4 M. G. Thuret on the Reproduction of certain Nostochinese. 



and to acquire a transparent membrane (fig. 2). At the same 

 time the green globules increase in magnitude ; but this time 

 their increase is in diameter, and no longer in the direction of 

 the length of the chaplet. They thus become discoid : finally, 

 they divide into two by a division which takes place in a direc- 

 tion opposite to that described above (fig. 3). Most of the glo- 

 bules divide once or twice in this way, and then the chaplet, 

 considerably enlarged, has entirely changed its aspect. It is 

 transformed into a transparent sac, of variable length, in which 

 the doubled granules are arranged in parallel superposed ranks, 

 often very distinctly and pretty regularly (fig. 4). But this 

 regularity soon vanishes; the rows become joined together 

 alternately; that is, the globule placed at the border of one 

 rank becomes adherent to the globule placed above it, and the 

 globule opposite, to that below it In this way a new chaplet 

 is formed, wound upon itself, in the interior of the sac. At 

 the first instant it is difficult to distinguish this arrangement 

 of the globules. Crowded in the narrow sac which the mem- 

 brane forms, and attached together more or less obliquely, they 

 often exhibit only a confused heap. But their enchainment 

 becomes more and more evident in proportion as the young 

 Nostoc grows ; the sac dilates, the new chaplet becomes elon- 

 gated, and its convolutions separate, and become more distinct 

 (fig. 5). For some time longer the young frond still presents 

 enlargements which correspond to the situation occupied by 

 the rows of globules : by degrees these traces are effaced, and 

 the frond becomes developed into a transparent rounded mass, 

 in the interior of which the chaplet turns and twists in all 

 directions. At this epoch we may already distinguish some 

 heterocysts among the globules. The figures comprised under 

 the numbers 2, 3, 4 and 5 will, I think, give a sufficient idea of 

 the different aspects presented in the transformation of a chaplet 

 of Nostoc into a new individual. This phenomenon presents 

 numerous variations of the details, upon which I consider it 

 useless to dwell here. I shall confine myself to the following 

 observations. 



In general the terminal globules of the chaplet do not undergo 

 the same modifications as the rest. They lose their colour like 

 the heterocysts, and remain attached to the extremities of the 

 chaplet, without taking part in the development I have just de- 

 scribed. Sometimes we find them a long time after, still ad- 

 herent to the surface of the young frond. 



Frequently, also, one of the intermediate globules becomes 

 transformed into a heterocyst. Sometimes even two or three 

 are produced at different intervals in the length of the chaplet 

 (fig. 4). The latter thus becomes divided into two or more 



