306 Phillips on a Comparative Study of the 
frequently, the case, then the heterocyst cell will develop 
into a spore also. The development of this spore into a 
trichome will depend upon the amount of material passed 
into it from the other cells. If it has formed in the middle 
of a trichome, and can be fed as the spore cell is, it will 
develop into a spore of as much strength as the other cell, or 
if it has only a few cells to feed it, it will be smaller and 
weaker. This would seem to argue that in Cylindrosper- 
mum, at least, the heterocyst is not a dead cell as it is usu- 
ally supposed to be. All stages in its development, as here 
described, have been seen and it is not a difficult task to 
reproduce the condition. Not infrequently the terminal 
heterocyst, in material grown on a soil saturated with a 
full culture solution, will divide once before its chromatin 
becomes diffused, thus forming a double heterocyst (Fig. 
56). After the formation of the spore, and the deposition 
of the plug which closes the pores at the ends of the spore 
cell, the chromatin which has been passing along the trich- 
omes may not cease to flow, but become heaped up within 
the cell next to the spore, which then begins to grow and 
becomes a spore (Fig. 64) like the first, and with all of the 
steps taken by it except the preliminary division which cuts 
off the heterocyst. 
In Spermosira the spores may form in any part of the 
trichome. One spore usually forms first in the manner 
already described for Cylindrospermum, except that the 
heterocyst cell is not cut off as in that organism. The 
chromatin here does not take the form of vesicles, but is in 
the form of more or less angular grains. After the forma- 
tion of the first spore, others commonly form on either side 
of it, until a long string (Fig. 64) has been developed. 
In Wollea saccata, the spore cell receives the material that 
is passed along the trichome just as in the other forms. It 
thus grows to twice or even three times the thickness of 
the vegetative cell and elongates until it is six to eight times 
the length (Figs. 84 to 86). The chromatin first becomes 
