290 Bower. — Some Normal and Abnormal 
by the transfer of the formation of flattened growths from the 
lateral branches to the main axis of the prothallus, and ac- 
cordingly the growth of the flattened surfaces has not re- 
mained so limited as is the case in the quoted species of 
Trichomanes ! This statement is supported by comparison of 
details in the two genera. 
The question now is, how do the details of structure of the 
oophyte in the species above described fall in with this 
scheme, and how far are the characters, which Goebel makes 
use of, constant in the species? The latter question vitally 
affects their importance as evidence in tracing the phylogenetic 
connections of the family. 
Apart from differences connected with abnormal de- 
velopment, the two species of Trichomanes above described 
differ from one another in details of conformation of the 
oophyte. In Tr. pyxidiferum flattened expansions have never 
been seen, the only departure from the simple filamentous 
(protonemal) form being in the case of the archegonium-bearing 
branches (archegoniophores). In the position of these latter 
there is some want of uniformity, the specimen in Fig. 13 
showing the archegoniophore to be merely a modification of a 
cell (or possibly two cells) of the filament due to septation, 
while the archegoniophore is usually a lateral appendage in 
this species ; this one specimen would therefore be a nearer 
approach than others figured by myself, or by other writers, 
to that possible type which Goebel has called ‘ die phylo- 
genetisch alteste,’ in which the sexual organs would be directly 
inserted on the branched filamentous oophyte, as in Vaucheria 
( 1 . c. p. 109). Thus, there is want of constancy in Tr. pyxidi- 
, ferum in respect of one of those characters involved in Goebel’s 
scheme. 
In Trichomanes alatmn [as also in Tr.incisum and Tr.sinuo - 
sum as described by Mettenius] great variety is to be found in 
the conformation of the prothallus. It is at times filamentous, 
and profusely branched ; at other times, and without any 
apparent regularity, it widens out into flattened expansions, 
which, as above pointed out, are sometimes lateral in origin 
