34 FRIDTJOF NANSEN. M.-N. Kl. 
surface salinity was remarkably high (about 347 and 34'8 0/00), in these 
regions where the sea-water was cooled towards its freezing-point, and 
where much new ice was being formed while Amundsen was there. 
The surface-salinity is during the summer and autumn very low in these 
same regions (see the isohalines PI. I), and sinks even below 340. It 
shows clearly how ice-formation is able appreciably to increase surface- 
salinity. 
When the cooling of the surface ceases in the spring, the coldest 
and heaviest water will gradually sink towards the bottom, and lighter 
and warmer water which is heated from above, wili replace it near the 
surface. The salinity of the top-layers will be reduced by the melting 
of the ice, and also by an increased quantity of fresh-water from 
the rivers and from the land. The salinity of the water near the bottom 
will be gradually lowered by intermixture with the overlying water; and 
it may therefore be expected that higher salinities, for the cold bottom- 
water on the shallow banks, will be found early in the spring and lower 
ones later on (cf. Wollebæk’s Station on May 31, 1900). During the 
course of the late summer and autumn, the cold bottom-water may 
gradually be washed almost entirely away on the shallow banks, by 
intermixture with the overlying warmer layers; but it will remain for a 
long time as a thin layer in the deepest hollows, and will there never 
entirely disappear; especially along the Novaya Zemlya coast}, In the 
deep central hollow of the eastern Barents Sea it will probably always 
form a bottom-layer of greater thickness (see above Figs..2, 3). 
The cold winter water of the southern shallow part of the Barents Sea. 
Amundsen’s Stations ı to 10 are taken on the southern extensive 
coast bank or plateau, with depths less than 100 metres (see Fig. I, 
p. 24), and on its northern slope. They are from an earlier part of the 
season (April and May) than are most other observations known from 
this region, and the vertical series shows very distinct traces of the 
vertical circulation during the previous winter and spring, which has 
produced almost homogeneous cold water between surface and bottom 
at the most typical Station, Stat. 6. Here the density (o;) of the water 
1 In the writer’s memoir on the N. Pol. Basin, where a similar explanation is given (cf. 
op. cit. p. 281) it is suggested that similar bottom-water may also find its way into the 
Kara Sea from the North, and thence through the Kara Strait into the submarine 
channel along the southwestern coast of Novaya Zemlya. This seems now, to be very 
doubtful. On the whole it seems that this cold bottom-water is much more local and 
forms much less of cold bottom-currents than then seemed ratherly likely, and also 
less than Knipowitsch seems to think. 
