SOME ENDOPHYTIC ALGE. 291 
through the frond, however, shows that as a rule in this host 
they are not intercellular, but run along in the outer layers of 
the cell-walls. 
Fig. 1 (Pl. 12) shows the appearance of Endoderma as seen 
from above in N. Hillie. The jointed filaments, of which it is 
composed, branch and spread out over the frond in a radiating 
manner and at first follow the outlines of the cell-walls as seen 
from above. Why they take this course is not clear, presumably 
it is because they find it easier to penetrate along the lines of 
junction of the cells; and this may possibly be accounted for by 
some slight difference in the character of the cellulose. In 
somewhat older stages, owing to repeated branching and further 
growth, the Endoderma-cells may lie so closely together that 
they form an almost pseudoparenchymatous mass (lower part of 
fig. 1). The averaze width of the cells in the younger portions 
is 5-6 p, and the length is generally two to three times as much. 
The terminal cell is usually considerably longer. In still older 
parts the cells become rounded and oval in shape, previous to 
zoospore-formation. 
A transverse section of the Nitophyllum frond shows that 
while growth is mostly parallel to the surface, the endophyte 
occasionally sends filaments down into the thallus as shown in 
fig. 2. This is somewhat unusual for an Endoderma, for as a 
rule the filaments do not penetrate below the onter membrane. 
Oltmanns (7. p. 805) for this reason classes Endoderma with his 
first group of Endophytes and not with the second, 
With regard to the minute structure of Endoderma, a single 
parietal chromatophore is found in each cell and there is one 
pyrenoid. (The presence of a single pyrenoid is a point of 
agreement with Reinke’s E. viride in Derbesia, whereas in 
Huber's E. leptochete in Ceramium there are three). The cell- 
membrane is not so obvious as in the case of other species, but 
by the use of reagents it can readily be brought into prominence. 
Endoderma grows by apical growth, as is the case in all endo- 
phytic filaments. 
When zoospore-formation is about to take place, the ordinary 
cells of the filament separate from each other and round them- 
selves to form zoosporangia. Apparently any or all of the older 
cells may in this way form fruit. The contents of the cell divide 
up by successive division until usually eight zoospores are formed ; 
LINN. JOURN.—BOTANY, VOL. XXXVII. x 
