AMELIORATIONS OF PRESENT ARCTIC CLIMATES 395 



Perhaps, at first thought, this process will seem to involve the 

 destruction of the warm middle layer, and its existence may seem 

 to be a reason to doubt the verity of the process. The middle 

 layer is of course constantly disappearing by the process but is 

 disappearing only httle by little very locally where crossed by the 

 briny film. The middle layer is, however, a current, and is being 

 continually renewed. It may thus be preserved, though continually 

 suffering loss. The detail of the operation is important to a clear 

 view. The thin, cold, briny sheet in passing down across the middle 

 layer of course exchanges temperatures and salinities with it at the 

 immediate contact surfaces, but very Httle elsewhere. This 

 exchange gives the contact portion of the middle layer greater den- 

 sity than the main portion of that layer, and hence this contact 

 film tends to go down with the briny layer and become part of the 

 basal layer rather than remain as a cooled portion of the middle 

 layer. The rest of the middle layer thus retains its temperature 

 almost unmodified. The gradual loss thus sustained by the middle 

 layer is suppKed by onflow as a mass from the original source, that 

 is, as interpreted below, from the Atlantic middle layer. 



The more special phase of the action has been here described. 

 There is a more general phase. After the surface of the water in a 

 lane freezes over, the vertical walls that border the lane remain, as 

 a rule, until the new ice is crushed by the closing of the lane. So 

 long as the prism of water in the lane thus lasts, any movement 

 of the ice over the water, or of the water under the ice, almost 

 necessarily forces any briny film that may have formed on the top 

 of the prism by continued freezing to move edgewise and thus to 

 encounter the border wall and be turned downward, much as in the 

 previous case. Now, the winds and tides are almost always causing 

 differential movements between the ice-cover and the water on 

 which it floats, and so this downward, edgewise, penetrating action 

 may not be so speci 1 as it seems. 



Atmospheric confirmation. — An interesting atmospheric abnor- 

 mahty has been observed in Grinnell Land by Moss and in Green- 

 land by Krogh, which falls in with this view of salt concentration by 

 freezing and seems to rather pointedly confirm it. These observers, 

 quite independently and at different dates, noticed that when the 

 wind at the localities of observations came from the northwest, 



