NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 211 



was considerably warmer than one would have expected by con- 

 sideration of the temperature in Stockholm. In the following twenty 

 years the coast water temperature was considerably lower than the 

 temperature in Stockholm would indicate. It is possible that peri- 

 odic fluctuations of long interval play a part in this, which produce 

 different effects upon the coast water and upon the air temperature. 

 But the variations which occur in periods of not very many years 

 are quite similar. A certain agreement is found between these 

 C-curves and the sun spot curves, but the eleven-year period in the 

 C-curves shows the tendency to a separation into several (three or 

 four) shorter periods. 



XII. THE RELATION BETWEEN METEOROLOGICAL VARIA- 

 TIONS AND VARIATIONS OF SOLAR ACTIVITY 



PERIODS FOUND IN METEOROLOGICAL VARIATIONS 



The result of the meteorological investigations which we have 

 thus far discussed shows that sometimes very great agreement 

 exists between stations which are far apart and situated in very 

 different regions of the earth. In these discussions we have treated 

 principally the variations of the curves after these have been 

 smoothed by twelve-month consecutive means. These curves show 

 principally the fluctuations of a few years, but they also indicate 

 the longer eleven-year periods. We obtained in this way a good 

 confirmation of Hildebrand's conception of the meteorological varia- 

 tions in their grouping about different action spheres. 



These fluctuations in the meteorological elements which we have 

 studied in different regions of the earth appear to be in a strong 

 degree periodic. Particularly there appears a period of about two 

 or three years in these curves most conspicuously. A correspond- 

 ing period is found frequently in the curves for sun spots (see 

 later fig. 95) and for prominences as well as in the disturbances of 

 the magnetic elements. 



By a proper formation of means for long periods of years we 

 have shown, as also have Koppen and others, that meteorological 

 fluctuations of about eleven years and of about five and a half years 

 occur. In other words, as we have already said, there appear to be 

 periodic variations of the meteorological elements of one-quarter 

 (and perhaps one-third), one-half, and a whole sun spot period 

 widely distributed over the earth. One cannot resist the conclusion 

 that these periods which have so close a relation to solar activity 

 have great influence on the condition of the earth's atmosphere. 



