14 



INTERNATIONAL ASPECTS OF OCEANOGRAPHY 



obviously are not free from marked periodic and 

 secular changes of oceanographic factors which on 

 the horizontal charts are expressed as local devia- 

 tions. In the higher latitudes, where such secular 

 changes are especially marked, the stations occupied 

 during the summer half of the year are strongly 

 predominant. But, as has been said, in addition 

 to all of these errors and disturbing factors, comes 

 the uncertainty contingent upon interpolation from 

 inadequate vertical observational intervals, and this 

 source of error is many times the most important. 

 All strongly discrepant values were placed in paren- 

 theses on the horizontal charts, as soon as they could 

 be attributed with some probability to one of the 

 designated sources of error, and in the construction 

 of the isotherms they were considered either not 

 at all or only with caution. The following table 38 

 gives a statistical summary of the number of sta- 

 tions with the serial measurement of temperatures, 

 made since 1873 by research vessels and cable 

 ships in the open Atlantic Ocean and available at the 

 Institut fiir Meereskunde at the beginning of 1934.' 

 Four layers, 200-1000 m., 1250-2000 m., 1250-2000 

 m., 2500-4000 m., and 4000-5000 m., are recognized. 

 The catalogue of sources is given in the Appendix. 

 The detailed station li.st and the four charts (supple- 

 ments XXXII-XXXVI) of the source material 

 make clear the status of the investigation of the 

 open Atlantic Ocean in the four principal layers 

 below 200 meters. Both of the uppermost layers 

 (200-1000, 1250-2000), considering the great extent 

 of the ocean, can be regarded as relatively well 

 investigated (although in the second layer there 

 are less than one half as many serial measurements 

 as in the uppermost layer). But for the deeper 

 layers, greater than 2000 meters, the only relatively 

 well explored regions are the regions investigated by 

 the Meteor, and the South Antilles Sea, the prin- 

 cipal region of work of the Discovery E.xpedition. 

 Of the 743 serial measurements which have yielded 

 the material for the layer between 1250 and 2000 

 meters the Meteor has supplied 275 series (includ- 

 ing the Greenland voyage), and the three ships of 

 the Discovery Expedition have supplied 254 series, 

 which are predominantly in the South Antilles Sea. 

 Then follows the Atlantis with 173 series of closely 



' For this opportunity I express the thanks of the Institut 

 fur Meereskunde to Professor H. Bigelow and Dr. Seiwell 

 for making available manuscript material of the Atlantis 

 Expeditions, to Professor Fleming and Professor H. U. 

 Sverdrup for similar material of the Carnegie Expedition, 

 and to Professor Helland-Hansen, Professor H. U. Sver- 

 drup, and Doctor H. Mosby for such material from the 

 expedition of the Nobvegia. 



spaced stations, along lines of special profiles in the 

 northwest Atlantic. As the charts show, the 

 Meteor in its investigation has placed great weight 

 on the investigation of the deepest levels (> 4000 

 meters), which as a rule have been neglected. Of 

 the 126 series which have supplied values for the 

 horizontal charts (4000 and 5000 meters), the 

 Meteor has contributed 77. 



B. SALINITY AND DENSITY 



In the source material of the salinities we have 

 recognized two fundamentally different groups: 

 (1) Those in which the salinities were determined 

 by physical methods — hydrometer, electrical con- 

 ductivity (salinity-tester) — which show great un- 

 certainties; and (2) Those which depend upon the 

 chemical method of chlorine titration and which 

 because of the standardization of the method 

 (normal water) are mostly comparable. The series 

 of salinities obtained by the use of hydrometers, 

 which constitute only about three per cent of those 

 for the uppermost layers and a still smaller propor- 

 tion for the lower layers, notwithstanding modern 

 methods of handling data, are eliminated from the 

 observational material above considered. After the 

 elimination of entirely defective values through the 

 correlation — salinity, such data are utilized only as 

 auxiliary points in regions that are poor in observa- 

 tions. The method of electrical conductivity (salin- 

 ity-tester), used on the vessels of the Ice Patrol 

 and on the Carnegie' apparently because of ther- 

 mal disturbances, is also not sufficiently accu- 

 rate to recognize the finer differences in salinity in 

 the greater depths. In depths of more than 2000 

 meters we have therefore placed in parentheses 

 those values obtained in this way. Our salinity 

 charts therefore represent only the distribution of 

 the chlorine content which has been transformed 

 into salinity according to the recognized relation of 

 chlorine to salinity.* As already noted, the salinity 



' According to a communication in a letter from Pro- 

 fessor Sverdrup the limit of error of the electrical method 

 in comparison with that of titration for chlorine reaches 

 about 0.04 Voo in salinity. On our salinity charts for 

 1.500-4500 meters in depth the Carnegie salinity in the open 

 Atlantic Ocean shows on the average around 0.03-0.04 Voo 

 too low, as does also a comparative consideration of the 

 curves S-f (t) of the Carnegie stations with the neighboring 

 stations of other expeditions. (In some places the devia- 

 tion of the Carnegie salinites varies between —0.10 and 



+0.02 Voo)- , , , 



» Since doubt has recently been expressed as to whether 

 the composition of sea water is sufficiently constant for such 

 a calculation, it has been proposed by Carter, Moberg, 

 Skogsberg, and Thompson, that it would be more precise to 

 abandon this transformation and in its place present charts 

 of chlorine-content. The author cannot agree to this step 



