NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 281 



means of the values thus obtained in curve V of figure 99. This 

 curve also shows a certain similarity to the curve S of the " solar 

 constant," but the agreement is not as good as that of the other 

 curves. 



This analysis of the conditions in 191 5 seems to prove that the 

 daily changes in the " solar constant " cause changes in the dis- 

 tribution of pressure which in the region of Norway occur as a rule 

 some days later. And the changes thus produced in the distribu- 

 tion of pressure cause changes in the temperature as a rule a day or 

 two later. The probability is thus that by daily measurements of the 



2000 

 80 

 60 



^0 



10 

 1900 



+st 



o"c 



-¥ 



solar constant it might be possible to predict meteorological changes 

 several days beforehand. 



As another interesting feature may be pointed out that the curves 

 of figure 99 show very distinctly a period of between 24 and 25 days 

 in the solar radiation, as well as in the pressure difference between 

 Christiania and Bergen and in the temperature at Bergen. There 

 are also indications of a shorter period of about 12 days, which is 

 especially conspicuous in curve V, representing the barometric vari- 

 ability from day to day at Bergen. 



For other years for which fairly complete series of observations 

 of the "solar constant " were obtained at Mount Wilson, we have 

 also computed the seven-day means of the " solar constant," of the 

 simultaneous temperatures at Bergen and of the pressure difference 



19 



