HICKSj ON THE DIAMORPHOSIS OF LYNGBYA, ETC, 159 



growth. The diametei' of the Lyngbya varies much even after 

 it has assumed the mature appearance. 



The formation of free gonidia from it, next claims interest. 



If we observe a batch of Lyngbya in its first appearance in 

 spring, and also at other times, we shall find many of the 

 threads throwing off from one extremity its terminal cells, 

 Avhich, relieved from pressure, become globular. Watching 

 these, and carefully tracing their history by keeping them 

 under continuous observation, I find that they undergo 

 segmentation in the same manner as the so-called Palmella- 

 cese. I have represented the principal varieties of their sub- 

 divisions at fig. 4. It will be seen that this process assumes 

 the same type as prevails in the gonidia of lichens, proceed- 

 ing in various manners till, in some instances, the subdivisions 

 are very minute. By means of this gonidial increase, con- 

 siderable surfaces are covered with a palmeloid growth, and 

 it has constituted one of the forms included under the term 

 Protococcus viridis ; and thus gives another example of the 

 temporary nature of that order. 



The mode by which these small cells assume the linear 

 form of segmentation is as follows : — Generally after proceed- 

 ing to their last stage, they assume an oval shape ; a septum 

 appears in the centre, and the divisions elongating form the 

 first appearance of a thread. These two cells segment simi- 

 larly in their centres, and thus a thread of many cells is pro- 

 duced (fig. 5). The contents are at first homogeneous, but at 

 a very early stage they may become fasciated (as at fig. 6), 

 or even with an apparent central nucleus (fig. 7). Sometimes 

 the linear segmentation assumes a form very like the threads 

 of a Nostoc (fig. 8). The length of the cells is very variable, 

 depending upon the rapidity of the process of linear sub- 

 division, compared with the rapidity of individual cell-growth. 

 Sometimes the rate of the former is so much in excess that 

 the cells are no thicker than the septa ; the thread appearing . 

 to consist of alternate narrow green and colourless bands ; 

 again, sometimes the cells are three or four times longer than 

 broad. All these various stages can be viewed at one time in 

 different threads in such a gentle gradation as to point un- 

 mistakeably to their common origin ; but, perhaps, the most 

 powerful confirmation is, that they may occur simultaneously 

 in the same thread. 



The method above described is probably that by which all 

 the algce revert from the segmenting gonidial to the linear 

 forms. 



Let us now recur to the mature thread of Lyngbya mu- 

 7'aJis. In certain circumstances there is a strong tendency 



