16 FRIDTJOF NANSEN. M.-N. Kl. 
Atlantic Current!, this relation is altered during the summer, and the 
former surface-water becomes often much lighter than the latter, and has 
consequently a tendency to spread out over it. When Amundsen crossed 
the Atlantic Current (Gulf Stream) west of Bear Island, on June 14—15, 
1901, the surface density (0;) was to a great extent between 27'80 and 
27'90, whilst further west near the ice east of Greenland the surface 
density in June sank frequently below 27'50, and 27'40, and later even 
much lower. 
If the observations in these regions be not fairly simultaneous, it is, 
therefore, very difficult to draw the isohalines, and it will necessarily 
become a matter of judgement which observations ought to carry most 
weight. It seems probable, however, that the isohalines of Pl. I, on the 
whole, give a fairly correct idea of the horizontal distribution of the sali- 
nity at the surface for the end of June and the beginning of July, 1901. 
The isohalines of 35'1, 35'0, 34°9, and 34'8 00 show the course of 
the Atlantic Current, or Gulf Stream, in this region. The waters with 
comparatively high salinity extend far westwards in the region east and 
north-east of Jan Mayen. This seems to be a very general feature occurring 
in most years, according to the observations made with the Michael Sars and 
others. But in the latitude of 73° N., or between 72 and 73° N., water 
with salinities below 34'8 and 34°7 00, and with comparatively low tempe- 
ratures, extends eastwards to about 10° E. Long. where, however, a narrow. 
branch of Atlantic water with comparatively high salinity and temperature 
extends northwards and follows almost exactly the edge of the continental 
shelf, west of Bear Island. It ends as a very narrow tongue west of 
Prince Charles Forland, off the Spitsbergen Coast. 
This Atlantic water seems to have a tendency to send off a branch 
of water with salinity above 34'8 00 westwards between 74 and 77 N. 
Lat., and to perform a partially cyclonic movement in this region, which, 
however, is not very clearly demonstrated by the observations in June 
and July, when this movement is obviously much altered; but it is pro- 
bably more prominent, even on the surface, earlier in the season and at 
the end of the winter? (see also the map of temperature and salinity at 
50 metres, PI. V). 
! Cf. for instance Amundsen’s surface densities (0,) in the Barents Sea, in April and May, 
1901, which were very frequently about or even above 28'00 (see Table I, o,), and 
considerably higher than those of the Atlantic current north of Norway at the same time. 
2 This is demonstrated by the observations which will be published in the Memoir by 
Helland-Hansen and the present writer, “On the Physical Oceanography of the 
Norwegian Sea”, Report on Norwegian Fishery- and Marine Investigations, Vol. II, No. 2. 
