298 Phillips on a Comparative Study of the 



71), which divides the cell into two, always, however, leav- 

 ing a small pore at or near the centre through which the 

 protoplasm of one cell is connected with that of the next. 



These observations, therefore, seem to show that two 

 methods of division occur in the Cyanophycese. In one of 

 these the nucleus does not pass beyond the net or spireme 

 stage, which constricts itself into two nearly, though not 

 necessarily equal parts. In the other method we have a 

 very evident primitive state of karyokinesis, but no longi- 

 tudinal splitting of the chromosomes, such as has been held 

 to be necessary to equally distribute the hereditary materials 

 in the higher plant cells. Whether there is any sequence 

 to the occurrence of these different methods of division, so 

 that one form might appear at one time in the life-history 

 of the plant and the other at another cannot yet be asserted. 

 There is no reason why such should not occur, however, 

 for the Cyanophyceae are evidently of a much more com- 

 plex organization than we have hitherto supposed. But the 

 fact that the two forms of division appear, and that each 

 starts upon the way to karyokinetic division, one stopping 

 in the spireme stage while the other goes on to the forma- 

 tion of a rudimentary spindle, is significant. The fact that 

 the division of chromatin is not always equal and never 

 longitudinal or qualitative, seems to suggest that this pro- 

 cess is not so essential to the division of the hereditary 

 material as has been held in the past, but that it has rather 

 been evolved as a convenient method of giving out to each 

 daughter cell the quantitative, rather than the qualitative 

 amount of chromatin which belongs to it. This is further 

 suggested by the method of division, which does not pass 

 beyond the net-spireme stage. It is not in any sense a 

 fragmentation of the nucleus, since it is invariably followed 

 by the formation of a division wall. This is laid down in 

 the same way as in the cells which show typical karyokine- 

 sis. It seems, therefore, that division in these forms is evolv- 



