468 Bower . — Studies in the Phytogeny of the Filicates. 
sequences, viz. (1) those with their sort superficial , and (2) those with 
a marginal position of the sori. The latter are characterized by the fact — 
now demonstrated for Thyrsopteris (‘ Phil. Trans./ vol. cxcii, p. 67), Cibotium , 
Saccoloma , Odontoloma , Davallia , and Lindsay a — that the receptacle 
originates from the actual margin, while the marginal initials appear them- 
selves to give rise in certain cases to the earliest sporangia. A similar 
condition was long ago demonstrated by Prantl for the Hymenophyllaceae 
(‘Die Hymenophyllaceen,’ 1875, Taf. V), and shown by him to hold with 
singular constancy in the Schizaeaceae (‘Die Schizaeaceen/ 1881, pp. 39-46), 
while Loxsoma appears also to share this character. These groups of Ferns, 
which show natural relationship in other features also, so that they have 
habitually been ranked together, constitute a great series characterized by 
having their sori marginal. 
In a second great series the sori are as constantly superficial, having in 
their origin no direct relation to the margin of the leaf. They include the 
Gleicheniaceae, Matonineae, Metaxya , Lophosoria , and the Cyatheae, with 
their related or derivative forms, such as the Woodsieae, Struthiopteris , and 
Onoclea , Peranema , and Diacalpe , the Nephrodieae, and their derivatives and 
relative forms. Also Plagiogyria , and the related Pterideae on the one 
hand, and the Blechneae on the other, are all, strictly speaking, types with 
sori of superficial origin. However nearly these may approach the margin 
of the fertile pinnae, in no case have they been shown to be actually derived 
from it. Indeed, to describe the sori of any of these Ferns as marginal is 
not in accordance with the observed developmental facts. In all cases their 
sori arise superficially from the flattened surface of the leaf. 
The value of the distinction upon which this grouping is based as 
a phyletic criterion will depend upon its constancy. In the series with the 
sori superficial the constancy of that feature is believed to be absolute, with 
the exception of certain anomalous cases to be noted below. But the same 
cannot be said for the series which has the sori typically marginal. It will 
be necessary to look into this matter in detail. At the outset it would 
appear to be biologically a natural and advantageous adjustment that the 
sorus, if actually marginal, should be deflected for purposes of protection 
towards the lower surface of the leaf. It will be seen that this has repeatedly 
happened. But it will be shown that comparative reasons indicate these 
changes as secondary and biologically adaptive. 
Looking back to the most primitive living types of Ferns, the Eu- 
sporangiatae, it is evident that the Ophioglossaceae have typically the 
marginal position of their spore-bearing members ; also, that the Marattia- 
ceae bear as typically their sori upon the leaf-surface. But the Osmundaceae, 
which in so many features occupy a position at the very base of the Lepto- 
sporangiate Series, are as a family indeterminate in this respect. Moreover, 
their sporangia are not disposed in definite sori. In Osmunda the sporangia 
