66 



SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. JO 



We may therefore reasonably assume that these curves are typical 

 representations of the real temperature variations of the ocean sur- 

 face of the Middle Atlantic Ocean for the period v^hich we have 

 investigated. We may also assume that the great and characteris- 

 tic features of these deviations are common to the whole ocean 

 surface. 



The results do not support the conclusion of Petersen that the 

 variations in the surface temperatures for the different months of 

 the year in the eastern and western parts of the Atlantic Ocean 

 tend to go in opposite directions with respect to an axis at 40° 

 west longitude. Referring to curves II and III for the ocean 

 eastward of 40° west longitude and curve IV of the ocean west of 



9* 99 1901) 1 



o o ■.20-59'^n:, UO-M'N. 



0---0-- -o20-39'^n''., ii3-'t8''N. 



Figure 30. Curves of the anomalies of the surface temperatures for 

 February 3 to March 4. 



40° west longitude in figure 29, we see that the principal features of 

 these curves are the same. As shown in figure 40 we find only in 

 isolated years, as February, 1905, and March-April, 1899, such an 

 opposition of temperatures as Petersen assumes. 



If we now consider the observed variations in the surface tem- 

 peratures in the 10° longitude fields for the Danish observations 

 northerly of 50° north latitude, we find here in the middle region 

 of the ocean between 20° and 40° west the same great general 

 features in the variations for February as for March and April. 

 This result is shown by our curves for 20° to 29° west longitude and 

 30° and 39° west longitude in figures 31 and 32. These may be 

 compared for example with the curves of figure 20 and figures 29, 

 30 and 48-W. A great depression may be seen in the February 

 curves for the year 1899 and a still greater one in the year 1904. 



