DAVIS STItAIT AND LABRADOR SEA 19 



method is not sufficiently free from erratic results for the purposes 

 of the International Ice Patrol and the new model bridjjje is looked 

 upon as an essential instrument. 



From the deeper layers in the vicinity southwest of Green- 

 land for which the unusually high salinities were found in 1933, 

 double samples were taken and were measured by silver nitrate titra- 

 tion in addition to the routine bridge measurements. Fourteen 

 samples were so measured, each sample being titrated twice, the 

 titration taking place within 48 hours after collection. In the case 

 of 13 of the 14 samples no third titration was necessaiy and the 

 titration values were consistently higher than the salinity bridge 

 values, the differences ranging from 0.03%o to 0.065%o salinity with 

 an average difference of 0.048%o salinity. In the case of the remain- 

 ing sample (Station 1764, 735 meters) the bridge gave 34.955%o on 

 July 14, the first titration gave 35.05%o on July 14, and the second 

 titration gave 34.99%o on July 14. As there was insufficient silver 

 nitrate solution prepared to make a third titration that day, the 

 sample was set aside and titrated again on July 16, when a value of 

 34.96%o was obtained. Not enough of the sample remained for a 

 fourth titration. 



The consistent discrepancy is somewhat puzzling. The persistence 

 of the difference, in magnitude and sign, makes it improbable that 

 the precision of the measurements is at fault. There seem but two 

 remaining explanations — (1) that the calibration of the bridge was 

 faulty and (2) that the relation of conductivity to total halogens 

 was different here than elsewhere. The fact that the same batch 

 of Copenhagen standard water was used for the measurements as 

 for the calibration of the bridge leaves no doubt but that the cali- 

 bration curve was correct at the salinity of the standard water. 

 Further, because the salinity of the 13 samples in question covered 

 but a small range of salinities (34.88%o to 34.93%o with an aver- 

 age of 34.912Voo') very close to the salinity of the standard water 

 (35.018%o) it does not seem possible that the calibration of the 

 bridge was at fault. This leaves as probable only the possibilities 

 that the conductivity varies among different tubes of the same batch 

 of standard w^ater or that the relation of conductivity to total 

 halogens was different in this water than elsewhere. 



The depths of the observations in 1934 were determined by the use 

 of unprotected thermometers. Five such instruments were used in 

 conjunction with protected thermometers. The shallow series always 

 carried an unprotected thermometer on its deepest bottle. At stations 

 where two series were necessary the deep series usually consisted of 

 seven bottles, the uppermost, deepest, and alternate intermediate 

 bottles being equipped with unprotected thermometers. The pres- 

 sure coefficients given in the Physikalish-Technische Reichsanstalt 

 test certificates for the instruments were used as given. 



The dynamic computations for the stations occupied in 1928 and 

 1934 have been made by means of anomaly tables published by Sver- 

 drup (1933) ; and for the years 1931 and 1933 after the manner de- 

 scribed by Smith (1926). The dynamic heights for those stations 

 shallower than the common reference depth have been computed by 

 means of the method described by Helland -Hansen (1934) for all 4 

 years. 



