Bower . — Studies in the Phytogeny of the Filicales. 461 
dition, which thus is seen in both species, is not uniform for either. Both 
are variable. It is, however, a condition which would naturally be antici- 
pated in Ferns of Gradate origin, in which the receptacle was flattened and 
the sorus (with a tendency to becoming a mixed one) compressed laterally 
between the indusial lips. The sporangia, in such a case, would be forced 
to open upwards, and for this a vertical annulus would be the proper 
mechanism, while the induration of its basal region would be mechanically 
ineffective. Such conditions as those seen in the Ferns named give a strong 
support to the view that the Lindsay a-Davallia type has been derived from 
some Gradate, and probably from a Dicksonioid source. 
Davallia. 
Having touched upon Ferns so nearly related to Davallia as those last 
mentioned, it is necessary briefly to allude to this genus. But it would be 
impossible in this memoir to do more than take a single sample of the species, 
and examine its sorus developmentally. A thorough comparative study of 
the sori of the genus is urgently wanted, which would probably lead to the 
recognition of many details conducive to a phyletic grouping of the Ferns 
of this affinity. This can only be suggested at present in the roughest out- 
line. It is already known that the sori of certain Davallias show a mixed 
condition, but that in their earliest stages the first sporangia arise in 
a median position, which may be taken as a reminiscence of a basipetal 
sequence in their ancestry (‘ Phil. Trans.,’ vol. cxcii, p. 76). This fact was seen 
in D. Griffithiana , which later shows very clearly a mixed condition (1. c., 
Figs. 134, 135). It must remain for further inquiry to show, for the various 
species, what balance there may be between the originally basipetal and 
the derivative mixed condition of the sorus in the genus at large. 
But the question of present interest for us is the position of the sorus 
at its initiation : is it in the first instance marginal ? This point has been 
examined in D. pentaphylla , Bl., a species in which, when mature, the sori 
are distinctly intramarginal, though not in so high a degree as in some 
species. Sections through very young sori prove that the origin of the 
sorus is nevertheless truly marginal, as in the related genera above described. 
Fig. 23 shows a very young state in D. pentaphylla, in which the receptacle 
lies between the two marginal flaps, and itself holds a marginal position. 
The two indusia differ, however, slightly from the first in bulk and in 
structure. That which is to be the lower (/) takes the precedence while 
young, but it is the less bulky, running out later into a single layer of cells 
(Fig. 24, /). The upper lip, which extends, as in Lindsay a and Saccoloma , 
as an apparent extension of the leaf-surface when mature, hangs back 
slightly at first (Fig. 24, u) } but is more bulky, consisting throughout of 
several layers of cells. It shows at first a marginal segmentation, but this 
soon ceases, and the greater part of the mature flap, which then appears like 
