NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 



71 



for February and March-April for the corresponding most easterly 

 fields as shown in figure 36. We have taken the yearly value for 

 1903 from Petersen's own drawing (1912). The correspondence 

 between the yearly curve and the curve for February and March 

 for Petersen's station No. I is obviously not so good as the agree- 

 ment between these yearly curves and the February curve for our 

 most easterly fields between 10° and 14° west longitude. See also 

 the curve for the field between 10° and 20° west longitude, shown 



1838 99 'iSOO 1 



Figure 35. Petersen's stations I to XII along the shipping course Channel 

 to New York between 11° and 60° west longitude. Curves for the yearly- 

 anomalies of the surface temperatures computed from September i of the 

 previous year to August 31 of the given year. 



in figure 20, 10° to 19° west, and see also the fields south of it 

 between 10° and 20° west longitude, 43° and 44° north latitude 

 shown in figure 36, 10° to 19° west. The reason for this we attri- 

 bute to the fact that our curves, which are determined from much 

 more extensive observational material, are more trustworthy than 

 the monthly curves for Petersen's station No. i. 



It is noticeable that the yearly curve for Petersen's most westerly 

 station, that is to say, No. 12 at 66° west longitude between 40° 

 and 41° north latitude, showing as it does a maximum in the year 



