276 DR, LILY BATTEN ON 
number of rhizoids, which may or may not develop dises at their distal ends, 
depending on the nature of the substratum. The formation of these rhizoids 
presents a feature of considerable interest, as a gradual increase of complexity 
is traceable, which can be correlated with the anatomy of the species. 
The rhizoids are thick-walled, and are formed principally by the lateral 
proliferation of primary pericentral siphons in ecorticate forms such as 
P. macrocarpa. In corticate forms, however, where the thallus is bounded 
by small cells, the rhizoids are formed from these latter, and not from the 
siphons themselves. Fig. 5 (Pl. 22) shows the development of a single 
rhizoid from a pericentral siphon in P. macrocarpa. The formation of rhizoids 
(rom the corticating cells of P. nigrescens is shown in fig. 6 (Pl. 22). When 
the rhizoids ramify among the small filamentous alge and debris on the surface 
of a rock, they do not develop discs at their extremities (Pl. 25. fig. 64). If, 
however, the tip of the rhizoid comes into contact with some firm object such 
as a fragment of rock or shell, or some larger alga, a flattened expansion begins 
to form. This eventually becomes lobed, but no septation takes place (PI. 22. 
fig. 2). The formation of a dise is a direct response to the stimulus of contact. 
If the rhizoid attaches itself to a sand grain, the dise bends over and partially 
encloses the grain. Derick (20. p. 251) states that the rhizoids of P. violacea 
are separated from the corticating cells by a wall, but I have never observed 
a wall at the base of a rhizoid in any species of Polysiphonia that I have 
examined. The rhizoids are unicellular, and are usually unbranched, but in 
P. urceolata an occasional tendency to branch may be observed although no 
septation takes place. 
Where Polysiphonia grows on another alga epiphytically, as, for example, 
P. violacea on Chorda filum, rhizoids are formed which adhere to the surface 
of the host by means of their expanded dises. Penetration does not take 
place, but the host is frequently a little distorted, and has fewer chromato- 
phores in the part of the thallus below the dises. 
The question naturally arises as to the exact means by which the dise is 
attached to the substratum. It must adhere either like a sucker, or by the 
purely mechanical means of fitting in among the crevices of the substratum, 
or by the secretion of a cement. Careful observation and sections have failed 
to reveal any lifting of. the centre of the dises either individually or in the 
aggregate, and the presence of large forms like 7". elongata on a perfectly 
smooth water-worn pebble near a jetty, seems to indicate, that purely mechan- 
ical adherence by fitting in with the crevices of the substratum is an in- 
adequate explanation. Where the substratum is irregular, this latter method 
no doubt aids in the attachment, but it is not sufficient when acting alone on 
certain substrata. It appears, therefore, that the thickened dises consist 
partly of a cementing substance which helps to fix the alga. It was 
suggested that there might be some chemical combination of the secretion 
with the substratum, but I have been unable to find any trace of chemical 
