in Rhodymeniales : II. Delesseriaceae. 19 1 
me that it is effected in precisely the way in which it takes 
place in other families of Florideae, that is to say, by ex- 
clusively apical growth. As to the growth in area, where, 
by inference, it is to be concluded that Schmitz believed 
growth by intercalation to take place, I have seen no 
evidence of this either, but do not claim to have given, as yet, 
adequate attention to the phenomena figured by Naegeli and 
Schwendener. 
From all this, however, it is manifest that in the way in 
which the thallus of D. sinuosa arises, it is more akin to Nito- 
phyllum than to the species of Delesseria already described. 
When the arrangement of the procarps in D. sinuosa comes 
to be considered, the divergence from Delesseria is equally 
striking. They are found to be distributed in great numbers, 
without any regularity, in the marginal portions of the thallus. 
The ultimate ramifications of the veins are obscure lines, 
traceable only with the aid of the microscope, where a row of 
axial cells gives off a pericentral cell above and below. 
Between these veins the thallus is only one cell thick. It does 
not appear that the procarps in their inception are related 
in any way. to the veins, although a strong vein always arises 
in connexion with a fertilized procarp. When the procarps 
arise, an axial cell cuts off a pericentral cell above and below, 
and from each pericentral cell there springs a 4-celled 
carpogonial branch, which curves in a characteristic manner 
before the trichogyne emerges slightly at the corresponding 
surface (Fig. 29). By the time the carpogonial branches are 
formed, the pericentral cells from which they originate divide 
again and give off externally other cells, and a minute 
swelling in the thallus is the consequence. The trichogynes 
emerge on the slopes of this swelling. When fertilization 
fails there is no further development, and great numbers of 
such unfertilized procarps may be found among the few 
which are fertilized. The first indication of the development 
of a procarp into a cystocarp is afforded by the staining 
properties of the central cell concerned. This extends 
gradually to the neighbouring axial cells. At this time too 
