The Freshwater Algae of Africa. 527 



bluish-purple of the chloroplasts (Fig. 15, A and Z?) ; the nuclei are 

 irregularly scattered among the chloroplasts. 



The longitudinal walls are in genera] very thin, recalling those of a 

 Vaucheria. In some cases they exhibited a faint longitudinal striation, 

 but this is very probably due to contraction. The septa are of a verv 

 peculiar character. In typical cases they appear to arise primarily by 

 the apposition of a complete ring of thickening on the inner side of the 

 longitudinal Avail (cf. Fig. 15, G). From this primary ring a con- 

 siderable number of thick finger-like processes arise (Fig. 15, G) and 

 gradually grow inwards towards the centre (Fig. 15, B). Whether 

 a complete constriction of the protoplast is thereby effected must 

 remain doubtful. In many cases, at least, there is an obvious 

 central aperture, through which the protoplast, though much narrowed 

 down, extends uninterruptedly (Fig. 15, if). It was not uncommon to 

 find that the protoplast had contracted considerably away from the 

 walls, both longitudinal and transverse, but in such cases a mass of 

 protoplasm often remained clinging on one or both sides of the septum, 

 a fact which also speaks for continuity of the protoplasm. I have 

 also been unable to settle whether the finder-like processes fuse 

 laterally or whether narrow interspaces persist between them ; there 

 is much to suggest that the latter is often the case, hut the point 

 could only be settled finally by investigation of fresh material. In 

 some cases the septa appear as in Fig. 15, E, and Pig. 16, H ; here it 

 looks as though a secondary thickening membrane had been deposited, 

 on either side of the septum, and possibly in these cases a true septum 

 is ultimately constituted. 



Apart from the normal " septa " just described, numerous irregu- 

 larities were observed. Very often the processes do not grow in a 

 regular radial manner towards the centre, but overlap and become 

 entangled with one another in various ways (Fig. 16, E, G, and H). 

 A not uncommon condition is that in which one or more of the pro- 

 cesses, either only on one or on both sides of the septum, after 

 growing hvwards in a radial direction, turns out at right angles, 

 leading to the development of a central hump, on one or both sides 

 (Fig. 16, E). Again, frequent instances were encountered, in which 

 successive processes arose near together from the longitudinal walls, 

 without forming a proper septum (Fig. 15, F). In extreme cases 

 (Fig. 16, /) this leads to the formation of very short "cells," often 

 only a very small fraction of the normal length. Short knob-like 

 ingrowths, of a similar nature to the processes, are to be encountered 

 at many points in the course of the " cells " (cf. Fig. 15, F). 



The processes themselves appear to be homogeneous, although 



