HYDROLOGY OF THE BRANSFIELD STRAIT 49 



and minus temperature maxima, or none at all. Thus both the November 1929 and 

 December 1930 observations show that the intermediate layer of warm deep water 

 occurs at the south-west end of the strait in greatest amount ; in the former year in 

 November there is evidence for its presence over a large area between King George 

 Island and Trinity Peninsula, whereas in December 1930 its existence over this area 

 is not shown by well-marked temperature maxima. This is one of the difficulties 

 encountered in combining material from two different years from an area where there are 

 large variations from year to year in the physical characteristics of the water masses. It is 

 possible that the warm intermediate layer present in the Bransfield Strait comes from both 

 the west and the east, but the flow from the western end is undoubtedly much the greater. 

 Owing to the irregularity and steepness of the bottom contour there are not many 

 observations below 600 m. to show the movement of the bottom water in the strait. As 

 previously mentioned the movement of this water must be restricted by the submarine 

 ridges. It might be supposed that the bottom water in the deepest parts of the deep 

 basins would be stagnant. This, however, is far from the actual case, as the oxygen 

 content at great depths shows. Thus at St. 543, approximately 15 miles from Cape 

 Melville, the oxygen content increases from a depth of 600 m. to the lowest observation 

 at 1500 m., where it is 6-43 c.c. /litre, equal to a saturation of 76-3 per cent. This 

 indicates that the bottom water must have been lately renewed, and movement of the 

 bottom water must take place whenever the older water is replaced by a fresh supply of 

 highly aerated water. In considering the movement of the bottom water within the 

 strait, its mode of formation must be sought. H. U. Sverdrup, in his report on the waters 

 of the North Siberian Shelf (1929), noted that instabilities were found in the surface 

 layer and that the accuracy of the salinity and density determinations was such that 

 these instabilities could not be otherwise than real. Moreover, the instabilities were 

 only observed in the cold season when freezing takes place, and never in summer. A 

 series of tables and graphs was given by the same author which show that the formation 

 of these instabilities is clearly connected with the freezing processes. He also states 

 (1929, p. 67) "The salinity of the water increases when ice forms, but the heavy brine 

 does not sink immediately to the level of equilibrium, the sinking may be very slow and 

 irregular and instabilities are so frequent that they are found in more than one-third of 

 all cases. It is characteristic that the greatest numbers of instabilities are found in the 

 earliest part of the winter when ice is freezing over many open lanes, while the greatest 

 numerical values of the instabilities are found in the middle of winter, when the 

 number of openings is smaller but freezing is more rapid ". It is probable that a simdar 

 state of affairs occurs over the continental shelf of western Graham Land and Trinity 

 Peninsula, and indeed over the shallower portions of the strait, in winter. In 1929, even 

 in November, the surface values of temperature, salinity and density (a,) at St. WS 482 

 were - 1-18° C., 34-51 °/ 00 and 2778, and in December 1930 at St. 547 the corre- 

 sponding values were - 1-02° C, 34-53 7oo and 2 779- These two stations were 

 situated on the continental shelf of Trinity Peninsula. The freezing-point of sea water 

 of salinity 34-50 °/ 00 is approximately - i-88° C, and thus the surface water at these 



