PHYSICAL OCEANOGRAPHY 237 



experiment, but are not in general use. The one consists in 

 measuring the refractivity of the water, i.e. the deflection under- 

 gone by a ray of monochromatic hght when passing from air to 

 water ; this quantity, again, stands in definite relation to the 

 salinity of the sample. The other method is based on the 

 electrolytic conductivity of sea-water, and has the advantage 

 that no sample need be brought up, a pair of electrodes being 

 simply sent down to any required depth and the readings being 

 taken on board. This method has been applied by Martin 

 Knudsen with good results in shallow water. 



The most convenient, and on the whole the most satis- Chemical 

 factory, method of determining salinity is a chemical one, and is '"^^^^o^^- 

 based on the fixed relation between the chlorine contained in 

 a sea-water and its total salinity. 



The amount of chlorine can be determined by a rapid and chlorine 

 easy method. When a solution of silver nitrate is added to ^^^'^^^'o"- 

 sea-water, the chlorine is thrown down as a white precipitate of 

 silver chloride. If a few drops of yellow chromate of potassium 

 are added it is easy to see when all the chlorine is precipitated, 

 for the silver nitrate will then act on the chromate so that the 

 yellow colour is changed into red. When the chlorine content 

 of a water-sample is to be determined, a certain quantity 

 {e.g. 15 c.c.) is measured off and poured into a glass; a few 

 drops of the yellow chromate solution are added as an indicator, 

 and then nitrate of silver from a burette, that is, a graduated 

 glass tube with a stopcock (for discharge) at the lower end 

 (see Fig. 164). When the red colour appears, the burette is 

 read off to find out how much silver solution has been added, 

 and it is easy from this value to calculate the amount of 

 chlorine. From Knudsen's Hydrographical Tables the salinity 

 or the specific gravity, corresponding to this chlorine-value found 

 by titration, may be determined. All this can now be done 

 quickly and accurately ; in fact, the salinity of a water-sample 

 is determined in less than five minutes to within about j-^ per 

 iiiille, i.e. i centigram of salt per kilogram of sea- water. The 

 modern method of chlorine titration is a great improvement on 

 former methods, and it has been much used in recent oceano- 

 graphical work, thousands of such determinations being now 

 made yearly. 



The density of sea-water depends both on the salinity and Density of 

 on the temperature ; the water is comparatively light when ^ea-water. 

 the salinity is low and the temperature high, and increases 

 in density with a rise of salinity and a fall of temperature. 



