688 
Kershaw.—A Fossil Solenostelic Fern . 
of the margins of the leaf-gap in this section cannot be accurately given, 
since the fossil is broken along the part indicated by dotted lines in e. The 
thickened margin is possibly almost complete, and did not differ much from 
its form in the previous section, d. 
The other margin to which the leaf-trace must have been attached 
is probably more seriously damaged. The next section,/, of which Fig. 3, 
PI. LVIII, is a photograph, shows the stele still more stretched, and the part 
b which is to form the branch is larger, but the actual leaf-trace is not 
preserved. 
Section g. Text-fig. and Fig. 4, PI. LVIII, shows the lateral branch now 
free from the main stele, though it is still joined up to the main stem 
by cortical tissues. The branch evidently became detached from the main 
stele by a gradual constriction of the xylem ring in quite a similar way to 
the departure of the branch in Microlepia. From an examination of the 
main stele it is obvious that at this level the leaf-trace had not yet departed, 
for the leaf-gap shows no signs of closing, and the free margin is still con¬ 
siderably thickened (Fig. 4, PI. LVIII). It is suggested from a comparison 
with Microlepia , which Sole?iostelopteris resembles in many respects, that the 
leaf-trace would probably be a hook-like mass of xylem attached at the end 
of the elongated portion of the stele on the side of the branch, at about the 
point where the xylem is broken (Text-fig., e and/). The branch was thus 
on the same side of the rhizome as the leaf-trace, and became free from the 
main stele shortly before it. The vascular anatomy of the lateral shoot 
(Fig. 4 and Text-fig.,^, //, i) is exactly like that of the main stem. 
Rootlets. 
Rootlet bundles are seen in many of the sections at different points 
on their way out through the cortex. In all cases they arise from the lower 
surface of the solenostele, i. e. the side remote from the leaf-gap. They are 
of the usual fern type—a diarch plate of xylem forming an almost circular 
mass with phloem on either side, a pericycle and clearly marked endodermis 
around (Fig. 8, PI. LVIII). 
While passing through the ground tissue of the stem the rootlets 
possess no ground tissue of their own, a feature noted by Gwynne-Vaughan 
in recent solenostelic ferns. The rootlet represented in Fig. 8 is the one 
seen at r in Fig. 3, in the sclerized tissue outside the solenostele. The 
groups of protoxylem (p. x.) show clearly at either side, also the development 
of metaxylem between. Unfortunately, in this particular rootlet, all tissues 
between the xylem and endodermis, which is quite similar to that surround¬ 
ing the stele of the stem, are lost. The diarch plate of xylem is placed 
with its long axis tangential to the circumference of the solenostele, as 
is usual in recent ferns. The mode of origin of the rootlet is easily seen in 
