55 



bottle and the middle bottle. During the post-season cruise the 

 nominal depths of observations were the same as listed above with the 

 addition of the depths 2,000 and 2,500 meters. With these added 

 depths the deep series usually consisted of seven bottles with an un- 

 protected thermometer attached to every second bottle. Two 

 thermometers were usually attached to each bottle. No bottles 

 were reversed until at least 10 minutes had elapsed after they were 

 in place. After being hauled to the surface and the bottles placed 

 on the rack on the after bulkhead of the pilothouse the thermometers 

 were allowed at least 30 minutes in which to attain temperature 

 equilibrium with the air before they were read. The thermometers 

 were read by either the oceanographer or his assistant. The readings 

 w^ere taken to the nearest 1/200° C. with a magnifying glass and were 

 corrected by means of individual graphs such as are described in the 

 Hydrographic Review, volume X, no. 1, May 1933, pages 126-130. 



As mentioned above, the thermometers were used in pairs and were 

 occasionally shifted with respect to each other so that during the 

 season each thermometer had been paired with several other thermom- 

 eters, thus furnishing a check against the possibility of any thermom- 

 eter or thermometers suffering a change of calibration during the 

 period of observations. An investigation of these data was the basis 

 for determining the reliability of the temperatures read and corrected 

 under the conditions existing on the General Greene. A total of 46 

 groups of readings of different pairs of thermometers was investigated, 

 each group having from 10 to 20 comparisons. For each group the 

 probable difference between the corrected readings of the pair was 

 computed by the method of least squares. These probable differences 

 ranged from 0.006° to 0.017° C. with an average value of 0.010° C. 

 It is therefore considered that the observed temperatures are accurate 

 to 0.01° C. 



The actual depths of the observations were determined from the 

 unprotected thermometers. After the actual depths of the observa- 

 tions had been determined, vertical distribution curves of tempera- 

 ture and salinity were drawn and from these were scaled the values 

 corresponding to the nominal depths mentioned above. These 

 scaled values were then used in the dynamic calculations which were 

 made by the oceanographer at sea as the data were collected. The 

 calculations were carried out using anomalies of specific volume and 

 dynamic depth after Sverdrup's recent tables.^ 



The current maps prepared during the season follow the practice of 

 past years in showing the topography of the sea surface with respect 

 to the 1,000 decibar surface. The post-season current map shows 

 the topography of the sea surface with respect to the 1,500 decibar 



' H. U. Sverdrup, Vereinfachtes Verfahren zur Berechnung der Druck- und Massenverteilung im 

 Meere, Publikasjoner fra Chr. Michelsens Institutt, no. 26; Oslo 1933. 



