FILAMENTOUS THALLUS OF DUMONTIA FILIFORMIS. 437 
vations on the relation of the ordinary thallus of D. filiformis 
to its “basal disc.” 
The mode of origin of the upright thallus from the creeping 
one seems of sufficient interest to warrant a somewhat detailed 
description. The creeping thallus is composed of more or less 
vertical closely-set rows of cells bound together by the cohesion 
of their gelatinous walls (¢f. figs. 1 & 2, Pl. XXXV.; figs. 5 & 7, 
Pl. XXXVI.). The individual articulations * or cells are connected 
end-to-end by the characteristic Floridean pits. When branch- 
ing takes place,’it does so by means of the formation of two 
successive oblique divisions at the tip of the terminal cell of 
the row. An illustration of this is given in fig. 3, A, s.d.; 
and when the process is complete, the branches are found to be 
connected by the characteristic pits to the parent cell, having the 
arrangement shown in figs. 3 A and 4 B, br. The branches are 
very nearly, sometimes quite, equal, and the septation which gives 
rise to them must follow in very quick succession, and might even 
be simultaneous. The term subdichotomous might very well be 
adopted for this kind of branching. It is not a true dichotomy 
of the terminal cell, for only portions of it are cut off at the tip 
by oblique septa instead of the branches being formed by a 
median longitudinal wall. 
The filaments of the creeping thallus increase in length by the 
transverse division of their terminal cell; and there is no sign 
of subsequent intercalary division of the cells except in certain 
special cases which will be subsequently described. As the arti- 
culations are connected with each other by pits in the longitudinal 
direction only, it is evident that the ultimate ramifications can 
be but very remotely connected one with another. Only one or 
two cases could be found in which an articulation had divided 
parallel to the long axis in order to give rise to a branch; but 
even then it was not divided by an organically median longi- 
tudinal wall. This is of importance in relation to Fr. Schmitz’s 
views on the structure of the Floridean thallus. They were ex- 
plained in his ‘ Untersuchungen ueber die Befruchting der Flo- 
rideen’+, and subsequently upheld with but slight modification 
in his “ Kleinere Beitrige zur Kenntniss der Florideen”{. The 
* This term, though not a happy one, will frequently be used in this paper as 
the equivalent of cell, since it is generally employed in that sense by algologists. 
t Berlin, 1893, p. 4 et seg. 
¢ ‘La Nuova Notarisia,’ serie 3, 1892, p. 111 et seq. 
212 
