16 TEMPERATURE. 



Method of taking Temperature Observations. — The screen was visited each morning between 

 8 and 8-30 a.m. As soon as the door had been opened the pen of the thermograph was 

 depressed, the time noted, and the dry bulb thermometer read. Thus the time and tempera- 

 ture recorded by the therraograi^h were controlled. The maximum and minimum thermo- 

 meters were then read and reset, immediately afterwards all three thermometers were read 

 again and theii readings recorded. 



MetJiod of Reducing the Thermograph Records. — The mean temperature at Cape Evans for 

 nineteen months was obtained from hourly values of the thermograph records, hence it is 

 important that these records should be correctly reduced. The method employed was different 

 for the two instruments. 



(a) The Screen Thermograph. — The papers supplied by the makers on which temperature 

 and time lines are engraved were used. If these lines had been correctly drawn and the 

 rate of the clock and the setting of the pen adjusted once for all it would only have been 

 necessary to read ofE the temperature at the correct time intervals. It is well known, however, 

 that these instruments can never be set with the required accuracy, hence the necessity for 

 some method of reduction which will allow for the errors of time and pen setting. It has 

 been stated above that every day the pen was depre.ssed at a recorded time. Hence on 

 every sheet there were seven time checks. Taking these checks into account it was easv 

 to mark the trace by a series of dots, each one of which corresponded to the exact hour. 

 At the instant the time mark was made each morning the temperature of the dry bulb 

 thermometer was read and recorded. A comparison of the reading of the trace and these 

 recorded temperatures ga^e the error of the thermograph. The mean error for the week 

 was calculated from the seven observations and applied to the trace throughout the week. 

 It will be seen that by this method the thermograph record was reduced to the scale of 

 the dry bulb thermometer. 



(b) The Hut Thermograph. — This was a daily instrument, a new paper being put on to the 

 drum each morning at 8 o'clock. These papers were blank. The pen of the thermograph 

 was depressed every hour by an electric current controlled by the standard clock so that 

 accurate time was shown on the record itself. The thermometer fixed near the bulb of the 

 thermograph was read every four hours and these readings were entered on to the trace 

 at the corresponding times. The deflection of the thermograph for a given change of tem- 

 perature had been previously determined and a glass scale constructed. This scale was used 

 for marking the po.sition of O^F. on the sheet at each of the eight points determined by 

 the eye-readings and through these marks a line was drawn from which the trace at each 

 hour was measured with the glass scale. Thus the recorded temperatures were reduced to 

 the scale of the thermometer attached to the bulb of the thermograph. 



The Accuracy of the Temperatures Observed. — It has just been pointed out that the two 

 thermograph records depended on the scales of the two thermometers. Both these thermo- 

 meter's, when tested at Kew before the expedition sailed, had no errors over the range of 

 temperatures to which they were exposed in the Antarctic. Unfortunately the}' were left 

 in the Antarctic so they could not be compared on the return. Still they were compared 

 in the Antarctic with other thermometers and there were no indications of any change 

 having taken place. Thus the hourly values of temperature may be considered to be quite 

 correct. 



The accuracy of the maximum and minimum thermometers must be considered in two 

 periods— (a) up to the end of February, 1912, and (b) after this date. 



During the former of these two periods the maximum and minimum thermometers were 

 read immediately after setting each morning and compated with the simultaneous reading of 



