COMPARATIVE DISCUSSION 487 
“these two great stocks appear to have been separate back to the 
beginning of the period when the palaeontological record begins.” If 
this were so, the anatomical difference would, in all cases, indicate a true 
phyletic distinction. It is necessary to obtain a clear idea whether or 
not this will hold good. 
The structural difference is based upon the greater or less dominance 
of the leaf in the whole shoot; the phyllosiphonic type going, as a rule, 
with a megaphyllous state. But megaphylly may have been attained along 
more than one line of descent. If it arose in more than one phyletic 
Fic. 268. 
Tmesipteris tannensis. Transverse section of the sterile region, high up. The proto- 
xylem (#7. xy.) is mesarch. The xylem of the stele is fading out, and being replaced 
by parenchyma ;: three of the tracheides (z. ¢v.) show incomplete.development ; there 
is no longer a complete ring, and the leaf-trace bundles (2. ¢.) enter the gaps which result, 
in much ‘the same way as in a phyllosiphonic type. There is no definite endodermis. 
X 150. 
line, then the phyllosiphonic state, which is its internal structural expres- 
sion, will also have originated more than once. If this were so, then the 
phyllosiphonic structure would not necessarily indicate affinity, and the 
Pteropsida, as based on the structural point, could not be held to be a 
natural group. The question will therefore be whether there is any evidence 
of the origin of a phyllosiphonic from a cladosiphonic state. It might be 
expected either in a shoot, with increasing proportion of the leaves, or of 
decreasing proportion of the axis. The latter is the state of the distal 
region of the shoot of Zmesipieris, and Fig. 268 shows the condition there 
seen: the two larger tracts of xylem are separate; but isolated elements 
showing imperfect lignification link them together : the cauline stele is here 
seen in course of disintegration into mere leaf-traces: these enter the 
