thermometers, it is considered that the observed temperatures listed 

 in the table of oceanographic data have a probable error of ±0.01° C. 

 Sahrities were determined, as in previous years, with a Wenner 

 salinity bridge. Standardizations were made with water from an oil- 

 sealed carboy of sea water. At least twice during each salinity run 

 a sample of Copenhagen standard water of the batch Pi 8 was meas- 

 ured as an unknown. At the end of each survey these measurements 

 of standard water were used to correct all sahnities determined durmg 

 the survey. Indicated corrections thus obtained were less than 0.005%o 

 for the three surveys made during the season and so no correction 

 was made. The salinities determined during the postseason cruise 

 were corrected to Copenhagen water for each run. The precision of 

 the salinity bridge measurements was about ±0.005%o. Since the 

 cahbration curve of the bridge was determined by silver nitrate titra- 

 tion, however, the accuracy of the salinities is no better than that of 

 the latter method. As all samples from any one station were measured 

 in^ the same salinity run any errors of standardization should not 

 affect differences in salinity from level to level at a given station. 

 Thus any cases of apparent instability arising from differences in 

 salinity of more than 0.0l7oo are considered to be real. It is of inter- 

 est to note that sev^eral such cases of slight instability were found in 

 1952 in the intermediate water of the Labrador Sea. This water has 

 usually been found to be of slight or indifferent stabihty. 



The dynamic topography found during the first survev has been 

 shown on two charts: figure 7 in which the heights of the sea surface 

 have been referred to the 1,500-decibar surface and which covers the 

 northern part of the surveyed area; and figure 8 in which the heights 

 of the sea surface have been referred to the 1,000-decibar surface and 

 which covers the southern part of the surveyed area. For the sake 

 of continuity there is an area in which the two charts overlap. The 

 1,000-decibar surface, which is adequate for the Grand Banks region 

 and preferable in view of the large proportion of shallow water sta- 

 tions there, is not deep enough for the Labrador Sea where the 1,500- 

 decibar surface more nenrly approximates the surface of no motion. 

 The area shown in figure 7 has infrequently been surveyed in summer 

 and has never before been surveyed in April. There is, then, no basis 

 for comparison to judge whether or not the current pattern found is 

 unusual. That the seaward margm of the Labrador Current is found 

 near the continental slope is to be expected from summertime obser- 

 vations. The northern limits of the North Atlantic eddy are inferred 

 to have been located at about 55° N., and therefore are approximately 

 m the same latitude found in summer. It was a surprise to find such 

 large gradients in dynamic height as are shown in the western margins 

 of the northward movmg water. As indicated by figure 7, the water 

 of most pronounced Atlantic characteristics was found at station 4724 

 where the salinities reached values above 35°/oo with a maximum of 



34 



