5 
the Ophioglossaceae . 
paratively small-leaved genera or species (e. g. Pteris cretica ) does evidence 
of the gap disappear, and the lateral bundles spring from the edge of the 
petiolar trace. Bertrand and Cornaille ( 1 . c.) have argued that the condition 
found in Osmundaceae is to be regarded as primitive, and the conditions 
seen in Polypodiaceae seem to be derived by reduction. It may be added 
that the structure of the midrib of the pinnae in Marattia is of the same 
general structure as the petiole of Todea, namely, the vascular supply of the 
pinnules arises by distinct gaps. The split or divided U seen in the petiolar 
strand of Botrychium virginianum recurs in numerous Polypodiaceae, 
e. g. the stipe of Onoclea , Pteris , spp., Asplenium , spp., so this feature 
forms no drawback to the derivation of Botrychium from fern ancestors. In 
Onoclea the leaf-trace is double near the base of the leaf, but higher up the 
two halves join to form a U-shaped strand ; the same sometimes happens 
in B. virginianum. In both these cases the leaf-trace arises as a single 
strand, while the supply of the fertile spike of Botrychium is double from its 
origin upward. The diagrams shown in Fig. 1-13 were made from a fairly 
small specimen ; in well-grown specimens the plan is less apparent on 
account of the parting asunder of some of the ‘divergents’ which compose 
the leaf-trace, though the plan is in reality precisely the same as in smaller 
plants. 
Of the species examined, B. virginianum seems to be the most fern¬ 
like. This appears not only in the mode of origin of the vascular supply 
of the pinnae, but in the amphiphloic feature of the leaf-trace and the 
distinctness of the divergents. Several varieties of B. tcrnatum have also 
been available for study, namely, intermedium , dissectum , and obliquum , 
which has lately been recognized as a distinct species. 1 These varieties 
merge into one another in a way that makes distinctions difficult. The 
anatomy of these varieties offers no basis for distinguishing them, and they 
may here be considered together. 
The leaf-trace of B. ternatum , like that of all other species, leaves 
a wide gap where it parts from the central cylinder. At first it is only 
slightly curved (Fig. 14), but it rapidly assumes the form of the letter C, 
and may almost form an O. From the free edges arise the two strands 
(Fig. 15) which supply the fertile spike. These strands may fuse more 
or less just above their point of origin (Fig. 16), but separate higher up. 
All of the strands so far mentioned are collateral, in distinction to those of 
B. virginianum which are concentric. The branches of the fertile spike 
show interesting transitions from the condition seen in the sterile part 
to that described for B. virginianum. In some specimens these branches 
manifestly arise from gaps, as is illustrated in Fig. 20. A small strand 
(hereafter called the marginal strand) arises from each edge of the main 
1 Lyon (Botan. Gaz., xl. 455-8, 1905) even proposes to separate it into a new genus, 
Scep/ridium, on the basis of gametophyte characters. 
