32 
Olive, 5Iitotic division of tlie nuclei of ilie Cyanopliyceae. 
proper staining with iron haematoxylin, tliat tlie chromosomes 
in fig. 17, e. g., are globular or nearly so, and tliat- tliey are 
quite separate and distinct froni tlie acbromatic portion. Tlie 
acbromatin in tbis instance is apparently most dense in tlie 
inimediate proximity of tlie chromosomes. (pp. 21, 22.) 
(3) Number of chromosomes constant. One of tlie 
strongest evideiices tliat a mitotic division must take place is 
tliat tlie number of chromosomes is constant for the same 
species. In Gloeocapsa and Nostoc , there are 8. In two of the 
species of Oscillatoria studied, in Phormiclium and Calothrix , 
there are 16; while in Oscillatoria princeps and in 0. Froelichia , 
there are probably 32. In side view, the longitudinal sections 
of cells with 16 chromosomes show usually 3 or 4: tliose with 
32, as many as 8. (p. 11.) 
(4) Longitudinal fission of chromosomes. This con- 
stitutes the most necessary accompaniment of mitotic division. 
In the case of Gloeocapsa , it is plainlv evident that a longitudinal 
fission of the spireme thread is taking place, or has alreadv 
been accomphshed, in figs. 67, 70, 71, 74, 75. It is equally 
evident that- an equatorial fission of the central body is occur- 
ring in figs. 6, 8, 10, 13, 34, 37, 80, etc. Since we find in such 
cases as figs. 8 and 13, indisputable evidence of the doubling 
of the number of chromosomes, one must conclude that each 
individual chromatin granule is divided longitudinally in a plane 
parallel to the cross partitions. The fission of all tlie chromo- 
somes in a single nucleus probably does not take place simul- 
taneously, for it appears to begin at the two extremities or 
outer edges of the spireme and thence to advance to the middle, 
in the same manner as the progressive fission of the central 
body (figs. 8, 13, 74, 75). (pp. 23, 29.) 
The fact, observed by many investigators, that the central 
bodies of the vegetative cells of the Cyanopliyceae have no 
nuclear membrane is due, in the opinion of the writer, to their 
continuous mitotic activity. It is probable that, if tlie nuclei 
could be. made to enter a resting condition, they would then 
form both karyolymph and membrane. This probability is 
rendered the more creclible by certain facts: first, that the nuclei 
of Oscillatoria , even in active filaments, occasionally seem to 
show nuclear sap and a poorly defined limiting membrane. Se- 
condly, by the fact, observed by Wäger, that the nuclei in 
young cells are often limited toward the cytoplasm by a „vacuolar 
membrane u ; and finallv, because in spores and lieterocysts we 
find nuclei in the usual resting condition. 
It will be of filterest, then, if the vegetative cells of the 
Cyanopliyceae are in a state of continuous mitotic activity, to 
determine how near to a resting condition tlieir nuclei come. 
In Gloeocapsa , in the very youngest cells (figs. 62, 64, 69. 76), 
one may see, more or less clearly defined, tlie eiglit irregularly 
disposed chromatin granules, imbedded in an achromatic sub- 
stance. Probably tlie next step in the karyokinetic clianges is 
