derive L as 0.025 860 0. Each of the nine values of /v were computed 

 and tiieir average taken as 0.094 6G1 74. Whence 



S-- 



9503.982 924 

 ■ m+ 195.773" 



■3.661 



(2) 



Bearing- in niiiid the stan(hir(Hzation (Ufference necessary in the use 

 of cahhration curves (1) and (2), the following is a suminarv of the 

 coniptirison of tlie results of the precise chlorinity determinations and 

 the salinities derived from the use of cahhration curves (1) and (2) 

 on the Coast Guard hridge and the salinities as measured hy the 

 WHOI hridges Xo. 1 and Xo. 2. 



From this tabulation it wiU be seen that these three bridges agree in 

 showing large differences from tlie chlorinity values for drums Xo. 

 14 and X'^o. 10. Hence these drums have not been used in the 

 computations. 



Applying a corr(M'tion of —0.002° oo to the salinities obtained by 

 WHOI bridge X^o. 2 and then taking the means of results from 

 bridges Xo. 1 and X'o. 2 gives a standard deviation from the salinities 

 derived from the chlorinities of ±0.0039 for the WHOI bridges whose 

 calibration curves are based on an assumed second degree relationship 

 between conductivity and salinity. This may l)e compared with the 

 standard deviation of the Coast Guard bridge salinities fi'om tlie 

 salinities derived from chlorinities of ±0.0042 where the calibration 

 curve is based on an assumed linear relationship between conductivity 

 and salinity. The errors introduced through the use of tlie simpler 

 expression, calibration curves (1) or (2), are therefore not considered 

 to be significant for the salinity range 31 to 36)2°/oo, and the use of 

 calibration curves of this type will be continued. 



From the February 1960, measurements it is now possible to correct 

 the pu])lished salinities for the postwar period 1948 through 1959 on 

 the basis of intercomparison of different batches of Copenhagen water 

 and the standardization values assigned these batches in arriving at 



37 



