THE LARVAL STAGES 99 



deeper than the lower Hmit of the vertical nets and that the eggs obtained [i.e. the few he found in his 

 samples] are the scattered product of dispersal of a much greater mass situated in still deeper water ? ' And 

 again, four pages later, '. . .the present observations on the vertical distribution of the Metanauplius 

 furnish very favourable evidence for the hypothesis that the eggs are to be found in deep water rather 

 than at the surface '. Rustad (1934) also supposed that in this species there would be a developmental 

 ascent, but from a very shallow depth, although he had no observations to show how it could actually 

 come about. 



Deep and shallow horizontal dispersal 

 In the south-western part of the Weddell Sea, as Brennecke (1921) has shown, the Antarctic continent 

 is bordered by a wide continental shelf where the depth is only a few hundred metres. Over this 

 shelf the water is cooled right through by convection in winter, and its salinity, already high, is 

 increased as fresh water is removed and salt left behind when sea-ice is formed. Owing to its great 

 density this cold highly saline water sinks from the shelf and, flowing down the continental slope, 

 makes its way northwards as a cold deep bottom current. The later work of Deacon (1937) seems to 

 show that the main, perhaps the only, source of this Antarctic bottom water as it has been called lies 

 in the southern part of the Weddell Sea where Brennecke originally found it, it being no longer 

 possible he says to assume as other authors have done that it ' is formed by the sinking of shelf water 

 all round the Antarctic continent, nor as Sverdrup (193 1, p. 102) supposes, that it is deflected to the 

 left on account of the earth's rotation as it flows northwards, turning towards the west. It is on the 

 contrary formed only in one region, the Weddell Sea, and its principal movement is towards the east '. 



Several reasons [Deacon writes] can be suggested to explain the very large formation of bottom water in the Weddell 

 Sea. The surface water in the southern part of the sea belongs to the current which flows towards the west along 

 the coast of the Antarctic Continent. Even in the Indian Ocean and the eastern part of the Atlantic Ocean this current 

 must have a high salinity in winter, and owing to the continued separation of sea-ice from it as it flows towards 

 the Weddell Sea its salinity must keep on increasing. Brennecke (1921) recorded surface salinities as high as 34-49%o. 

 The deep [i.e. the warm deep] water in the south and western parts of the Weddell Sea also has a lower temperature 

 and salinity than it has in the open sea in any other part of the Southern Ocean ; like the surface water it travels 

 along the continental slope from the western part of the Indian Ocean, and owing to the continued mixing with the 

 surface and bottom waters it forms a weaker barrier between the surface and bottom layers in the Weddell Sea than 

 it does anywhere else along the edge of the continent. 



The efltect of the earth's rotation on the current towards the west in the southern part of the sea and towards 

 the north along the east coast of Graham Land also tends to make the water sink in the coastal region. This effect is 

 also likely to be more powerful in the Weddell Sea than in any other Antarctic sector, because of the exceptionally 

 high latitude to which the sea penetrates and because the westward movement is not confined to the surface layer 

 as it principally is in the other sectors, but extended to the deep and bottom layers. 



The temperature distribution suggests that the principal movement of the bottom water from the Weddell Sea 

 is towards the east across the Atlantic Ocean ; but the low temperature of the bottom water farther north shows there 

 is also a strong northward movement. 



In his most recent account of the formation of the bottom water Deacon (1959), referring to the 

 work of Fofonoff (1956), has written: 



. . .he [Fofonoff] shows that water cooled to freezing point in the Antarctic must have a salinity of at least 34-5 i%o 

 before it can form mixtures with the warm deep water that are heavier than either type of water alone. It follows 

 that as soon as freezing water on the continental shelf round Antarctica reaches such a high salinity it can form 

 mixtures that sink down the continental slope to feed the bottom current. If it attains a salinity as high as 34-63 %„ it 

 is heavier than any mixture it can form with the deep water and must sink directly down the continental slope 

 unless it is confined in a depression on the shelf. 



Water with a sahnity less than 34-5 i%o, even if cooled to freezing point, cannot form mixtures with the warm 

 deep water that are heavy enough to sink below it. They must therefore float above it and feed the surface current. 



9-2 



