40 



DISCOVERY REPORTS 



allows the computation of the dynamic heights of the sea surface and other isobaric 

 surfaces at shallow stations relatively to those at deep stations. Prior to the construction 

 of the vertical section of specific volume anomaly, the curves of the distribution of 

 temperature, salinity and density with depth were drawn. In a few cases the values of 

 temperature and salinity were smoothed when this was necessary, and this was done in 

 such a manner that the interpolated values of temperature and salinity were in good 

 agreement with the density and specific volume anomaly curves. The dynamic depths 

 of the various isobaric surfaces are calculated in the usual manner, adding to the known 

 values of the specific volume those which have been artificially obtained in the manner 

 described above. Fig. 51 illustrates the method for obtaining the extra values of 

 specific volume anomaly. 



From the dynamic calculations one station could be selected at which the level of the 

 surface of the sea was lowest, i.e. the most dense water occurred at this station. By 

 taking the dynamic depths above the datum level it was then possible to obtain relative 



7 



Fig. 5 1 . Sketch showing method of extrapolation used in 

 obtaining specific volume anomalies. 



heights of the sea surface at the various stations and find the difference between the 

 other stations and the station selected. Occasionally a datum station was chosen which 

 was not that of the most dense water. When the height of the sea surface above the 

 datum isostere at a station was lower than that of the datum station, this was expressed 

 as a minus difference of level in dynamic centimetres. 



The two principal lines of stations in the Bransfield Strait were from King George 

 Island and from Livingston Island to Trinity Peninsula. By use of the Helland-Hansen 

 and Nansen method the relative height of the sea surface at stations along each line was 

 found. It was then necessary to find a connection between the two lines in order to be 

 able to construct dynamic charts. This has been done by combining a station from one 

 line with a station from the other, either by the Jacobsen and Jensen method, if the two 

 stations in the link were selected so that the difference in depth of the stations was small, 

 or by a direct comparison in the case of the deepest stations where the values of the 



