09 
axis of the cell, usually in its upper part, and giving off a number of lobes in 
several directions towards the periphery of the cell. These lobes proceed further 
along the periphery of the cell and may together form a more or less interrupted 
cylindrical parietal layer. In the middle of the central body lies a pyrenoid, which 
is thus situated in the axis of the cell. This form of chromatophore shows a par- 
ticularly fine development in Ch. immersa, where the lobes are very long and di- 
stinct (fig. 57); but it must be confessed that in this species the pyrenoid is not 
always central (fig.57 B,C). In two species, Ch. Dumontiæ (fig. 52) and Ch. cytophaga 
(fig. 50), which also have stellate chromatophores, I have not been able to see any 
pyrenoid and must therefore suppose that it is wanting. In other species the chro- 
matophore is an entire or somewhat lobed parietal plate containing a pyrenoid 
which is thus excentric in the cell. The pyrenoid is always prominent in the 
interior of the cell, and it is sometimes so large that it reaches almost to the op- 
posite part of the chromatophore; when seen in profile, however, it is always easy 
to determine that it is parietal (figs. 30, 34, 54). Only in some species with very 
thin filaments it may be difficult to decide if the pyrenoid is axile or parietal, and 
transitions may perhaps occur. In Ch. Polyidis the chromatophore has a very pe- 
euliar structure, which I have unfortunately not been able to fully elueidate; it 
seems to be single but becoming very much branched (fig. 60). A third (or fourth) 
type of chromatophore occurs in Ch.efflorescens and pectinata, where each cell con- 
tains usually more than one spiral-shaped or more irregular band-like chromato- 
phores (figs. 64, 66). — In pyrenoids of Ch. immersa treated with picric acid an an- 
gular body, probably a crystalloid, was observed (fig. 57). 
The cells always contain a single nucleus lying almost in the central part of 
the cell, thus at a lower level than the pyrenoid. In some cases it is easily visible, 
even in the living state (fig. 30 C), in other it is concealed by the chromatophore; 
in Ch. immersa it is even sometimes found in a hollow in the mass of the chro- 
matophore (fig. 57). 
In nearly all the species hyaline, unicellular hairs occur at the ends of the 
filaments, which, when the filaments develop farther, are pushed aside, while the 
filament continues its way in the same direction as before, but really sympodially. 
This development has been pointed out by Kyrın (1906 and 1907) in some species, 
and I have found the same in all the species with hyaline hairs examined by me. 
The hair arises as the terminal cell of the filament, being however much narrower 
than the usual cells and containing no chromatophore but protoplasm and a nucleus. 
In the out-growing hair the protoplasm is collected towards the upper end of the 
cell and decreases in bulk on the lengthening of the hair. In some cases, how- 
ever, e. gr. Ch. rhipidandra, the hair is not pushed aside but retains its terminal po- 
sition, and the filament then makes a bend for each hair it produces, with the 
result that the sympodial nature of the filament becomes very evident (figs. 20, 21). 
But even in the cases where the hair is early shed, this process often causes a 
more or less pronounced obliquity of the upper end of the cell (fig. 18). The 
11* 
