Smith. — Cytological Studies in the Protococcales. I. 461 
always has lobed projections and never angular ones as in Hydrodictyon and 
Scenedesmus . At times the lobes are deeply constricted and the pyrenoid 
dumb-bell shaped, or there may be several lateral projections, of various 
sizes, that are almost completely cut off from the main body. These 
irregular pyrenoids are the rule rather than the exception, a few variations 
in shape being shown in Text-fig. 1. Usually there is one large pyrenoid 
near the centre of the cell, but not infrequently other pyrenoids are found 
scattered about in the cytoplasm. These vary in size from those that are 
scarcely visible to those that are 10-12 microns in diameter. The smallest 
ones are generally spherical and some distance from the large irregular 
ones, a fact which suggests their origin de novo rather than a formation by 
a division of existing pyrenoids. In the development of the cell there is 
a single pyrenoid at first, but later on others may appear. If there is 
a division of the pyrenoid it is not into two equal parts. While I am not 
willing absolutely to deny that new pyrenoids in Characium Sieboldii are 
Text-FIG. i. Outline drawings of pyrenoids showing their irregular contour, (x 2,ooo.) 
formed by the division of an already existing pyrenoid, I am of the opinion 
that in a great majority of cases the pyrenoids are formed de novo within 
the cell. 
Wille (12) states that the existence of starch in Characium is question- 
able. Fig. 12 shows the ring of starch plates usually found around 
a pyrenoid, but these differ from the starch plates in most other algae in 
that they are very much thinner. Owing to the comparatively large size of 
the pyrenoid, many starch plates are seen when it is viewed in cross-section. 
In certain preparations these starch plates are found just inside the plasma 
membrane (Fig. 12). Similar starch plates in Hydrodictyon have been 
called ‘ stroma * starch by Klebs ( 4 ), but Timberlake (10) has shown that 
these ‘ stroma 5 starch plates in Hydrodictyon are of pyrenoidal origin. In 
Hydrodictyon the ‘ stroma 5 starch plates are found close to the pyrenoid 
starch plates, but in Charachnn they are always just inside the plasma 
membrane. The reason for this may possibly be that the plates are so light 
that surface tension causes them to migrate to the periphery in much the 
