512 Annals of the South African Museum. 



Reproduction takes place by breaking up of the large colonies into 

 four-celled groups or by the liberation of single cells (Fig. 8, e). 



The cell-contents were very poorly preserved. I have been unable 

 to recognise a pyrenoid, nor am I able to state the nature of the 

 chloroplast ; the latter apparently readily fragments into two or more 

 parts (Fig. tf, a and g). 



As regards shape of cells, this species comes very near to S. bijugatus 

 (Turp.), Kuetz., but it differs from all known species of the genus in 

 the cohesion of successive generations of cells to form giant colonies. 

 Moreover, the method of division is different. 



Selk (Beitr. z. Ivenntn. d. Algenfl. d. Elbe u. ihres Gebietes, 

 Jahrb. Hamburg. Wiss. Anst., xxv, 1907, 3 Beiheft, p. 6 et seq.) 

 appears to have observed a similar form bearing the characteristics 

 of 8. hystrix, Lagerh., although it is not quite clear from his descrip- 

 tion whether the colonies are produced in exactly the same way. He 

 describes it as a forma cohaerens of S. hystrix, regarding it merely as 

 a growth-form. 



Genus SELENASTRUM Reinsch. 



1. Selenastrum gracile, Reinsch, Abhandl. Naturhist. Ges. Nuern- 

 berg, iii, 1866, p. 65, Tab. IV, tig. 3, a-b. 

 Samples 14, 15. 



Genus COELASTRUM Naegeli. 



1. Coelastrum sphaericum, Naegeli, Gatt. einzell. Algen, 1841', p. 98, 

 Tab. V, C, tig. 1. 



Sample 39 (rare). 



Recorded from the Karroo by G. S. West. 



2. Coelastrum microporum, Naegeli, in A. Braun, Alg. unicell., 

 etc., 1855, p. 70 ; Senn.Bot. Zeit., lvii, L899, p. 53, PI. II, tigs. 11-17. 



Sample 57 (very rare). 



3. Coelastrum morus, W. & G. S. West, Algae from Central 

 Africa. Journ. Bot., xxxiv. 1896, ]». 381, Tab. 361, tig. 4. 



Forma capeusis, uov. form. (Fig. 9 ) 



Cellulis varie dispositis, verrucis paucioribus et inter se remotis. 

 Diam. cell., 10-11 /x ; altit, verrucae, 2-25 /v.. 



Samples 14 (rare), 15 (rather common, epiphytic, and even endo- 

 phytic in the stratum of Phormidium valderianum ). 



There seems little doubt that this form was actively multiplying 

 within the stratum of the Phormidium in sample 15, although some 



