86 



sea. Copenhagen standard water was used for every series of measure- 

 ments and either Copenhagen standard or a substandard was used in 

 each cell of the salinity bridge once every 10 or 12 measurements. 

 The routine bridge measurements were performed by the oceanog- 

 rapher's assistants, Bernard S. Loebig, yeoman fu-st class, and Edwin 

 H. MacDonald, quartermaster third class. The titrations were 

 made by the former. In the routine measurements each sample was 

 measured twice on the salinity bridge and where the two measurements 

 did not agree to within 0.02 %o in salinity a third measurement was 

 made. The observed salinities are considered to be accurate to within 

 0.02 %o in salinity and to be precise within 0.01 %o in salinity. 



At the 167 stations occupied during the season, serial observations 

 of temperature and salinity extended to depths of between 1,000 and 

 1,400 meters where the depth of water permitted. The levels from 

 which it was attempted to obtain measurements were 0, 25, 50, 75, 

 100, 150, 200, 300, 400, 600, 800, 1,000, and 1,400 meters. The division 

 of bottles between deep and shallow series, the method of determining 

 the depths of observation, the location of unprotected thermometers, 

 the time allowances and precautions tat en in reading the thermom- 

 eters, as well as the method of correcting the thermometer readings 

 were all as described in United States Coast Guard Bulletin No. 24. 

 On the post-season cruise the observations were carried to as near the 

 bottom as practical at all stations, the nominal depths of observations 

 being similar to those listed above from the surface down to 1,000 

 meters and at 500-meter intervals below 1,000 meters. 



As has been mentioned in recent bulletins, the thermometer readings 

 are corrected by means of individual correction graphs such as are 

 described in the Hydrographic Review, volume X, no. 1, May 1933, 

 pages 126-130. In order to facilitate the construction of such correc- 

 tion graphs for protected thermometers table 1 is presented here. 

 This table gives the values of {T' — t) corresponding to each 0.01° cor- 

 rection for all integral values of Vo from 90° to 145°, thus covering the 

 ranges of thermometer constants and temperature conditions usually 

 encountered in the practical use of protected thermometers. The 

 table is based on a reciprocal differential thermal coefficient of expan- 

 sion between mercury and thermometer glass of 6,100. 



