NO. 



4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 57 



together and partly because the ocean is here to a great extent af- 

 fected by coast water. However, these curves show certain great 

 features which give the impression of being in agreement with the 

 true conditions. By more accurate studies one also finds that these 

 are by no means accidental. We will later return to this con- 

 sideration. 



In figures 21 and 22 we give curves showing the temperature 

 anomalies for February and for March-April for the chosen six 

 10° longitude fields. These curves are smoothed according to the 



Figures 21 and 22. Smoothed curves (according to the formula i 

 (a + 26 + c) for the temperature anomalies for February (fig. 21) and for 

 March-April (fig. 22) for the same fields as in figure 20. 



formula b' = ^(a + 2b + c). The illustrations show the general fea- 

 tures of the temperature variations in our period very distinctly 

 and regularly. 



In figures 23 to 26 are shown the form of the isopleths according 

 to the results of our investigations along the ship route from the 

 Channel to New York. These relate to the first decade group in 

 February and to the second decade group in March- April. We give 

 for each year and for each decade group the mean of the tempera- 

 tures for all fields of 4° of longitude between 10° and 70° west 

 longitude. The fields are indicated above on the axis of abscissae 

 of the figure, and the years at the side as ordinates. In figures 



