GEOGRAPHICAL DISTRIBUTION OF TEMPERATURE— SEA-LEVEL. 21 



is almost exactly a replica of the first. It will also be noticed that if the temperatures 

 recorded during the day (shown by the dots on the diagram) only bad been considered the 

 mean temperature of the period would have been much too high. Without the independent 

 evidence of the low night temperatures shown in the upper curve which were obtained with 

 a sling thermometer, one would have been in doubt as to the accuracy of the minimum 

 temperatures shown in the lower diagram found by exposing the minimum thermometer under 

 the sledge. 



These two curves show clearly that the daily variations of temperature arc nuich too 

 great to be neglected, and that true mean temperature cannot be obtained without taking 

 this variation into account. Nearly all the temperature observations made on sledge journeys 

 have, therefore, been plotted and the mean daily temperatures determined by reading the curve 

 at four-hourly intervals. 



The only exceptions are the tempeiatures of the first Return Party and of Cherry-Garrard's 

 Relief Party. In these two cases minimum temperatures were not I'ecorded while observations were 

 taken in the morning, in the evening and near to midday. If the midday temperatures had been 

 taken without counterbalancing night temperatures the means would have been too high, hence in 

 these two cases the morning and evening temperatui'es only have been used in obtaining the mean 

 daily temperatures. 



Geographical Distributiox of Temperature. ; ■; 



Sea-level Tempmdures. 



Winter Temperature on the Barrier.— V\ hen the Discovery was in the Antarctic, 1902-04, 

 the regular temperature observations were made in a Stevenson screen erected on the sea 

 ice near to Hut Point, and other temperature observations were made at a position a little 

 to the south of Cape Armitage (marked A on figure 2). The distance between the two 

 positions was only one and a half miles, yet the temperatures recorded at Cape Armitage 

 during the winter were verv much lower than those observed at the same time at Hut Point. 

 Differences of from 10° to 20° F. in the recorded minimum temperatures were quite common, 

 and differences of even 25°F. between the two places occurred at times. These large 

 differences could not be understood and Mr. Dines, who discussed the Discovery temperature 

 observations, was inclined to believe that ' the thermometer screen at the Discovery was placed 

 too near the tide crack, and that the temperatures are unduly high on account of the 

 influence of the sea-water in the crack.' Captain Scott strongly repudiated this explanation, 

 and those who are familiar with ice conditions will agree with him that even if there had 

 been a tide crack near the screen, which there was not, it could not have affected the 

 temperatures to anything like the extent shown by the observations. 



The explanation \nll be apparent at once when we have considered the temperature 

 observations made on the journey now to be described. 



A party led by Wilson left Cape Evans on 27th January, 1911, to visit the emperor pen- 

 guin rookery at Cape Crozier. The path they took is shown on figure 2. They went first 

 over the sea ice to Hut Point, and while on this part of their journey the temperatures 

 they measured were similar but slightly higher than those measured simultaneous!}' at Cape 

 Evans. At 1-30 p.m. on the 28th, when they were just off Hut Point, they measured - 26°F. 

 which was the temperature at the same time at Cape Evans. They then rounded Cape 

 Armitage and apjiroached the edge of the Barrier, which at this point was only 12 feet 

 above the sea ice. Wilson in his report states: 'Coming down the snow slope oft' the Barrier 

 was a stream of very cold air w-hich we felt first when we were only a few yards from the 

 foot, and lost very soon after reaching the top.' At 7 p.m. they camped on the Barrier, 

 and at 8 p.m., when meteorological observations were taken, the temperature was found to 



