88 FRIDTJOF NANSEN. M.-N. KI. 
must sink deeper and deeper, gradually to approach a position of 
equilibrium, whilst near the surface the heavy water will be replaced 
by lighter water coming in laterally. 
Summary. The process according to which the heavy bottom water 
is formed must consequently be the following: the heavy surface water coo- 
led down to a density greater than 28°10 must sink, but as long as the surface 
cooling continues, new heavy water is continuously formed to replace it and 
sink in its turn. The water probably does not sink in vertical direction, 
but spreads out laterally, and finds its way in under the lighter cold 
water ofthe Polar Current towards the west and north, as also in under 
the lighter warmer water ofthe Atlantic Current towards east and south. 
Warm water runs in everywhere under the cold toplayer of the 
Polar Current where-ever there is a chance, and forms the intermediate 
warmer layer between the cold polar water and the underlying bottom- 
water, having a density just between those of these two waters. But 
the intermediate warm water-layer is kept out in the above region whilst 
the formation of bottom-water is going on near the surface because this 
sinking water is heavier, at all depths from the surface downwards. 
There is consequently no room for the warm intermediate water to run 
in, and it would have to rise to the surface, where however, it will be 
cooled down and transformed into cold bottom-water. 
As soon as the cooling at the surface ceases in April and May, no 
more cold water is formed to replace that which sinks, and lighter water 
will consequently be found in its place near the surface. The later the 
date of observation in the warm season, the deeper will therefore the 
isotherms of —1° C. and —1'2" C. and the isopycnal of 28:10 probably be 
found. For it is clear that conditions like those seen e. g, in Section 
IX (Pl. X), at Amundsen’s Stations 20, 21, and ı3, cannot be stable, 
and they cannot last long unless there is some external source of con- 
tinuous renewal; the isopycnal of 28:10 is too steeply inclined towards 
both sides. This heavy water must evidently sink to greater depths and 
spread out towards the sides, in endeavouring to attain a position of 
equilibrium. The sinking of the isopycnal towards a position of equili- 
brium will continue during summer, and autumn until the cooling at the 
surface during the winter has again reestablished the former conditions 1. 
i Prof. O. Pettersson (op. cit.) believes that the bottom-water of the Norwegian Sea 
is Polar water coming from the North Polar Basin with the East Greenland Polar Cur- 
rent, but its low temperature and comparatively high salinity prove that this is im- 
possible. Besides, if the cold heavy water at Amundsen’s Stations where actually 
moving southwards from the North Polar Basin it would be extremely difficult to ex- 
eo 
