NOTES ON SPECIES 159 



he really saw was C. valdiviae. However, he decides to retain Karsten's name for the 

 species, on account of his being the first to describe it in accurate detail, and this seems 

 to be the best course in view of the frequent references to C. valdiviae in subsequent 

 literature. 



The size variation within this species is enormous ; in spring it has frequently been 

 seen to be reproducing by both microspore and auxospore formation in the same 

 samples. The spineless chain form encountered so abundantly round South Georgia 

 during the abnormally warm season 1929-30, in sub-Antarctic surface water, and rarely 

 farther south, should, we think, be regarded as a vegetative phase of this species, as 

 intermediates with the terminal frustules bearing the characteristic whorl of bristles, 

 have been observed in Bransfield Strait. Possibly this form of growth is induced by a 

 temporary shortage of silica. 



This is one of the most important of Antarctic plankton diatoms, being almost uni- 

 versally present, and only surpassed in numbers by the vast swarms of small Chaeto- 

 cerids, Thalassiosira antarctica and Nitzschia seriata, met with in old water of the 

 western Weddell Sea type, and, less noticeably, in old Bellingshausen Sea water ; and by 

 the large masses of Chaetoceros criophiliim and Rhizosolenia styliformis present i n the eastern 

 Weddell Sea. All these have a much more localized distribution in time and space. 

 Taking the year as a whole, Corethron valdiviae formed over 90 per cent of the phyto- 

 plankton in Bransfield Strait. It was the dominant form in the Bellingshausen Sea, 

 except in the extreme west, and also at several stations in water of the eastern Weddell 

 Sea type. In common with most of the other species it appears to reach its maximum in 

 late spring, but maintains a high level of abundance throughout the greater part of the 

 season. 



Family BACTERIASTRACEAE, Lebour, 1930 



Genus Bacteriastrum, Shadbolt, 1854 



Bacteriastrum varians, Lauder. 

 Pavillard, 1924, p. 1086. 



In his revision of this genus based mainly upon material collected in the Mediter- 

 ranean and North Atlantic, Pavillard has shown that the forms there met with should 

 not be referred to this species. He goes on to say that it is characteristic of the Indian 

 Ocean and eastern tropical waters generally, and probably does not occur in the 

 Atlantic at all. While this is almost certainly correct so far as the North Atlantic is con- 

 cerned, we believe it to be present in the South Atlantic. Forms agreeing with the 

 revised description of this species were met with in small numbers at three stations in 

 sub-tropical and tropical surface waters, between 31 and 21 1" S in long. 30° W. It was 

 abundantly present in waters of the Agulhas stream, which is, of course, of Indian 

 Ocean origin, and was also observed to the west of Saldanha Bay in the Benguela cur- 

 rent. It would therefore seem quite possible for this species to spread into the tropical 

 and sub-tropical South Atlantic via the Agulhas, the Benguela, and westwards in the 

 south equatorial current. 



