518 Annals of the South African Museum. 



circumference (Fig. 11, h). Sooner or later, however, this chloroplast 

 becomes drawn out at either end into a number of processes, which 

 extend the entire length of the cell, so that at this stage the chloro- 

 plast fills the whole of the latter (Fig. 11, a, c-g) ; it will be plain 

 from the figures that this development of the chloroplast does not 

 go hand in hand with the gelatinisation of the longitudinal walls 

 (cf. especially Fig. 11, g). The number of processes of the chloro- 

 plast seems to vary between four and six, and they often run to a 

 more or less marked point (cf. Fig. 11, c). At this stage the chloro- 

 plast is very similar to that of many Chaetophoraceae, especially 

 that of Draparnaldia, in which, however, the processes are generally 

 more numerous. Even in the older stages the main body of the 

 chloroplast retains its girdle-form (Fig. 11, a, c). In the young 

 stage the chloroplast appears to have but a single pyrenoid, but 

 later on there are certainly several small ones (Fig. 11,/). The 

 round nucleus is relatively large and occupies a central position 

 (Fig. 11, a, c,f). 



In the young filaments the cells are several (3-4) times longer than 

 broad (Fig. 11, b, e, g), but, as gelatinisation of the longitudinal walls 

 proceeds, the cells become shorter, their breadth at this stage often 

 exceeding their length (Fig. 11, a, c). The longitudinal walls may at 

 first be quite thin (Fig. 1], g), but thickening sets in sooner or later 

 (cf. Fig. 11, d and e). The original septa, produced before this 

 thickening occurs, are even in the mature filaments found extending 

 right across from one edge to the other, but the septa subsequently 

 formed do not do so (Fig. 11, c). In the fairly young filament there is 

 often distinct constriction between some of the cells (Fig. 11, b ande). 

 These points of constriction persist, even in the mature condition, 

 but there are no constrictions evident between the later-formed cells 

 (Fig. 11, a, c). The original points of constriction appear, in the older 

 filaments, to mark places of weakness, at which the filaments readily 

 split, and this may possibly represent one method of reproduction 

 of this species (Fig. 11, c). The thickening of the longitudinal 

 walls is very markedly mucilaginous, but the bounding cuticle always 

 remains distinct (cf. Fig. 11) ; there is no indication of stratification 

 in these walls at any stage. The transverse walls do not in general 

 become markedly mucilaginous, although slight gelatinisation is not 

 infrequent. 



As far as I am aware the type of chloroplast characteristic of the 

 mature filaments of this species has not hitherto been observed in the 

 genus Ulothrix, or for the matter of that in any member of the 

 Ulotrichaceae This, taken together with the very marked gelatinisation 



