604 CYCLES OF ORGANIC AND INORGANIC SUBSTANCES 



increase the precision of measurement by the new techniques and 

 to estabHsh with high precision the relations between the classical 

 quantities and conductivity but also to ensure that new research 

 ships should be able to maintain a vertical wire under all condi- 

 tions when work is possible. 



Salinity and Temperature-Salinity Diagrams 



Our approach has been experimental. We have concentrated on 

 five positions which, with the help of the Decca Navigator system, 

 we are able to maintain within a circle of 1-mile radius. The main 

 station Cavall in 4700 m of water is well out in the northern Bay 

 of Biscay, 70 miles from the continental slope. This station is 

 worked very intensively. Only the waters at Ca\'all below 1250 m 

 will be discussed. The Gulf of Gibraltar water around 1000 m 

 will be ignored. 



Our salinities on deep samples from the cruises on R.V. Sarsia 

 in March, April, and September, 1958. were determined by precision 

 titration by Dr. Riley and Mr. Culkin at the University of Liver- 

 pool (Bather and Riley, 1953). The analyses in April were not so 

 good as the rest, almost certainly due to changes during storage 

 and transportation, and will not be used. The March and September 

 figures were \'ery consistent (Fig. 2). There is a change in both 

 scales of four times at 5.5° and salinity 35.21%. The slope is un- 

 changed as between 5.05° and 8°, as had been established in earlier 

 years. It is seen that the water between 5.05° and 8° is in neutral 

 adiabatic equilibrium at a potential density 27.80 sigma-theta. 

 There is no hindrance to complete mixing with the lower layers 

 of the Gibraltar water, which I shall not discuss today. Although 

 there are large changes in potential temperature and salinity, 

 they completely and exactly compensate each other. 



At depths greater than 1850 m, breakpoints were selected by 

 inspection, linear curves being then fitted by the method of least 

 squares. Around the curve covering depths between 1850 and 2600 

 m, the standard deviation was twice as great as for other ranges 

 of depth, and indicated a source of A'ariability other than experi- 

 mental error. When the oxygen obser\ations, to be discussed 

 later, were taken into account, it became reasonable to construct 



