NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 95 



If we consider now the distribution of anomalies in the single 

 decades from decade to decade, as shown in our isopleth diagram 

 of plate 27, we may possibly see indications that in the ocean east- 

 ward of 50° west longitude a certain displacement toward the east 

 in the greater and smaller anomalies takes place from decade to 

 decade, but this displacement is not marked and is very irregular, 

 which last is no doubt partially due to the inaccuracy of the material. 



RELATION BETWEEN SURFACE TEMPERATURE AND AIR TEMPERATURE 



The question whether the variations which we are investigating 

 depend upon changes of the temperature of the masses of water 

 brought on by the ocean currents or whether they depend upon 

 other causes must be settled by the relation between the tempera- 

 ture of the air and the temperature of the ocean surface. Since 

 we have seen the march of the variations in the temperature of the 

 air and o'f the ocean surface, at least on the whole, go in the same 

 direction, we must expect that if the variations in the temperature 

 of the ocean surface are the primary cause this would precede the 

 variations in the temperature of the air and call them forth. 



Since the temperature of the air is on the whole lower than that 

 of the ocean surface, it follows that the addition of colder water- 

 masses by ocean currents tend to bring the temperature of the ocean 

 surface more nearly towards that of the temperature of the air, so 

 that the difference between these would be less than the normal. 

 If the temperature of the ocean surface in consequence of the trans- 

 portation of warmer masses of water is raised by the ocean cur- 

 rent, the temperature of the surface will depart from that of the 

 air and the difference between them will become greater than nor- 

 mal. This is, however, not always the case as one may see by 

 our figures (pis. 42 to 45, curves W and L, and W-L, also figs. 48 

 to 52). We shall now examine how the special minimum of the 

 years 1903 and 1904 goes with respect to this view. 



In the field 50° to 55° west longitude the difference, surface 

 temperature minus air temperature, in February, 1903, was nearly 

 normal, as shown in plate 42. In March-April, 1903, the difference 

 is considerably smaller than normal (see pi. 42). In February, 

 1904, the difference is a little greater than normal, but in March- 

 April, 1904, it is considerably less than normal. Thus it seems to 

 appear on the whole as if the possibility that the variations in the 

 temperature depend upon variations of the quantity of heat brought 

 bv the water-masses is not shut out. 



