Olive, Mitotic division of tke nuclei of tke Cyanopliyceae. 
43 
Fig. 56. Sliowing cyanopliycin granules only, stained witli safranin. 
Fig. 57. A cross section, sliowing in tlie central body both ckromatin and 
acliromatin. 
Fig. 58. A spore (?) wliicli shows a nucleus in resting condition. 
Fig. 59. Another spore, sliowing abont 8 cliromatin grannles. Flem. 
Fig. 60. A similar preparation. 
(Figs. 61—76, Gloeoeapsa polydermatica Kützing, all stained with Heid. 
Fig. 61. A dividing cell, witli its surrounding gelatinous wall. Tlie central 
body is too deeply stained. 
Fig. 62. A yonng cell, sliowing abont 8 cliromatin grannles and tke 
acliromatic portion. 
Fig. 63. A dividing cell, sliowing a deeply stained epireme tkread. In tke 
vacnolar spaces in tke cytoplasm, are tke „slime globnles“ (of 
Sckmitz). 
Fig. 64. A yonng- cell, in wliicli tke central body sliows clearly 8 cliromatin 
grannles and an acliromatic portion. 
Fig. 65. A cell, apparently in a similar condition to tkat sliown in fig. 64. 
Fig. 66. Skows abont 8 cliromatin grannles, and at one side of tke central 
body, a large globale. It is possible tliat sncli globnlar bodies are 
slime globnles; altliongli tliey do not seeni to be stained as tke slime 
in Oscillatoria and in otlier instances witli metkylene blne. 
Fig. 67. An older cell in wliicli abont 11 or 12 ckromatin grannles are 
sliown. 
Fig. 68. A spireme stage, in wliicli tke simple spireme tkread kas a dis- 
tinctly spiral form. 
Fig. 69. A yonng cell, sliowing only abont 7 ckromatin grannles. Certain 
fibrons projections from tke central body, wkick extend into tke cyto¬ 
plasm between tke food grannles, are somewkat stained by tke kaema- 
toxylin. 
Fig. 70. A cell in a state of division in wkick tke spireme tkread is double. 
We can now connt abont 16 ckromatin gramdes. 
Fig. 71. Sliowing tke pecnliar manner in wkick tke divided spiremes sepa¬ 
rate, tke one being drawk into tke npper dangkter cell, tke otlier 
into tke lower. It is possible tkat tkere are cytoplasmic übers, corre- 
sponding to tke mantle übers, attacking tke spireme to tke end of 
tlie cell, and exerting a pnll as tke cell elongates. Tkese are not 
evident, kowever. Jndging from tke ügnre, tkere appears to be 
an actnal Üowing of tke spireme snbstance into tke dangkter cell. 
Tke central spindle between tke separating spiremes is obvionsly very 
little developed. 
Fig. 72. A spiral spireme tkread in wkick we can connt abont 7 or 8 
ckromatin grannles. 
Fig. 73. A constricting cell in wliicli tke two dangkter spiremes kave com- 
pletely separated. Abont 8 ckromatin grannles can be connted in 
eacli dangkter cell. 
Fig. 74. A spireme wkick appears to be Splitting at tke two ends. 
Fig. 75. Anotker instance, in wkick tke Splitting of tke spireme at tke 
two ends is even more obvions. 
Fig. 76. A step fnrtlier advanced tkan in üg. 73, in wkick tke constriction 
plane kas completely divided tke cells. wkick kave yet become ronn- 
ded off. 
(Figs. 77—103, Cylindrospermum stagnale Bornet and Flakanlt.) 
Fig. 77. A preparation freslily stained witli metkylene blne. Tke keterocyst 
sliow two niinnte, dark blne slime globnles; tke spore, one (sometimes 
several); tke vegetative cells eack one to several. Tke grannlation 
in tke cytoplasm of tke spore is liere qnite evident. 
Fig. 78. A ülament, similarly stained, in wkick tke spore cell is no larger 
tlian tke vegetative cells. Tke keterocyst bears at its end several 
bacteria-like bodies. 
