NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 21 



The curves of surface temperature minus air temperature show, 

 in general, a similar march to the other curves. First they rise 

 from the east to the west, then show a well-marked minimum 

 at the " wedge ", then a new rise and a sudden fall towards the 

 American coast. The difference between the temperature of the 

 water and the air is in the easterly part of the region rather small, 

 about 1.2° C. in February and 0.7° C. in March- April. Near 43° 

 west longitude the difference is 3.6° and 3.0° C, while at 49° west 

 longitude (in the "wedge") it is only 1.8° and 1.4° C. In con- 

 trast with the temperatures of the water and the air, this difference 

 reaches an absolute maximum west of the " wedge " in 63° west 

 longitude, giving in February 5.6° C, and at 61° west longitude in 

 March-April with a difference of 4.2° C. 



It is apparent that the difference between the temperatures of the 

 water and the air in the first decade group is on the whole consider- 

 ably greater than in the second. This is because the water reaches 

 its temperature minimum considerably later than does the air. This 

 feature is yet more clearly shown by comparison of the single decades. 



In order to study the developments from decade to decade, we 

 have combined the observations in the northerly steamer route in 

 larger fields of 20° in longitude. The results are given in the fol- 

 lowing tables where our decades are designated by Roman figures. 

 The temperatures are given as mean values of the eleven-year nor- 

 mals for all the chosen 2° fields (see fig. i) which come within the 

 20° fields above mentioned. 



In these tables we give the mean values for the three combined 

 great fields. These mean values, therefore, indicate the tempera- 

 ture relations for the whole width of the Atlantic Ocean from the 

 beginning of February until the middle of April. They are graph- 

 ically represented in figures 11 and 12. 



The surface tetnperattire for the whole region sho'ws a long ex- 

 tended minimum. The three decade values marked II, III and V, 

 from the middle of February on to the second half of March are 

 11.84°, 11.82° and 11.83° C. The variations of these numbers are 

 less than the margin of error. In general one can draw the con- 

 clusion that at this time a well-marked vertical convection exists. 

 Great water masses, from the surface to very considerable depths, 

 are being cooled, so that the variation of temperature of the surface 

 is strongly damped. The consequence is that no well-marked tem- 

 perature minimum can be recognized. This strongly supports our 

 preliminary assumption that the surface temperatures in the second 



