NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 223 



We see here also, what we have hitherto often found, that no gen- 

 eral rule can be laid down. Some parts of the curve go with the 

 sun spot curve and others oppositely to it. It is the curves for 

 the winter months which lend to the curve of the year its dis- 

 tinctive character. We see that until about 1853 this curve went 

 generally oppositely to the sun spot curve, but after this time it 

 went at least as much directly with the sun spot curve. 



In figure 81 we give values obtained in the same way by combined 

 two-year and three-year consecutively smoothed means for Stock- 

 lato 1820 mo mo ibso mo i&ju mo i89o iqoo 1910 



Figure 81. Anomalies of the air temperature in Stockholm and Batavia 

 for February, July, and for the whole year in combined two- and three-year 

 smoothing. 



holm for February, for July, and for the whole year, in comparison 

 with the sun spot curve. It is clearly shown here that the curve 

 for July goes to a great degree opposite to the curve for February, 

 while this latter in general goes inverted to the sun spot curve, but 

 partly also directly with it. The curve for February has the great- 

 est similarity to the yearly curve. 



In the same figure we give also curves obtained in the same way 

 by smoothing the temperature at Batavia for February, July, and 



