SCIENTIFIC RESULTS 197 



temperature and salinity, and the stratified condition of the water 

 under natural conditions are all difficult to reproduce exactly and 

 failure to do ,so ma}^ be vital. For instance, Pettersson used a block 

 of ice in his tank experiment which extended one-third the depth 

 to the bottom, while under natural conditions the proportions of 

 mean draft of ice to the average depth in the North Atlantic is 1 

 to 1,000. 



In the formation of land ice the liberation of the latent heat of 

 freezin<r is due to the atmosphere: wliile the meltino; of land ice (as 

 icebergs) in the sea, withdraws tlie latent heat energy mostly from 

 the water. The reaction, therefore, is somewhat ditl'erent from that 

 in case of sea ice where the processes of latent heat, both freezing and 

 melting, begin and end in the sea itself. While the mean draft of 

 icebergs is about sixty times greater than the mean draft of the sea- 

 ice cover it is only one-fifteenth that of the mean depth of the 

 Atlantic, consequently the thermal effect of the melting of tli,e 

 icebergs is similarly confined to the upper stratum. 



If the effect of the two phases of latent heat, as we just qualified, 

 were always exerted within the same general region, heat energy 

 would be liberated and consumed in the same place. It is because 

 of the transportation by the w'inds and the ocean currents of (a) 

 the ice itself, and (h) the surrounding chilled thaw water that 

 distant temperate climes are actually affected. As preAdously stated, 

 the pack ice, drifting southward in two prominent streams, the East 

 Greenland and Labrador, seems to exhibit a marked tendency to 

 remain inside the continental edges of the respective shelves. This 

 may be i^artly illusory, for any ice wdiich drifts out across into 

 warmer oceanic w^ater quickly disappears through rapid melting, 

 but only partly so, for both the ice and the current trend on shore. 

 Icebergs are no exception to the above rule, despite the phenomenal 

 drifts occasionally reported, the number of such are comparatively 

 few. This coastal concentrati(m of icebergs in Davis Strait was 

 strikingly emphasized in the sunnner of 1928 when out of a total 

 of over 2,000 bergs sighted by the Marion expedition, only 3 were 

 observed outside the 5t)0-fathom contour (see fig. ?-3, p. 142). Thus 

 only a very small percentage of the northern ice actually floats out 

 into the Atlantic basin to melt directly in contact with the ocean 

 water itself. Consequently, while the ice melts in the surface layers 

 of the coastal zone, it cools oceanic waters only indirectly by the 

 mixing of the two waters, and the extent of this mixing, considering 

 the ocean generally, is surprisingly small. 



We found (see p. 190) that glacial ice in the Davis Strait-Baffin 

 Bay region is only about 2 per cent of the volume of sea ice. In 

 this connection the ice area and the melting area of these regions 

 were calculated as 467,300 square miles. (See fig. 121, p. 200.) Em- 

 ploying the foregoing data we are uoav ready to determine quantita- 

 tively the magnitude of the chilling effect of ice melting on the w^ater 

 masses of the North Atlantic. The fact that melting area and ice 

 area are approximately equal (see p. 190) allows a direct comparison 

 between the relative thickness of the ice and the depth of the surface 

 layer through which the chilling effect of ice melting directly per- 

 meates. We have taken 80 feet as the depth of the surface layer 

 which is continually kept in turbulence by waves and local currents. 

 Assuminc; that the meltine: of 1 cubic foot of ice will cool 80 cubic 



