Vol. 2] Gardner. — Cytological Studies in Cyanophyceae. 271 



Ehrlich's haematoxylin is excellent for this purpose, since it does 

 not stain anything deeply except the granules (when they are 

 present in the young spores) and the chromatin. It has been 

 found by the use of this stain that these granules disappear be- 

 fore the spore reaches maturity. They do not disappear at once, 

 but become gradually smaller and finally disappear entirely. 

 These granules arise de novo in the cell, and are not handed 

 down by division and subsequent growth. It so happens, how- 

 ever, that if there are but a few in a cell they are so oriented that 

 some will get into each daughter cell, for no healthy vegetative 

 cell is without them. 



I have at times seen appearances which indicate that the a 

 granules divide, but the evidence is not sufficient and conclusive. 

 There often appears to be a connection between two granules, 

 but I presume that this is simply the deeply stained protoplasm 

 the color of which is not washed out. The new granules usually 

 arise in close proximity to the older ones. One may often see a 

 series of several sizes together ranging quite gradually from larg- 

 est to smallest. Cylindrospermum is an excellent genus in which 

 to study them in the vegetative cells, and Anabaena variabilis 

 furnishes a good example of their disappearance during spore 

 formation, for here the spores mature gradually in a series. 



Many of the granules appear hollow, or at least only the outer 

 rim stains. These are usually the larger ones, the smaller ones 

 appearing to be solid. The stain which I have so far found to 

 be unerring in differentiating them is a saturated aqueous solu- 

 tion of Bismarck brown used on living material. This makes it 

 exceedingly quick and simple to examine a species to detect if 

 they are present or absent. Living material mounted on the slide 

 as previously described may be stained, dehydrated and mounted 

 in balsam within a short time, varying from a few minutes to 

 half an hour, according to the species, but there is scarcely ever 

 any danger of over-staining. The granules of material killed in 

 potassium iodide-iodine or in alcohol stain equally as well with 

 Bismarck brown as the living material. In some species the gran- 

 ules may become swollen when stained. Besides Bismarck brown, 

 there are other stains which will stain equally well and differen- 

 tiate these granules in some species. Methylene blue in aqueous 



