78 



TEMPERATURE. 



Table 3G. 



The data used in tbe above table are only for one year, but the interdiurnal variation 

 shows, as did the reduced range, that at Cape Evans the non-periodic temperature changes are 

 largest in the winter and smallest in the summer. So far as the observations go the same is 

 the case at Cape Adare and Framheim. Thus in the yearly variation the non-periodic changes 

 follow closely the mean temperature. In every month the mean temperature at Framheim 

 is lower than either of the other two stations, and the interdiurnal variability is greater. 

 At Cape Evans and Cape Adare the same relationship does not hold so closely, but in six 

 months out of ten Cape Evans has both the lower temperature and the greater variation. 

 Cape Adare was very much nearer to permanent open water than Cape Evans, therefore the 

 horizontal temperature gradient must have been much larger at the former than at the latter 

 station, this was particularly the case in June when the mean temperature at Cape Adare 

 was actually lower than at Cape Evans and the open ocean very much nearer. In spite of 

 this. Capo Evans had on the whole the greater temperature variation, which shows that the 

 horizontal temperature gradient is much le.ss efficient in producing temperature clianges than 

 the vertical gradient. The large non-periodic variability at Framheim is particularly instructive. 

 The low temperatures on the Barrier have already been ascribed to the great radiation which 

 is possible over its surface of snow. The large variability of temperature signifies that tbe 

 low Barrier temperature is confined to a low layer above which the air has a much higher 

 relative temperature. In other words the large interdiurnal variation of temperature at Fram- 

 heim strongly supjiorts the view already expressed that there is on the average a greater 

 temperature inversion over the Barrier than over McMurdo Sound, so that the temperature 

 of the upper air over the whole region is much more uniform than the temperature near the 

 ground. 



If we examine the figures in table 35 for the period April to October we find that with 

 few exceptions at all stations the rises of temperature are greater than the falls, but their 



