NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC IO9 



by weak full drawn lines and the average result for both months 

 by strong dotted lines. There are also given curves for the anomalies 

 of surface temperature (curves W), for the anomalies for the air 

 temperature (curves L) and finally the anomalies of the surface 

 temperature minus the air temperature (curves W-L). The figures 

 at the middle and right hand side relate to the route New York 

 to the Channel. The figures on the left relate to that from New 

 York to Portugal, although the three most westerly io° fields of 

 the route New York to the Channel are included in these figures 

 while the values for the three easterly fields are the mean values of 

 all 10° longitude fields between Portugal and 40° west longitude. 

 These are made up of all fields between 37° and 35° north latitude 

 and between 10° and 20°, 20° and 30°, 30° and 40° west longitude. 

 If we examine these charts for the different years closely and com- 

 pare them with the curves of the figures, we see that on the whole 

 there is a good correspondence between the anomalies of the sur- 

 face temperature and the air pressure gradients. This comes dis- 

 tinctly to view both in the charts and in the figures. Particularly 

 good is the agreement in the years when the air pressure gradients 

 were large, that is to say, when the air circulation was strong, as 

 for example, in the years 1899 and 1903. The years 1898, 1906, 

 1907 and 1908 may also be included in this remark. The agree- 

 ment is less good in the years when the air pressure gradients are 

 less and in consequence the wind was weaker. In this connection 

 we may particularly mention the years 1900 and 1902 when the 

 agreement was less satisfactory. 



THE WINDS ARE THE PRINCIPAL CAUSE OF TEMPERATURE VARIATIONS 



ON THE SURFACE AND IN THE AIR UPON 



THE NORTH ATLANTIC 



Already the charts and curves of these plates have sufficed to 

 show that the wind in most years has a very strong influence on the 

 temperature variations in the field we have investigated. We obtain 

 perhaps the strongest impression that this must be the case if we 

 examine the curves of plates 42 to 46 which give for each of our 

 10° longitude fields and for the Danish fields the variations from 

 year to year in the anomalies of the pressure gradients for the dif- 

 ferent months (January to March), of the surface temperatures, of 

 the air temperatures, and of the surface temperatures minus the air 

 temperatures. These curves show us that the agreement is not 

 particularly good in the most western and most eastern fields. On 



