6. Since ice floes move in different directions even when their drift speed is the same (see 

 figures 136 and 138), they rotate when they come into contact with each other. Of course, rotation 

 of the floes is even more certain when the floes move at different speeds (figure 137). 



7. If the individual ice fields are so large that one must consider the difference in the wind 

 direction at their edges, the fields will rotate as follows, as is evident from an examination of the 

 figures: floes situated both to the right and to the left but near the path of the cyclone center will 

 move counterclockwise, those situated on the periphery of the cyclone region and to the right and 

 left of the path of the cyclone center will move clockwise. 



8. As can be seen from figure 138, which depicts the curvatures of equidistant lines parallel 

 and perpendicular to the movement of the cyclone at the initial moment, there are three main 

 directions of movement of the floes (on the map) during the passage of a cyclone: 



1) Upward. 



2) From right to left in the upper region of cyclone influence. 



3) From left to right in the lower region of cyclone influence. 



Reduction of the horizontal distance between scattered floes may be expressed only in con- 

 centration. However, if the ice is close-packed and thick enough so that the wind characteristic of 

 the cyclone can produce only slight hummocking, the floes which are already in the cyclone region, 

 on drifting from left to right, will encounter floes before them which have not yet entered the 

 cyclone region. Thus, floes may be set in motion and begin to hummock long before the wind 

 begins. This phenomenon, hummocking preceding a wind, has frequently been noted by polar re- 

 searchers (see Section 95). 



The phenomena in the region above the sphere of influence of the cyclone may prove to be 

 still more interesting. Here, under these same conditions (great concentration and thickness of 

 the ice, limiting intense hummocking) ice drift and hummocking may occur during a complete still. 

 Such cases have also been observed frequently by polar researchers. 



Thus far we have examined the movement of ice during the passage of a cyclone over an un- 

 limited ice area. 



Let us assume that the vertical dashed line AB (on the left) in figure 138 is the edge of ice 

 before the passage of the cyclone. Of course, after the passage of the cyclone the position of this 

 ice edge will be depicted by the extreme left-hand curve of figure 138. In section AC , the ice edge 

 will be scattered, while in section CB it will be concentrated. 



Now let us assume that the path of the circular cyclone with a constant pressure gradient 

 intersects a straight channel filled with ice at an angle of 45° (figure 139). Naturally, in this case 

 the movements of the floes will be restricted by the coastline and this will intensify the scattering 

 effect in some places and the hummocking in others. Let us assume, further, that during 

 hummocking the length of the ice fields in the direction of movement of the cyclone and at the wind 

 force characteristic of this cyclone will not be reduced by more than 1/4. 



Figure 139 shows the path of the center of the cyclone, the limits of its influence, the regions 

 of compression and scattering of the ice and also the direction of rotation of the ice fields. By 

 modifying the position of the coastline and the directions of the cyclone paths relative to it one may 

 approximate natural conditions. Under the given assumption, even a case as schematic as the one 

 depicted in figure 139 can be applied, for example, to the Kara Sea. If the paths of the cyclone 

 centers move from west to east somewhere in the region of Cape Zhelaniya (Mys Zhelaniya), the 



380 



