NAVAL RESEARCH LABORATORY 



The point to be brought forth here is that Knudsen's Hydrographical Tables (in wide- 

 spread use by oceanographers) have been shown to hold very well over the normal range 

 of the concentration of sea water but are not necessarily valid for highly diluted or 

 concentrated sea water. The diluted samples (from the Baltic Sea) used in preparing 

 Knudsen's tables had been diluted by river water containing relatively large quantities 

 of dissolved solids. Thus, the equation relating salinity to chlorinity shows a salinity 

 of 0.03 ppt for zero chlorinity. The salinity determined by titration of sea water that 

 had been diluted by water containing a lesser quantity of dissolved solids would be smaller 

 that that obtained from Knudsen's tables. Conversely, if the sea water were diluted with 

 water containing a greater quantity of dissolved solids, the salinity obtained from Knudsen's 

 tables would be too low. 



The validity of the chlorinity-salinity-density relationships as established by the 

 International Commission depends upon the ratios of the more abimdant substances in 

 sea water being virtually constant. However, "chlorinities'" determined by titration of 

 sea water diluted by melting sea ice were consistently higher than those computed from 

 density measurements (10,11). In this case, the diluting water was essentially distilled 

 water. The dependence of the Cl-S-density relations on the dissolved solids content 

 of the diluting water and the restricted application of these relations to highly diluted 

 water occurring naturally or prepared in the laboratory should be kept in mind. 



The sea water samples used in this investigation were prepared by evaporation 

 under vacuum or by dilution with distilled water of "mid-ocean" water. Previous work 

 has shown the velocity of sound in sea water to be an approximately linear summation of 

 the velocities of the components (7), and thus the validity of this work also depends on 

 the constancy of the ratios of major constituents of sea water. 



VALIDITY OF THESE MEASUREMENTS AT SONAR FREQUENCIES 



A question arises concerning the validity of utilizing at sonar frequencies these 

 velocity measurements obtained at megacycle frequencies. Arelaxational phenomenon 

 exis ts for sea water with a frequency of about 200 kc at 22.5" C. It is true that a small 

 disp^erslbn of^unS velocffy accbmpanTes~a"reiaxational absorption, but the generally 

 accepted work of Kneser (13) shows this dispersion to be very small indeed. He gives 

 the expression 



V^ - Vo =— OmV V 



00 TT 00 



where V o is the sound velocity at very low frequency, V is the sound velocity at very 

 high frequency, and am is the maximum value of the amplitude coefficient of absorption 

 per wavelength for the relaxational frequency N^. This can be written approximately as 



V TT 



Using the approximate value of 50 x 10"* obtained by Wilson (12) for am. and 1.5 x 10^ 

 cm/sec for V, V -V '5' 0.024 meter/sec, a value which is well within experimental error. 



ACKNOWLEDGMENTS 



The author gratefully acknowledges the assistance received in the course of this 

 investigation by many members of the Sound Division of the Naval Research Laboratory. 

 He is especially indebted to R. J. Urick, who suggested the problem, and gave helpful 



^\^i^^Wfc^i»»rA<i''**«£«*Js'A?iT-v5*ya^{f^ . 



