18 discovery reports 



ment is towards the south-west, but the observations in this part of the ocean show 

 that the boundary between the easterly and westerly currents is not so well defined as 

 it is in the Indian Ocean: the easterly current is well marked north of 60-65 S and 

 the westerly current near the continent, but the intermediate region, although proved 

 by the upwelling of the deep water to be a divergence region, appears also to be one of 

 irregular movements. 



The observations in section 6, Plates VIII and IX, and in a section drawn by Mosby 

 from the 'Meteor' and 'Norvegia' data (Mosby, 1934, Figs. 17-19), between io° W 

 and io° E point to the existence of a divergence region in 61-66 S. In 15-20 W the 

 observations made at Sts. WS 552-5 (Station List, 1932) indicate a sharper divergence 

 between 64 and 68° S. The data in section 5, Plates VII and IX, show that in 22-23 ° W 

 there is an ill-defined divergence region between 64 and 70 S. The longitudinal section 

 made by Brennecke through the Weddell Sea (1921, pi. 4-9) shows that there is a di- 

 vergence region in 62-70 S ; the observations are, however, so far apart that they 

 probably miss such irregularities in the temperature and salinity distribution as are 

 shown in section 5. 



Along the western shores of the Weddell Sea the westward-flowing water turns north- 

 ward ; a good indication of the path that it takes is given by the drifts of the ' Endurance ' 

 and the ' Deutschland '. The ' Endurance ' drifted first to the west between 76 and 77° S, 

 then north-west as far as 74 S, and farther north the drift of the ship and later of the 

 ice-camp was north-north-west or north. The 'Deutschland', beset 100-150 miles 

 farther from the coast, drifted north-west as far as 72 S, northwards between 72 and 

 66° S, and then north-east. 



The observations made during the drift of the ' Deutschland ' show that the move- 

 ment of the ice was caused only by the prevailing winds (Brennecke, 1921, p. 200), and 

 they suggest that the bending of the westerly current towards the north off the east 

 coast of Graham Land is entirely due to the existence of a prevailing northerly wind. 

 A section was made across the northerly current by Nordenskjold (1917, pi. 2). Within 

 200 miles of the coast the isotherms slope steeply towards the west, pointing to the 

 existence of a strong movement towards the north ; farther off-shore they are almost 

 horizontal and indicate a lesser movement. As it flows northwards into the region of the 

 westerly winds the current is turned towards the east ; only a very small current turns 

 round the northern end of Trinity Peninsula towards the west. 



Summarizing, it will be noted that there are as yet few series of observations that can 

 be used to fix the boundary between the easterly and westerly currents. In the neigh- 

 bourhood of the boundary the surface currents will no doubt be irregular, and there is 

 evidence of particularly wide zones of irregular movements north of the Weddell and 

 Ross Seas. With the help of a large number of observations it would be possible to 

 determine an average position, which itself would probably vary from winter to summer, 

 as suggested by the observations made during the drift of the 'Belgica'. 



The approximate position of the boundary as far as it is known at present is shown in 

 Fig. 4. In the Atlantic and Indian Oceans the boundary is found in about 65 S, but 



