1906. No.3. AMUNDSEN'S OCEANOGRAPHIC OBSERVATIONS IN 1901. 99 
as by intermixture with the overlying warmer layers. The probability 
would then be that there is some kind of communication between the deep 
basins of the Norwegian Sea and the North Polar Sea; and that there 
can be no high ridge between Spitsbergen and Greenland as the writer 
assumed in the Memoir cited. 
This ridge, if it exists, must in that case be very low, rising perhaps 
to depths where the bottom-water of the Northern Norwegian Sea, be- 
tween Spitsbergen and northern Greenland, has a temperature of about 
—1° C. This low ridge would then prevent the coldest bottom-water 
of the deepest basin in the Norwegian Sea from running into the North 
Polar Basin. It has, however, been seen that the bottom-water is pro- 
bably heated from about —ı'3° C. to about —ı'ı° C. on its way from 
74° N. Lat. to the southern part of the Norwegian Sea. And a similar 
heating must be considered likely on the much longer way through the 
North Polar Basin, which must be considered as like a great fjord, 
where movement of the deep water is extremely slow. The higher 
temperature is, therefore, no hindance in the way of assuming that the 
bottom-water of the latter is the same as that of the Norwegian Sea. 
It is only the higher salinity which seems to stand in the way of accep- 
ting the above explanation. 
The writer has, therefore, again examined whether there is no 
possibility that the values of salinity in the North Polar Basin, as given 
in the Memoir, are not in spite of every possible care much too high}, 
All his own observations made with the hydrometer, in 1894? have 
been revised and the readings corrected for the absolute minimum and 
maximum corrections ? of the instrument (Hydrometer Äderman No. 2). 
er: 
17-3 
= = 
sc) and salinity have been com- 
19 8 
The values of specific gravity (S 
puted by means of Knudsen’s Tables #, 
{ Op. cit. pp. 146 et seg. 
2 Op. cit. pp. 168—184. 
3 I. e. the instrumental errors at maximum and minimum of surface-tension of the sea- 
water. Cf. op. cit. p. 330. 
4-The coefficient of thermal expansion for the glass of the hydrometer has been assumed 
to be between 0'000026 and 0'000028. 5 
As stated in the memoir, the salinities, somputed from the specific gravity (s ae 
ETS OC: 
by Tornøe's formula [Salinity °/,, = (s ae =) 1315] are, for values about 
350 39, about 0°16 9/4, higher than the values obtained from the same specific gra- 
vities by Knudsen’s Tables. If the observations with the hydrometer be computed by 
Knudsen’s Tables, they will, however, also give somewhat lower specific gravities than 
published in the memoir; this is especially the case*with observations taken at somewhat 
low temperatures. And thus the final values of the salinity will be still more reduced. 
