Smith. — Cytological Studies in the Protococccdes . II. 475 
there may possibly be a specially active period of nuclear division just prior 
to zoospore in Hydrodictyon. The appearance of the nuclei of Pediastrum 
at this time is quite different from the appearance at other periods. Before 
this period of nuclear division sets in there is a conspicuous nucleolus, the 
chromatic material being quite faint and lying near the nuclear membrane 
(Figs. 19, 22) ; but after the rapid nuclear divisions the peripheral chromatin 
granules are very conspicuous (Figs. 9, 13). 
While these nuclear changes have been going on the pyrenoid has 
disappeared. In the various algae whose spore formation has been studied 
the behaviour of the pyrenoid is not uniform. In zoospore formation 
in Hydrodictyon ( 7 , 17 ) and Cladophora, and in spermatogenesis in Sphaero- 
plea ( 5 , 6), the pyrenoids generally disappear before spore formation sets in. 
On the other hand, in the zoospore formation in Characium ( 15 ) and Tetra- 
spora ( 9 ), and colony formation in Scenedesmus ( 13 ) and Tetradesmus ( 12 ), 
in which the products of cell- division are the morphological equivalents of 
zoospores, the pyrenoid does not disappear, but remains unchanged in one 
of the daughter-cells. I find in Pediastrum that it is only in exceptional 
cases (Fig. 12) that the pyrenoid persists after the process of cleavage com- 
mences. Stages were found in the disappearance of the pyrenoid, and 
it was noted that the ring of starch plates disappears while the body 
of the pyrenoid becomes diminished in size, frequently becoming quite 
angular just before its final disappearance (Figs. 11, 12). Cases were not 
found similar to those described by Askenasy ( 1 ), in which the first cleavage 
plane cuts through the pyrenoid, dividing it into either equal or unequal 
parts. 
The first cleavage plane may be so placed as to include either the long 
or the short axis of the cell. Whether it is formed by a furrowing in of the 
plasma membrane or by the formation of a row of vacuoles could not 
be satisfactorily determined, but I am inclined to think it takes place by 
the former method. In a very few instances cleavage planes that did not 
extend clear across the cell were found between two neighbouring nuclei 
(Fig. 13). These suggest conditions similar to those found in Hydrodictyon 
by Timberlake ( 17 ), in which cleavage furrows appear at various places 
between the nuclei. Unlike the furrows in Hydrodictyon those in Pediastrum 
do not branch. The protoplasts formed by the cleavage furrows are always 
angular, and the number of nuclei in each varies from one to four (Figs. 14, 
15, 17). The process of cleavage continues until the entire contents of the 
cell have been cut up into angular uninucleate pieces (Fig. 17). Mention 
has been made of exceptional cases in which the pyrenoid persists after the 
beginning of cleavage. In two instances, when the triple stain was used, a 
small red-staining body was found in one of the uninucleate protoplasts 
(Fig. 16), which may possibly be the last vestige of the pyrenoid. The 
nuclei in the cells that are undergoing cleavage have much the same 
