Bower. — Studies in the Phytogeny of the Filicales . 469 
are in serried ranks which are essentially marginal, though they may spread 
on to the surfaces. In Todea they are seated on the lower surface. The 
Osmundaceae are , however , the only living family which shows this indefinite- 
ness . Certain isolated exceptions occur elsewhere, but no other whole family 
is characterized thus. All are either characteristically marginal or superficial. 
It may be held from these facts that while the Eusporangiate families 
had already resolved this question of the position of their spore-bearing 
organs relatively to the leaf-margin, the Osmundaceae represent a very 
ancient type of Leptosporangiate Ferns in which the question was still 
undecided. This family may in fact in some sense suggest, or even represent, 
the indeterminate ancestry from which the Leptosporangiate Ferns, with 
their two distinct sequences, the marginal and the superficial, originally 
sprang. 
We have seen that the superficial position of the sori is, with certain 
exceptions, constant in those Ferns which have adopted it. The exceptions 
must be considered in their bearing upon the constancy of such characters 
for phyletic comparison. Examples are found in Polystichum aculeatum , (L.) 
Schott, var. 37, anomalum , Hk. & Arn., and certain aberrant forms of 
Scolopendrium vidgare. The former of these is a variety of P. aculeatum , 
which grows on the Horton Plains, Ceylon, and was described in 1856 by 
Hooker and Arnott. The peculiarity consists in the appearance of the 
sori ‘ usually on the superior face \ Hooker states (‘ Sp. Fil.,’ iv, p. 27) that 
‘ the species is in cultivation at Kew, and retains its usual peculiarity of bearing 
the sori on the upper or anterior side of the frond Here it would appear 
that there has been in some way a transference of the stimulus, whatever it 
be, to sorus-formation to a spot where it is not typically present ; for the 
sori themselves appear to be quite normal, except for their position. A 
somewhat similar state, though accompanied by malformation of the leaf, is 
to be found in some of the extravagant monstrous forms of Scolopendrium 
vulgar e. Such cases, however, do not appear to me to throw any satisfactory 
light on the phyletic story. They may serve some time as points of attack 
on the question what it is that determines soral development at all. But 
the very suddenness of their appearance and the isolation of their occurrence 
stamps them as anomalies, rather than as dependable phyletic signs. 
Very much the same may be said of the occasional occurrence of the 
sori in Deparia Moorei superficially. In the species D. prolifera they 
appear to be as a rule marginal, though in some Sandwich Island specimens 
they incline to the lower surface, as in Davallia } A marginal position is 
the rule also in D. Moorei ; but isolated sori are found not uncommonly 
upon the upper surface, and even at a distance from the margin. The 
absence of intermediate steps in D. Moorei suggests again an anomalous 
transfer of the stimulus to soral formation in the young primordium, rather 
1 Compare Mettenius, Farngattungen, vi, p. 63, and PL VI, Figs. 17-20. 
K k 2 
