NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 6 1 



and the eastern fields and the middle ocean region is also clearly 

 marked. 



In figure 2"] we give to tenths of degrees the differences between 

 temperatures in the first decade group, February, and the second 

 decade group, March-April, for the same 4° longitude fields along 

 the steamer route Channel-New York. The bold-faced figures in- 

 dicate here increase of temperature from February to March-April, 

 while the inclined ones indicate cooling. We see that here also a 

 certain regularity or system prevails with regard to the place and 

 time of the distribution of these temperature differences. We find 

 for example that in the year 1903 the temperature diminished from 

 February to March-April over great regions in the middle part of 

 the ocean, while in the year 1904 there was an increase of tempera- 

 ture from February to March-April. In the year 1905 again there 

 was a diminution of temperature during this time interval over the 

 greater part of the fields. This also was the case in the year 1906 

 in the western half of the region, but not in the eastern half. In 

 the first years, 1898 to 1900, there was in all fields a general rise 

 of temperature from February toward March-April. This was also 

 the case in the last year, 1910. 



What we have said about the curves for the anomalies of the 

 surface temperatures in our northern region Channel to New York 

 is in general true of most of the corresponding values and curves 

 in the southern region between 27° and 45° north latitude and be- 

 tween 10° and 40° west longitude. In consequence of the small 

 number of observations we have, as already remarked, reduced the 

 observations within this region to larger fields of 10° in length and 

 2° in width, and in this way twelve such 10° longitude fields were 

 obtained (see fig. i). 



In figure 28 we show graphically by curves the values obtained 

 for the surface temperature in all the years within these twelve 

 fields. The variations of curves agree in all their principle features 

 so well together that they must certainly be closely representative 

 of the truth. It is moreover worth noting that the agreement 

 between the curves is particularly good for those fields which ad- 

 join one another in the direction north to south between the same 

 10° longitude intervals. The agreement is not so good for the 

 fields taken from east to west. Finally there is a good agreement 

 between the curves for these three groups of 10° fields and the 

 curves of the next northerly or northeasterly lying 10° fields of 

 the northern region of our observational material, that is, Channel 

 to New York (see figs. 20 and 28). 



