6 INTRODUCTION, 



which are permanently smaller. The case is the same with 

 those genera in which the separation of the cells is complete. 

 I admit that the division of the cells just described apparently 

 differs greatly from that in other simple Algse ; but I believe 

 that the process in all is essentially the same, and that what- 

 ever differences exist are modifications necessarily resulting 

 from the different forms of the cells. In the examples already 

 given the cell itself consists of two distinct portions, having 

 a constriction between them ; hence each of the new-formed 

 portions is similarly distinct from the older one which forms 

 it and to which it is united. 



In order fully to elucidate the subject, cells may be distri- 

 buted into three principal kinds, distinguished by their form : 



1st. Bipartite cells, already described, and more or less 

 constricted at the middle ; 



2nd. Cells globose or rounded at the ends, or having the 

 extremities attenuated ; 



3rd. Cylindrical cells. 



Bipartite cells belong only to the Desraidieee ; cells globose 

 or roundish at the ends are seen in the Nostocs and PalmeUeee ; 

 attenuated cells in the Desmidiese ; and cylindrical ones in 

 the Conjugatse, Tiresias, &c. 



It is obvious that the new portions must arise from the 

 whole of the junction margin of the original valves ; conse- 

 quently when the junction occupies only a part of the breadth 

 the new portion will be narrower than the old ; but when the 

 junction of the valves is as broad as the cell, the new portion 

 will from the beginning be of the same breadth. From this 

 important fact, w^e may explain the different sorts of division. 

 Since in the two latter kinds of cell the valves are united by 

 their entire breadth, the new portions cannot be distinguished 

 by their size, we must therefore have recourse to other aids 

 to enable us to trace the changes and satisfy ourselves of 

 their real identity with that already described ; and I hope to 

 be able to show that this identity does exist. 



In Nostoc and Anabaina the cells are globular, and as there 

 is no constriction we might remain ignorant of the real me- 



