24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 'JO 



observations, as is the case with our material. Nevertheless it is 

 also well known that inversions of the air temperature frequently 

 occur, and it not infrequently happens that in February after a rise 

 of temperature, a new fall occurs, so that the decade mean values, 

 even after a long- series of observations, do not show a perfectly 

 smooth march (see Hann 1911, p. 91). 



Local irregularities of this kind, however, in a great degree dis- 

 appear when, as we have done, the final mean values for a very great 

 region are considered. We have taken the mean values for not 

 less than forty-eight 2° fields in our computation of the values 

 which occur in figure 11. In the study of the peculiarities inside the 

 three 20° fields of longitude (see fig. 12) we find that the irregu- 

 larity depends very largely upon the results of the middle fields 

 which have a very marked secondary minimum in the fifth decade. 



The difference between the temperatures of the water and of the 

 air grows gradually less on the whole from the beginning of Febru- 

 ary to the end of April (see fig. 11, W-L). In the first three weeks 

 of March, however, the difference remains about equally great, 

 because then both the air temperature and the water temperature 

 are substantially unchanged. The difference amounts on the whole 

 to about 3° at the beginning of February, and not much more than 

 1° in the middle of April. In this there is, however, a good deal 

 of local difference. 



The tables show the difference in the temperature behavior in the 

 three parts of the region as covered by the 20° fields of longitude, 

 namely: the easterly part, from 10° to 30° west; the middle part, 

 30° to 50° west; and the western part 50° to 70° west — that is to 

 say, west of the " wedge ". The results are expressed graphically 

 in figure 12. 



In all decades the water and the air are both warmer in the 

 middle part of the North Atlantic Ocean. The water is coldest 

 toward the eastern part while the air is coldest toward the western 

 part of the region. The difference between the air temperature 

 and that of the water is greatest in the west and least in the east. 

 The reason for this may be easily understood, for the middle part 

 is under the control of the Gulf Stream, and is there not so strongly 

 cooled. In the western part the cold water of the American coast, 

 in large part the water of the Labrador current, mixes with the 

 Gulf Stream water, so that the mean temperature is made lower. In 

 the eastern part the water masses of the Gulf Stream have finally 

 become cooled. The wind blows on the whole from America toward 



