the individual life, after the first stages are past, 
reflects the probable story of development of the complex adult condition 
in the race. 
It is in this way, through the seedling, that the Marattiaceae may best 
be approached. They have in the mature stem a still more complicated 
system of vascular strands than other Ferns; but in their seedlings the 
ontogeny opens in all cases with a monostelic state, with a solid xylem-core. 
Complications soon arise: in Kaulfussia and Archangiopteris a cylindrical 
dictyostele is formed, not unlike that of other Ferns; but in Angiopteris 
there may be as many as three or four concentric, meshed zones in the 
stock, while the leaf-trace is also disintegrated into numerous strands. It 
is important to note, however, that in the related fossils the leaf-trace is 
habitually a single connected strand, while greater coherence is also seen 
in the vascular tracts of the axis in the fossils than in the living species. 
These facts all indicate that in the evolution of Ferns there has been 
a progressive amplification and disintegration of the vascular tissues; and 
they lead back towards a type, which seems to have been a common one: 
the original type was characterised by a radial shoot traversed by a protostele, 
from which the successive leaf-traces came off each as a simple strand, and 
with the minimum of disturbance of the axial stele. If this were the 
original type of shoot in the Filicales, it is plain that the foliar gap, to 
which Jeffrey attaches so much importance as the distinctive character of 
his phyllosiphonic type, must have been a secondary development: it is 
