Filicales. 23 
arose by septation of a unilocular sporangium ; for in the Lycopods 
the archesporium arises from a row of cells and additions seem to 
be made to it later (3). Yet among the Lycopods there can be no 
question of septation, since the sporangia are solitary. It would 
seem that in bulkier sporangia, such as those of the Equisetales and 
Lycopodiales, the archesporium arises from more than one cell. In 
0phiogloss iu n more than one cell still contributes to the formation 
of the archesporium (4). In the Marattiaceae this contribution has, 
in accordance with the smaller size of the sporangia, become 
inconstant and irregular; in the leptosporangiate ferns, the sporangia 
of which are as a rule the smallest found among the Pteridophyta, 
the archesporium originates from a single cell. This is presumably 
due to their reduction in size. 
Besides the fossil and developmental evidence, Professor Bower 
relies on an analogy with the anthers of certain Angiosperms that 
have clearly become multilocular by septation (4). He states that 
these show a close parallel to the details described for Dancen, and 
afford strong support of his theory that the synangium is a septate 
sporangium (5). But as he himself remarks, in a later paper, 
examples of fusion of sporangia also occur among the Angiosperms 
(7). In view of this fact and of the remoteness of the Angiosperms 
from the Marattiaceae the analogy with the septate anthers should 
not be allowed to carry much weight. Thus, though it is not certain 
whether free or coherent sporangia are the more primitive, the 
evidence on the whole points to the priority of the sorus with free 
sporangia. 
It seems absolutely useless to look for the nearer allies of the 
Marattiaceae among orders so unlike them as the Botryopterideae, 
Hymenophyllaceae, Gleicheniaceae or Schizaeaceae. The polycycly 
of the Matonineae certainly recalls that of the Marattiaceae, but it is 
associated with solenostely and almost certainly arose independently 
of the polycycly of the latter order. 
The origin of the dietyostelic polycycly of the Cyatheaceae and 
Polypodiaceae may be traced within the limits of these orders, and 
as it is impossible to derive the Marattiaceae directly from either of 
them, it is clear that the close similarities of the stelar structures 
are due to parallel development. Among the differences forbidding 
such a derivation may be mentioned the position of the protophloem 
(exarch in the Cyatheaceae and Polypodiaceae, endarch in the 
Marattiaceae), the polyarchy of the roots of the Marattiaceae, and 
the contrast afforded by the free annulate sporangia of the Tree 
Ferns and Polypodiaceae, with their basipetal or mixed succession 
