352 Bower . — The comparative examination of the 
that the structure of the wings in these plants conforms, at 
least in the upper parts of the frond, to that of the Polypo- 
diaceae. I have also noted that in Aneimia the wings of the 
basal part of the leaf are of the more bulky type : thus the 
Schizaeaceae appear, as regards the wings, to show no advance 
of complexity upon the Polypodiaceae. 
Passing on now to the Osmundaceae, the structure of the 
young wings has been most exactly observed in Todea bar- 
bar but observations have also been made on Todea superba 
and T . hymenophylloides ( =pellucida :), and on Osmunda re- 
galis : it may be stated at once that (excluding the £ filmy ’ 
Todeas) the result is to find the wings in this family more 
complex than those of the true leptosporangiate Ferns. 
Taking first Osmunda regalis , it has been stated by Kny 1 
that marginal cells are present on the wings, from which, as 
in the Polypodiaceae, segments are cut off obliquely and 
alternately. This may be the case in young plants, but 
numerous sections which I have made from various parts of 
the leaf of this species fail to bear out the statement for well- 
grown plants : the construction of the young wing is found in 
old plants to be essentially similar to that to be described for 
Todea barbara ; sometimes it is true that a series of larger 
cells may be seen from without to occupy the margin, while 
in section a wedge-shaped cell may be seen to lie at the 
extreme outer edge of the wing, but this is not constant, and 
when so found it is of relatively small size, while the segments 
never show that regularity of arrangement which is so charac- 
teristic of the Polypodiaceae. Knowing as we do that the 
apex of the leaf of Osmunda is occupied by a three-sided 
apical cell, and that the position of the wings corresponds to 
two of the angles of that cell 2 , and that the edges of two 
rows of segments contribute to the formation of each wing, it 
would be contrary to all expectation that a definite marginal 
series should be found. In fact, I look upon the absence of 
1 Parkeriaceae, p. 40, footnote. 
2 Compare Phil. Trans., 1884, Plate 37, Figs. 4, 5, 6. 
