250 BOTANICAL GAZETTE [APRIL 
sometimes the appearance of a zone or band of darker staining mate- 
rial extending throughout the pyrenoid body (jig. 22). The ends of 
this zone may be:split so as to give a Y-shaped figure in cross section. 
In a Closterium that had been fixed at night I have observed a large 
vacuole in the pyrenoid (jig. 23). 
In a great number of cases the denser portions are numerous and 
are scattered apparently irregularly throughout the pyrenoid. A 
careful examination will show, however, that each dense area is 
opposite a starch grain, while the lighter portions are opposite the 
clefts between the grains, appearing merely as continuations of the 
latter (figs. 10, 11). This concentration of pyrenoid material so 
near to the place where starch is being laid down in the starch 
grain is an interesting fact, indicating that even though the rudi- 
ment of the grain may be formed as TIMBERLAKE has suggested 
some kind of organic connection is still retained between the young 
grain and the part of the pyrenoid nearest to it. The same relation 
might be expected to exist also if the grain is wholly a deposition 
product of the pyrenoid. How long this connection is retained can- 
not be determined. In any case, this increased concentration of 
the pyrenoid material in certain definite regions would seem to indi- 
cate that the pyrenoid is not a mere mass, but that it has a structure 
correlated more or less wholly with its function. 
In many cases the pyrenoids seem to be in the process of cutting 
off pieces or of breaking up into a number of pieces. These may be 
in the form of a segment split off from one side (figs. 16, 17), or the 
entire pyrenoid may break up into lens-shaped segments (figs. 253 
26). Where the body of the pyrenoid itself shows a more densely 
staining band in cross section, these segments are cut off from one 
or both sides of this region (fig. 16). If the pyrenoid is made up of 
denser bodies placed opposite starch grains, the cutting off may 
occur along any of the lighter, less dense lines (fig. 10); however, 
hardly ever more than two of these segments are cut off and usually 
only one of them at any one time. The process may be traced in all 
its stages, from a lightly stained zone in the pyrenoid body to a small 
cleft, and finally to the separation of a free segment. TIMBERLAKE 
was able to find in the pyrenoids of Hydrodictyon that the segments 
changed their staining reactions from the bright red of the pyrenoid 
