46 



top) one of the thermometers was of the unprotected type, in order 

 to determine the accurate depth where the water bottles tripped and 

 study the curvature of the wire. This procedure proved to be of 

 great importance, because wind and current caused considerable 

 wire angles on many stations; several times we measured more than 

 40° from the horizontal. 



All thermometer readings were made by help of a special-made 

 reading lens. The accuracy of the temperature observations is 

 0.01° C. 



The pressure coefficient ("Standerhohung") for the unprotected 

 thermometers changed a little from time to time during the season. 

 Corrections and comparisons were made whenever we had calm 

 weather and vertical wires. 



The sea water samples were titrated with silver nitrate on board by 

 Bernh. S. Loebig, yeoman, first class, who for several weeks before 

 the inauguration of the ice patrol had been trained in titration and 

 other oceanographic methods with the senior physical oceanographer. 

 Each water sample was titrated twice on board; if the difference in 

 salinity for two titrations of the same sample was 0.02 per mille or 

 more, a third titration was made. The accuracy of the salinity deter- 

 minations is 0.01 per mille. 



The salinometer worked well; but it usually took a long time 

 before the reading settled. The cells are too voluminous; it was 

 impossible to have a water sample from the citrate bottle determined 

 twice with the salinometer. The apparatus was, therefore, only used 

 during our time in port for comparing with the results of the titra- 

 tions. Even in the roughest weather we never had any difficulties 

 in performing the titration work at sea. 



All corrections and the hydrodynamical calculations were made 

 on board on the cruises; the oceanographer also prepared the current 

 maps, which at the end of each survey were delivered to the ice-patrol 

 cutter at sea. 



The oceanographic data and the dynamic calculations will be seen 

 on pages 51-64. The current maps (figs. 18 to 21) show the dynamic 

 topography of the sea surface constructed in the usual way after the 

 Bjerknes's theory of circulation. The sea-surface relief has been 

 obtained by comparing it with that of a plane at a depth of 1,000 

 decibars. The djmamic height contours are drawn for every 2 

 dynamic centimeters, with a heavy line for each 10 dynamic centi- 

 meter. Where the depth of the deepest observation of temperature, 

 salinity, and density is less than 1,000 meters, the method of J. P. 

 Jacobsen and J. C. Jensen has been used (published in Rapport et 

 Proces verbeaux des Reunion, Vol. XXXIX). This method gives 

 good results, except in some cases when the difterence in depth between 

 two stations in different water masses is large. The positions of the 



