448 Bower .— Shi dies in the Phytogeny of the Ft tic ales. 
position, as their oldest type, while Cryptogramme will retain the position 
in near relation to it, as has been correctly recognized by Diels. 
As regards the arrangement of the species within the genus, though 
there are no quite distinctive features of priority, it appears that P. semi- 
cordata is in certain characters more primitive than P. pycnophylla and its 
closely allied species (or forms), P. scandens and P. gtauca. This is 
indicated by the prevalence of dichotomy of the axis in the former, while it 
is, as far as observed, absent in the latter, its place as a method of amplifi¬ 
cation of the shoot being taken by stolons in the Eastern species. The 
more marked serration of the leaves in P. semicordata , while in the Eastern 
species they are more nearly entire, may be held to be a nearer condition 
to a more richly branched ancestry, while the leaf-form of the Eastern 
species is more completely condensed ; the Eastern species show a fuller 
development of the pneumatophores, which are a special adaptation ex¬ 
tended here to the upper leaf, while they are present only on the leaf-base 
in P. semicordata ; and this goes along with the more highly developed 
covering of mucilaginous hairs in the Eastern and Western species. These 
are only indications it is true, but they all point to P. semicordata as being 
the more primitive species. On the other hand, the division of the leaf- 
trace in P. semicordata into three strands in the basal region of the leaf 
must be borne in mind, though the bearing of the fact is obscured by their 
subsequent junction below the lowest pinnae into a single strand again. 
There is also the fact of the absence of the median protoxylem-strand, 
which is present in P. pycnophylla and glauca , but again it is not clear what 
interpretation is to be put upon this fact. 
Lastly, the observations on Plagiogyria have their bearing on the 
relations of those three grades of soral condition characteristic of the 
Simplices, Gradatae, and Mixtae. When these types were first recognized 
by me, it was clearly laid down that the three divisions illustrate steps in 
the evolution of the Filicales, but that they were arrived at not by pro¬ 
gression along a single line of descent, but by similar adaptations making 
their appearance in more than one evolutionary sequence, and the results 
were grouped according to their common adaptation (‘ Studies,’ v, p. 123 ). 
It was specially insisted on that, though at that time no evidence was 
available to show that the ‘ mixed ’ condition of the sorus had been acquired 
by modification of the ‘gradate’ condition, ‘it is possible that a mixed 
condition of the sorus may have arisen also by interpolation of successive 
sporangia without order in a typical simultaneous sorus, such as that of 
Angiopteris or Gleichenia ; but evidence of this has not come to hand ’ ( 1 . c., 
p. 124). Since this was written in 1899 such evidence has been acquired in 
the case of the genus Diptei is, already quoted; here in the broad-leaved 
D. conjugata sporangia of different ages are found side by side in the same 
sorus, though in other species of the genus, having the more primitive 
