MEAN MONTHLY TEMPERATURE. 



•79 



frequency is less. Both these conditions are consistent with rapid rises of temperature and 

 slow falls. This is exactly what would occur with the removal and formation of cold layers, 

 for such a layer can be removed almost instantaneously by the setting in of a wind from 

 any direction, but the fall of temperature would take place much more slowly, as the air cools 

 under the abnormal radiation when calm weather sets in again. 



Mean Monthly Temperature. 



Description of the Data Available. 



McMurdo Sound. 'Discovery' Expedition, Hut Point 1902-04.— On pages 373 to 384 of the 

 volume of meteorological results of the National Antarctic Expedition 1901-04 are given a 

 complete set of two-hourly temperature observations made at Hut Point from February 9, 

 1902, to the end of January, 1904. Thus with the exception of eight days in February, 1902, 

 the record is complete for two whole years. The mean values given on these pages are used 

 in the following and not those given on page 464 of the same work which appear to have 

 been computed from defective data ; the difference, however, is small. 



' Nimrod ' Expedition, Cape Royds, 1908-09. — THe meteorological observations made on 

 Shackleton's expedition have not yet been published, but mean monthly values of the tem- 

 perature recorded at Cape Royds were given to Professor Hann who has published them on 

 page 690 of volume III of his ' C'limatologie.' No information has been given as to the 

 method of obtaining the monthly values, but probably they are based on two-hourly readings 

 of the thermometers. 



'Terra Nova' Expedition, Cape Evans, 1911-12. — There is a complete set of hourly data for 

 the period February, 1911, to August, 1912. In addition, observations of daily maximum and 

 minimum temperatures are available for 18 days in January, 1911, and for the remainder of 

 1912 after the end of August. The maximum and minimum temperatures can be used for 

 finding the true mean temperature if the corrections to be applied are known. Fortunately 

 it is possible to determine a close approximation to these corrections by comparing the 

 observations of mean maximum and mininuim \\ith the true mean from observations made in 

 three years when the observations are complete. 



In the following table are entered the differences between the true mean temperature 

 based on either hourly or two-hourly observations, and the mean of the maximum and 

 minimum temperatures, from which the corrections to be applied are found. 



Table 37. 



Corrections to he Applied to the Mean of the Maximum and Minimum Temperatures in Order to 



Obtain the True Mean Temperature. 



