COMPARATIVE DISCUSSION 489 
more active development in Bofrychium finds its correlative in the more 
active thickening seen in Sphenophyllum. These again may be mere 
analogies, but they are cumulative, in that they run parallel with others. 
Lastly, it has already been shown that as regards structure of the 
roots there are unmistakable points for comparison of the Ophioglossales: 
with the Lycopodiales on the one hand, and on the other with the Fili- 
cales, especially the Marattiaceae: the former comparison is in respect 
of the simpler, monarch types, the latter as regards the more complex: 
the latter branch monopodially as a rule, behaving thus like the roots 
of the Filicales: the monarch roots, however, show dithotomous branching. 
Unfortunately, the Psilotaceae, which show so many other points for com- 
parison, are rootless, while the roots of Sphenophyllum are so imperfectly 
known as to give little help. Though the facts relating to the roots are 
not in any way decisive, they indicate, what emerges from so many other 
comparisons, that Opfzoglossum shows characters approaching the strobi- 
loid Pteridophytes, while He/minthostachys compares rather with the Filicales, 
and Botrychium takes a middle position. 
In the embryology two distinct types have been recorded for the 
Ophioglossaceae, the one with and the other without a suspensor. That 
without a suspensor corresponds in its essentials to the type prevalent 
in those Pteridophytes which have the usual octant division. But there 
are modifications here in accordance with the underground origin from 
a large mycorhizic prothallus, which nourishes itself saprophytically : the 
chief of these is the deferring of the period of functional activity of the 
shoot: consequently it is differentiated late, and though the root is not 
initiated early, as compared with other embryos, it very markedly precedes. 
the appearance of the axis and cotyledon in Ofhzoglossum, and in less. 
degree in Botrychium. This appears in an extreme form in those species. 
described by Campbell, and especially in O. pendulum, where it is possible 
that the primary shoot is permanently replaced by adventitious root-buds, 
similar to those common in the genus. These modifications in time of 
development make the reference of the parts to definite positions in the 
embryo somewhat difficult. But it seems certain, nevertheless, that in the 
less extreme forms the axis arises from the epibasal hemisphere, in close 
proximity to the intersection of the primary octant-walls. The cotyledon 
appears between the stem-apex and the root, but it is late in origin. In 
O. vulgatum it appears simultaneously with the axis, and the relation is 
so close in B. virginianum that Jeffrey states that the cotyledon, like any 
other leaf, is derived from the shoot meristem. This is interesting in 
its bearing on the theory of the cotyledon, which has been held to be 
simply a leaf of the shoot showing anticipatory development (see p. 186-7). 
The foot which is not largely developed originates from the hypobasal 
hemisphere. The position of the first root appears to be indeterminate, 
as it is in some other embryos, a fact which is interesting as upholding 
the view that it is a mere accessory to the shoot. It is referred by 
