254 



DISCOVERY REPORTS 



Table 20. Sample C of discoloured water near Sandwich Harbour. Collected immediately after Sample B 

 {Table lg), but in a streak of noticeably more turbid water 'Khaki-coffee coloured'. 0-25 ml. of well- 

 mixed sample examined by drop method and counts summed. Factor to express results as approx. nos. 



cells /ml. x 4 



Estimated 



Area in optical section of the organisms in 1 ml. in Sedgwick- Rafter counting-chamber: about 5160 W.I.S.U., or 2 + 

 parts per thousand. Volume about 12-5 parts per million. 



These tables allow comparisons, not only of the constituents, but also of the relative density of the 

 microplankton in each sample in terms of numbers, area in optical section, and volumes, per millilitre. 

 It is seen that sample C (Table 20) was by any standard nearly twice as dense as B. Sample A was 

 much thinner, but the ratio varied little in terms of numbers, area or volume. 



As to the constitution of the samples it appears that the small dinoflagellates Peridinium triquetrum 

 and Prorocentrum micans were the main constituents of the thinnest of these samples (Table 18). 

 with quite a high proportion of Asterionella and other inshore diatoms. The two richer samples 

 (Tables 19 and 20) show a definite 'bloom' of Peridinium triquetrum, the other constituents providing 

 less than 12% of the total number of organisms estimated. 



Peridinium triquetrum, a neritic species of very wide distribution, has been recorded 'blooming' in 

 association with a fish mortality on at least one previous occasion, at Rostock in the Baltic on 

 10 October 1917 (Lindemann, 1924, who called it by its old name Heterocapsa triquetra). Linde- 

 mann's plankton observations were only made after a considerable search for other possible causes 

 of the catastrophe had proved fruitless. The Peridinium triquetrum was co-dominant with Krypto- 

 peridinium foliaceum and other microplanktonic organisms were present only in quite insignificant 

 proportions. 



Our Benguela current findings, added to Lindemann's observations, might seem to suggest 

 Peridinium triquetrum as a probable ' cause ' of the Walvis Bay mortalities, but it has to be remembered 

 that in both instances the mortality had begun before the observations on the bloom were made. The 

 bloom may or may not have been present when the fish were killed. Braarud (1945) has shown that 

 very high production of P. triquetrum (and sometimes of Prorocentrum also) may occur in heavily 



