180 DISCOVERY REPORTS 



(Sandstrom, 1919). Other methods, such as the isentropic analysis used by Montgomery (1938), have 

 in certain cases an advantage, but after experimentation with both methods, and bearing in mind that 

 the observations do not properly satisfy the requirements of either, a straightforward presentation of 

 the dynamic height anomalies has been made. A few words may be said about the theoretical implica- 

 tions of this method in so far as the present work is concerned. 



SALI 



NITY '/v, 



WS996 



20- 



5 J 



345 

 _l ■ ' ' 1_ 



350 



35 5 



-J 1 I 



WSI05& 



15- 



10" 



345 350 



J I I I I 1 1 I 



355 



_j 1 



WS986 o 



,00/ ( 



/ 



'JOO 



W5 968 



C D 



Fig- 33- Temperature-salinity relationships at 'William Scoresby' stations, (a) WS 996 and WS 1000 (broken line) off 

 Orange river mouth, survey I. (b) WS 1052 (broken line) and WS 1056 off Orange river mouth, survey II. (c) WS 976 

 (broken line), WS 979 (dotted line) and WS 980 off Walvis Bay, survey I. (d) WS 986 and WS 988 (broken line) off Sylvia 

 Hill, survey I. Positions of stations are shown in Figs. 1 and 2. 



First it should be noted that Bjerknes's theorem applies to currents in which a state of 'stationary 

 motion ' is maintained. In other words, it assumes that while the current flows the distribution of 

 mass in the water remains unaltered. This implies that no mixing, or change in properties of the 

 water particles along their path, may take place, and that no vertical translatory movement should 

 occur. On these assumptions the horizontal movement of the water can be related to the distribution 



