323 Mr. J. Ralfs on the Nostochinesp. 



or cells are commonly more or less orbicular, but sometimes dis- 

 ciform or cylindrical, and in a few instances confluent. At first 

 they are uniform ; they divide in the same way as they do in 

 other simple Algse, and during division are geminate. 



I believe that the filaments multiply in consequence of the 

 division of the joints being here and there complete in the same 

 way as in Oscillatoria ; because we find that the filaments in 

 an early state (as Mr. H assail has correctly noticed) are but spa- 

 ringly scattered through the mass, whereas in mature specimens 

 they become very numerous. Mr. Hassall considers that this 

 increase results from the separation or dislocation of the enlarged 

 cells. I believe, on the contrary, that when the enlarged cells 

 are fully formed the plant approaches maturity, and the filaments 

 cease to multiply. The following facts will, I think, prove that 

 the multiplication of the cells has no essential connection with 

 the enlarged joints. In Cylindrospermwn they terminate the 

 filament, yet here the filaments are as numerous as in the other 

 genera, nor do we find them of more unequal lengths, which 

 would however necessarily be the case if the enlarged terminal 

 joints separated to form new filaments ; and further, the enlarged 

 joints themselves are not developed until after the filaments have 

 become numerous. 



As I have stated above, the joints or cells are at first uniform 

 in size and figure; but in the mature plant they are of three 

 different kinds, all of which are generally, though not I believe 

 invariably, present in the same filament. These are — 



1st. Ordinary joints or cells. — Of these 1 shall only observe, 

 that they divide by separating into two hemispheres, each gra- 

 dually acquiring a new portion in the room of that from which 

 it has separated, and that until this process is complete they ap- 

 pear geminate. As this geminate state is common to every spe- 

 cies during growth, it cannot be employed as a specific character, 

 and as the other cells do not divide, the elongation of the fila- 

 ment must be solely due to the ordinary ones. Their endo- 

 chromes are slightly granular, which gives them a dotted ap- 

 pearance. 



2nd. Vesicular joints or cells'^. — These make their appearance 

 at an early period. They are generally orbicular, but sometimes 

 elliptic, and are usually larger than the ordinary cells. They are 

 evidently analogous to the vesicles present in Rivularia and in 

 some of the Oscillatoriecs. They do not contain granular matter, 

 but are filled with a bluish fluid, which eventually acquires a brown 

 tint. When mature they are frequently furnished with diverging 

 hairs or cilia. They seem to be connected with the adjacent 



* By Mr. Thwaites they are named, frc^n their usual position, " connect- 

 ing cells;" they have also been called " Heterocysts." 



