Bower.— Studies in the Phytogeny of the Filicales . 467 
retained the median dehiscence with modifications which are peculiar to 
itself. In fact it exhibits a state transitional between a Schizaeaceous and 
a Dicksonioid type. Such a progression would in the peripheral series of 
Ferns run parallel to that seen in the superficial series, as exemplified by 
Gleicheniaceae, Lophosoria , and the Cyatheoids. 
Other lines of comparison accord with this conclusion. The dermal 
appendages on the rhizome of Loxsoma are stiff brown bristles (‘ Borsten ’), 
with longitudinal, as well as transverse, septation towards the base. They 
resemble the stiff brown bristles of Thyrsopteris . But flattened ramenta, 
which are so marked a feature in the Cyatheae, are absent in both. The 
solenostelic structure described in detail by Gwynne- Vaughan (‘Ann. of Bot./ 
vol. xv, p. 71) corresponds to what is seen in many relatively primitive 
Ferns. It may, however, be specially noted that it is characteristic of the 
creeping Aneimias of the section Aneimiorrhiza (cf. Boodle, ‘ Ann. of Bot./ 
vol. xiv, p. 359). The undivided leaf-trace of Loxsoma , in particular, has 
been compared by Gwynne-Vaughan with that of the Dennstaedtiinae on 
the one hand, and with that of Aneimia on the other. But, as regards the 
Cyatheaceae in the widest sense, to which von Goebel would refer Loxsoma , 
it may be remarked that in them the tendency is to subdivision of the leaf- 
trace, extending in Cibotium and in Cyathea to the extreme leaf-base. In 
Loxsoma , however, the leaf-trace is undivided, and continues upwards in that 
condition, as it does also in Aneimia. Thus, the rough vascular structure 
would indicate a nearer similarity to the Schizaeaceae than to the Cyatheaceae 
in the widest sense. Accordingly, in attempting to place Loxsoma phyleti- 
cally, it should, in my opinion, be regarded as the sole representative of 
a distinct tribe ; and its position appears to lie about the limit between the 
Simplices and Gradatae. And as its sorus is marginal, it takes its place 
in the marginal series, between the Schizaeaceae and the Dicksonia-Denn- 
staedtia series. It may be held generally to be an up-grade type, though 
the biological requirements of the sorus have led to a reduction of one side 
of the annulus. 
A fossil referable to a somewhat similar position has recently been 
described by H. H. Thomas ( Stachypteris Haiti , a new Jurassic Fern, 1 Proc. 
Camb. Phil. Soc./ vol. xvi, p. 610), in which spike-like marginal sori are 
found. There is some uncertainty as to the details of its sorus, but sufficient 
is known to countenance its reference also to an intermediate position 
between Schizaeaceae and the Dicksonioids, somewhat similar to that now 
ascribed to Loxsoma. 
On the Phyletic Value of Soral Position 
A considerable number of relatively primitive Ferns have been dis- 
cussed in the above pages, chiefly as regards their external morphology, 
their coarse anatomy, and their soral condition. They fall into two 
K k 
