ON THE ACCURACY OF THE SALINITY VALUES 



In the preceding chapter Soule has shown that the 

 readings of the salinity bridge were converted into terms 

 of salinity by means of a calibration curve which was 

 obtained by measuring samples in the bridge and titrat- 

 ing the same samples by the ordinnry silver-nitrate 

 method. Owing to this procedure the salinities obtained 

 from bridge readings should, on an average, be equal to 

 salinities determined by titration, but the accidental er- 

 rors of the values would be somewhat greater, amounting 

 to +0.04 per mille, mainly since minor deviation from a 

 constant temperature exercised a considerable influence 

 on the bridge readings. A comparison between the Car - 

 negie salinities and those from other expeditions indi- 

 cates, however, that the Carnegie salinities are, on an 

 average, somewhat too low. This result is arrived at by 

 a study of the conditions at great depths where the salin- 

 ity is very uniform and where no variations from year to 

 year have been detected. 



In his discussion of the deep water of the North At- 

 lantic WiJst (1935) writes (in translation): "In our sa- 

 linity charts at 1500 to 4500 meters the Carnegie salini- 

 ties in the open North Atlantic Ocean appear to be on an 

 average 0.03 to 0.04 per mille too low, as also shown 

 from a comparison between the TS-curves at the Carne - 

 gie stations and neighboring stations from other expedi- 

 tions (in single cases the deviations of the Carnegie sa- 

 linities vary between -0.10 and 0.02 per mille." 



In the Pacific Ocean the salinity of the deep water 

 has been determined on two later expeditions, the Dana 

 expedition in 1928 to 1930, and the Bushnell in 1934. 

 The Dana observations have not been communicated in 

 detail, but some of them have been used in special pub- 

 lications. In Schott's (1935) "Geographic des Indischen 

 und Stillen Ozeans," salinities and temperatures are 

 given at the depth of 3000 meters at a station in latitude 

 20° south and longitude 174° east, and at depths of 3000, 

 4000, and 5110 meters at a station in latitude 19° south 

 and longitude 163° west. These values can be compared 

 with Carnegie observations at depths greater than 3000 

 meters at stations 87 to 91 which are located between 

 latitudes 15° and 18° south and longitudes 145° and 160° 

 west. We find: 



The U.S.S. Bushnell undertook oceanographic work 

 in the North Pacific, occupying eighteen stations between 

 Adak, Aleutian Islands and Oahu, Hawaiian Islands, to 

 depths of 2500 to 3500 meters. The observations! below 

 3000 meters can be compared with the Carnegie obser- 

 vations below 3000 meters at stations 122, 142, 144, and 



146, of which station 122 is located near the Aleutian Is- 

 lands and stations 142, 144, and 146 nearer the Hawaiian 

 Islands. We obtain the following table. 



Stations 



Lati- 

 tude 



°N 



Longi - 

 tude 



°W 



Mean 

 depth 



Mean 



temp. 



°C 



Mean 



salinity 



o/oo 



No. ob- 

 serva- 

 tions 



Bushnell 50-22 158-170 3396 1.501 34.675 13 

 C arnegie 46-32 140-174 3637 1.528 34.644 11 



If we consider 34.67 as the characteristic salinity of 

 this region we find that the single Carnegie salinities 

 deviate from -0.07 to 0.00 per mille from this value. 

 Thus, the range of variation is less than in the North 

 Atlantic, indicating that the accuracy of single determi- 

 nations was greater during the latter part of the cruise 

 than during the first part. 



Compiling these different comparisons we obtain the 

 following differences between the salinity of the deep 

 water as determined on other expeditions and on the 

 cruise of the Carnegie : 



North Atlantic 

 0.03 to 0.04 



South Pacific 

 0.027 



North Pacific 

 0.031 



From the systematic character of these differences 

 we must conclude that the Carnegie values of the salini- 

 ty of the deep water are about 0.03 per mille too low. 

 It follows that all salinity values between 34.6 and 35.0 

 per mille are too low by the same amount, but it has not 

 been possible to find the cause of this systematic dis- 

 crepancy, nor has it been possible to decide whether or 

 not a similar discrepancy is present at other values of 

 the salinity. 



All tables and graphs had been prepared in final 

 form before this systematic discrepancy was discovered, 

 for which reason the original Carnegie values have not 

 been changed, but in the te.xt attention has been drawn to 

 the discrepancy in all cases in which the exact value of 

 the salinity of the deep water has been discussed. 



It may be added that the discrepancy will not influ- 

 ence the results of the dynamic computations, if it has 

 the character of a constant difference, but if the differ- 

 ence depends on the absolute value of the salinity, an 

 error is introduced in the results of such computations. 

 This error will not be serious since it will only influ- 

 ence the data from the upper layers and will no doubt be 

 smaller than uncertainties arising from lack of knowl- 

 edge as to periodic or aperiodic variations in these 

 layers. 



^ These were kindly placed at the author's disposal 

 by the Scripps Institution of Oceanography. 



LITERATURE CITED 



Schott, G 1935. Geographie des Indischen und Stillen 

 Ozeans. p. 203. Hamburg. 



Wiist, G. 1935. Wissensch. Ergebn. d. Deut. Atlantischen 

 Exped. Meteor 1925-27, vol. 6, no. 1, p. 230, footnote. 



77 



