508 Annals of the South African Museum. 



to be three-lobed; basal lobe massive, rounded, apical lobes usually 

 widely divergent, separated by a wide and deep cleft somewhat 

 V-shaped, but with rounded base. Typically this cleft is about one- 

 third the total length of the cell in depth, more rarely it may be 

 quite shallow (fig. 7, C 3). The apices of the lobes are more or 

 less narrowly rounded, sometimes slightly recurved. The parietal 

 chloroplast is extensive, slightly lobed, lining most of the cell 

 except a small part beneath the cleft ; in this the nucleus lies (i.e. 

 approximately in the centre of the cell). The single large pyrenoicl 

 lies in the massive basal part of the chloroplast. The cells are rich 

 in starch, and contain, in addition, refrangent globules, probably 

 volutin, generally most conspicuous in the apical lobes. The wall 

 is delicate, closely adpressed to the protoplast, only becoming 

 apparent on staining, e.g., with iodine or methylene blue. With the 

 latter the gelatinous matrix stains faintly bluish purple, the protoplast 

 deep blue. 



The cells divide to form 4- or 8-celled groups (fig. 7, D) ; the empty 

 wall of the parent cell is often visible near such groups of young cells. 

 The three-lobed structure is apparent at a very early stage (fig. 7, D), 

 but sometimes the young cells appear almost straight (fig. 7, E). 

 As the cells thus formed grow they diverge from one another, and all 

 trace of grouping is soon lost. 



This species appears to be nearest K. Malmeana (Bohlin) Wille, 

 (cf. Brunnthaler, 1915, p. 181, fig. 266), but the cells are typically much 

 more deeply incised and more distinctly three-lobed, with rounded 

 apices, than in that species, although some cells occur in which the 

 anterior end is only slightly concave (fig. 7, C, 3). The mature cells 

 here are larger than in most species and the colonies very much larger 

 than anything as yet recorded for the genus ; the number of com- 

 ponent cells, too, is exceptionally great. 



Large cells in which the nucleus or pyrenoid had already divided, 

 presumably preparatory to cell division, were occasionally seen 

 (fig- 7, F). 



On and inside many of the colonies biciliate motile cells were often 

 found, but, as these are identical with the zoospores of Sphaerocystis 

 Poweri, it is almost certain that they belong to that species, and 

 have penetrated the matrix of the Kirchneriella colonies. In the 

 same way they may enter older Volvox colonies of either of the 

 associated species (cf. Plate XXX, B). 



