146 THE PHENOMENON OF 



has been formed, this must be repeated, but in the 

 reverse direction, by the cell at the opposite end, as well 

 as by certain others lying in the midst of the filament. 

 This first division into two unlike cells may be regarded 

 as the rudiment of an apical growth, which, however, is 

 arrested at the first stqp, since, after the production of a 

 single cell of the second order, the further division of the 

 cell of the first order of the second degree (the apical 

 cell), produces merely cells of the first order. Anuhainu 

 (to which may be added Sjjhmrozyga and Spermosira*) 

 and Cylindrospermum are distinguished from Nostoc by 

 the occurrence not merely of two, but of three modifi- 

 cations of cells, since the cells multiplying by continued 

 division (the cells of the first order) do not all behave 

 alike in the last generation, but partly expand into larger, 

 thick-coated seed-cells, doubtless destined to a long 

 period of rest, and partly remain small and thin-coated, 

 and then perhaps perform the function of a second kind 

 of reproductive cells, corresponding to the more gonidium- 

 like germ-cells of Nostoc, but, perhaps, at least in many 

 cases, die away as sterile cells, when the filament breaks 

 up. In Anahaina several of the intermediate cells of 

 each section are developed into the larger seed-cells ; 

 in Cylindrospennmn the lateral cells immediately bound- 

 ing the termmal or interstitial cells. 



The filaments of the Nostochinese exhibit, otherwise, no 

 perceptible contrast at the two extremities, and their in- 

 tertwined and irregular position admits of no distinction 

 into an upper and lower end.f In the group of the 

 Scytonemese and Rivulariese (excluding the Masticho- 

 trichese) a contrast of this kind is distinctly present, but 

 still without any root-like formation of the lower end, 



* These three genera are so loosely defined by Kutzing, that we cannot 

 form any clear idea of his distinctions. Anabaina seems to me to be only 

 sterile Spherozt/ga, and Spermosira to differ merely by shorter and more dis- 

 coid cells, and Nodularia may be a sterile condition of the last form. 



f I found indications of such a contrast in Cylindrospermum. Immicola, Kg., 

 and nther species, in which one end decidedly comes to its full development 

 later than the other. 



