PHYTOPLANKTON OF BRANSFIELD STRAIT iii 



were recorded, as against two on the more southerly stations, albeit in very small num- 

 bers, a fact which lends support to the view that the remnants of a formerly more varied 

 phytoplankton persisted longest on the northern side of the strait. The three stations 

 in the middle of this line furnished by far the richest catches of the whole survey, with 

 C. valdiviae again predominating, the numbers at the richest of them (St. WS 391) being 

 estimated at over 37 millions. The surface salinity and temperatures at the four northerly 

 stations on this line are rather higher than those commonly met with to the north-west 

 of the Bellingshausen Sea ; yet there is little doubt that this surface water originated 

 there, the apparent anomaly being explicable on the grounds of vertical mixing as 

 the water enters the Bransfield Strait over the comparatively shoal stretches between 

 the islands across its western end. At the most southerly station on this line the 

 marked fall in temperature and higher salinity again indicated the influence of the 

 Weddell Sea eddy ; but, as on the eastern line, this was apparently without effect upon 

 the nature of the phytoplankton, C. valdiviae still being entirely dominant in a more 

 moderate haul. Indeed, it would almost appear that the peculiar type of old Bellings- 

 hausen Sea surface water filling the greater part of the strait was here diluted, as it were, 

 by admixture with old Weddell Sea water, itself almost devoid of plankton. 



The full analyses of the phytoplankton material collected on this line will be found 

 in Table XXXIV, from which it will be seen that of the rarer forms Rhizosolenia 

 graciUima was again most frequent on the northern side of the strait. 



This survey was concluded with a series of six stations between the islands at the 

 western end of the strait. Actually the majority were worked in south to north order, 

 but the order adopted in Table XXXV appears to facilitate the consideration of the 

 survey as a whole. The positions are shown in Fig. 49, from which it will be seen that 

 the first of these stations, St. WS 394, was worked just to the north of Deception 

 Island, between it and Livingston Island. Here Corethron valdiviae and Thalassiosira 

 autarctica predominated in a very poor catch. 



At St. WS 399, worked in the comparatively deep channel between Smith and Snow 

 Islands in the extreme north-west of the Bransfield Strait, a moderate haul was ob- 

 tained; in it Corethron valdiviae was again the most numerous form, but there was 

 also a larger variety of species than at any other station worked during this survey. 

 Chief among these were Nitzschia seriata and Rhizosolenia graciUima, with Chaetoceros 

 atlanticus, Ch. dichaeta, Ch. neglectus and Ch. schimperianus , forms which were rarely 

 or never encountered elsewhere in the area at this time. Farther south, between Smith 

 and Low Islands at St. WS 398, the poorest catch of the whole survey in point of total 

 number was obtained. Here there was still a fair variety of species, but Corethron was 

 again strongly dominant, the other forms occurring only in very small numbers. 



At each of the three more southerly stations, bridging the gap across the western end 

 of the strait from Low Island to the northern end of Brabant Island, a closely similar 

 phytoplankton was encountered, with C. valdiviae very strongly dominant. The size of 

 these hauls decreased steadily towards the southward, indicating that the main influx 

 of Bellingshausen Sea water into the Bransfield Strait takes place at the north-western 



