20 
Olive, Mitotic cLiVision of tlie liuclei of tlie Cyanopliyceae. 
ancl at tlie otüer, tliey join witü a cüromosome. Tlie extension 
of tlie cell in lengtli by osmotic forces, possibly combined witb 
tlie actual sbortening of tlie übers tüemselves is tlie probable 
cause of pnlling of tlie divided cliromosomes apart, 
It remains, tlien, to apply to tlie übrons portion between 
tlie separating cliromosomes in ügs. 6, 8, 10, 14, 17, 37, tlie 
term „connecting übersor „central spindle“ in order to com- 
plete our conception of tlie acliromatic ügure, in wüicü we liave, 
especially in tlie sliort-celled species of Oscillatoria , a spindle“ 
wüicü is not at all spindle-sliaped, but' is ratüer in tlie form of 
a niore or less tüick disc, Tliis disc-sliaped central body, as is 
seen, e. g., in üg. 10, is ünally ent in two equatorially by tlie 
ring formecl wall wbicli grows in from tlie outer wall; tliis pro- 
cess will be fnlly discussed later. It is obvious tüat a single 
centrosome would not sufüce for sneü a pecnliar, broad-poled 
ügure. As a matter of fact, liowever, no structures resembling' 
centrosomes liave been observed in any of tlie species examined. 
As we süonld expect in sncli a long-celled species as Cylin- 
drospermum 1 we find tlie nucleus also greatly elongated; and 
liere, furtliermore, tlie wliole karyokinetic ügure lias usually tlie 
spindle sliape seen in tlie liiglier plants, instead of tlie üattened- 
disc süape of tlie short celled Oscillatorias (ügs. 77. SO, 84, 85, 
89, 90). He gl er süoavs in Anabaena also central bodies wüicü 
are similar in form to tliose of Cylindrospemnum. 
Proof tliat tlie process of dhisiou is mitotic. 
In tlie foregoing discussion, we liave spoken of tlie dividing 
central body as a „mitotic figure“, and it üal been pointed out 
tliat tliis ügure possesses botli cliromatin granules and an aeüro- 
matic, übrous substance. Simply sliowing tüat cliromatin and 
aeüromatin are present far from proves, h oivever, tüat tlie pro¬ 
cess of division is one of mitosis; altüougli it would seem tüat 
merely tlie fact tüat tlie number of cliromatin granules in every 
cell is constant, süould furnisü sufficient proof. But if one 
sliould judge solely from sueü appearances as are süovm in ügs. 
2 and 3, and from many drawings given by Kolli and otliers, 
Ave may, in fact, Avitli equal rigüt, decide witü Wäger tüat tlie 
division is direct. 
Tlie nucleus of tlie Cyanopliyceae must certainly divide by 
one of tlie tivo metliods—eitlier by mitosis or by amitosis. If 
amitotic, as claimed by Wäger and otliers, tlien tlie wliole mass 
of tlie central body must undergo a simple constriction, and 
tliere süould be no spindle, and no spireme arrangement of tlie 
cliromatin. Tlie most essential act accompanying mitotic division, 
on tlie otlier liand, is tüat tlie cliromatin granules are eacli split 
in two, so tliat tlie ; ,daugüter nuclei receive precisely equivalent 
portions of cliromatin from tlie motüer-nucleus u (Wilson, 1900, 
p. 70). Usually, moreover, dming tlie preliminary stages of mi¬ 
tosis, tüe cliromatin granules are arranged along a more or less 
convoluted tliread, wliicli, wüetüer continuous or discontinuous. 
