Thanks to the work of Nansen we know that underneath the cold and non-saline surface water 

 of the central part of the Arctic Basin lies the warm Atlantic water. The speed with which this warm 

 water moves to the east may be determined from the following considerations, which are partially 

 confirmed by the studies of Dobrovolski: It is believed that about 4 years is required for polar ice 

 to move from east to west — from Bering Strait to Greenland Sea — and it is natural to assume that 

 the same time interval is necessary for movement of Atlantic water in the reverse direction, i. e. , 

 from Spitzbergen to Bering Strait. 



We have as yet by no means discovered how the subsequent eastern passage of deep Atlantic 

 water along the continental slope of the Siberian coast affects ice conditions in the corresponding 

 regions. There is no doubt, however, that the temperature of this water does have an influence on 

 ice abundance, for these reasons: first, because this water is involved in one way or another in the 

 vertical circulation which accompanies ice formation, and second, because this water (after mix- 

 ing with other water) comes out on the surface of the sea in certain definite areas, as a result of 

 driving phenomena which are caused by one factor or another. 



This discussion must be supplemented by the following. The presence of anomalous warm or 

 anomalous cold water in one sea region or another brings about an anomalous distribution of 

 meteorological conditions, in particular pressure patterns and as a result, a distribution of winds. 

 But if the sea temperature is connected on the one hand with ice abundance and on the other with a 

 pressure pattern, there should then be a certain dependent relationship between pressure and ice 

 abundance. Therefore the efforts to find this relationship from observations of ice abundance and 

 pressure distribution are quite understandable. 



Meinardus came to the conclusion that a weak air circulation in the north Atlantic from 

 August to February accounts for the comparatively small quantities of floating ice at Newfoundland 

 in the following spring, and the converse is also true. Meeking considered that the pressure 

 gradient across the ice current along the shores of Labrador in the preceding winter is the main 

 factor which determines the boundaries of floating ice in the northwest part of the north Atlantic. 



Lesgaft, studsdng ice conditions in the Kara Sea from 1869 to 1911, determined that with 

 comparatively favorable ice conditions in the north and in the southeast of the Barents Sea, favor- 

 able conditions are established likewise in the northern part of the Kara Sea, north of the northern 

 end of Novaya Zemlya and in Matochkin Shar. Likevwse, imfavorable ice conditions in the northern 

 and southeastern parts of the Barents Sea are associated with unfavorable conditions in the northern 

 part of the Kara Sea. From such facts Lesgaft concluded that in these parts of the Arctic Ocean the 

 ice conditions are regulated by one general principle. Ice conditions in the southern part of Kara 

 Sea do not depend on the ice conditions of the Barents Sea and are determined by the pressure 

 gradient at Cape Karmakula-Obdorsk. 



Vize subsequently showed that an increase of pressure in northern Greenland and north of 

 Iceland in June or July corresponds to a great quantity of ice in the Barents Sea in the following 

 August, the converse also being true. 



The authors mentioned above have noted that small ice quantity is the result of an appropriate 

 distribution of atmospheric pressure. Other authors, for example Brooks and Kennell, found on 

 the contrary a connection between ice abundance in the polar seas and the subsequent pressure dis- 

 tribution in Western Europe. It must be noted that these are only apparent contradictions, since, 

 as we have often noted, the general circulation of the atmosphere is very closely connected with 

 the general circulation of heat in the ocean. 



463 



