71 
meste opveies af don Forøgelse af samme, der skyldes de 
der herskende lavere Temperaturer. 
Det fremgaar altsaa, at Differentserne mellem Salt- 
gehalten i de atlantiske Overfladelag og de paa Hunden 
hvilende iskolde Yandmasser gjennemsnitlig kun ere meget 
smaa, om de end paa de Puncter, hvor Våndet i de dybere 
Lag besidder en overveiende polar Karakter, turde være 
adskilligt mere fremtrædende. Disse Differentser af hen- 
imod 0.02 °/p ere dog mere end tilstrækkelig store til i de 
nederste mere alkjølede Lag af det atlantiske Vand at 
fremkalde et, som det vil sees af Tabellen, meget tydeligt 
om end svagt Maximum af den specifiske Vægt, livad der 
nærmest bevirkes derved, at Søvandet ved Afkjøling under 
0° nærmer sig sit Tæthedsmaxinium og derfor i Nærheden 
af dette for mindre Temperatur variationer kun forandrer 
sit Volum med næsteh umærkelig smaa' Værdier, saaledes 
at en selv meget ringe Forøgelse af Saltgehalten under 
disse Omstændigheder faat en overveiende IudHydelse lige- 
overfor en Grads Forandring af Temperaturen. 
Det er saaledes saa langt fra Tilfælde, at der i de 
specitiske Vægter af de forskjellige Vandlag ligger nogen 
Hindring for Antagelsen af, at det atlantiske Vand skulde 
synke .gjenuem det koldere Polarvand, at man tvertom af , 
disse maa, slutte, at .saa maa være Tilfælde, dersom ikke 
andre i Havet herskende Strømme virke hemmende paa en 
saadan Bevægelse. Man tænke sig f. Ex. ved Siden af 
hinanden i Havet to Vamdsøiler af 2000 Favnes (3658 
Meters) Dybde, hvori Temperaturens Variation med Dybden 
for Simpelheds Skyld kan antages at være den samme, 
hvorimod Saltgehalten i den Iste helt igjemiem sættes til 
3.52 °/ 0f medens den i den 2den paa Strøget fra Overdåden 
til 500 Favnes (014 Meters) Dyb gives Værdien 3.52 °/ 0 
og fra 5U0 til 2000 (914 til 3658) Værdien 3.50 °/ 0 , saa- 
ledes som Forholdet ifølge Observationerne virkelig synes 
at stille sig paa enkelte Steder i det undersøgte Hav. Det 
er da umiddelbart indlysende, at en saadan Fordeling af 
Saltgehalten vil have en Synkning i den Iste Søile til Følge, 
saaledes at Våndet i denne vil søge at udbrede sig langs 
Hunden og fordrive det omliggende specifisk lettere Vand. 
Den Hastighed, hvormed en saadan Bevægelse foregaar, 
vil naturligvis rette sig efter Differentsen mellem Trykkene 
i samme Niveau i begge Søiler, en Differents. som ved 
Bunden i 200U Favnes (3658 Meters) Dyb efter Beregning 
beløber sig til henimod 32 mm Kviksølvsøile. 
For nærmere at begimnde den forhen iremsatte Hvpo- 
these, om at det paa Bunden hvilende iskolde \ and paa 
de i Kartet med rød Farve betegnede Steder skulde have 
atlantisk Oprindelse, vil jeg benytte mig af de i en tidligere 
Afhandling 1 beskrevne Observationer over de i Søvandet 
iudeholdte Kvælstofnuengder, hvis Anvendelse i saadaut 
Øiemed allerede paa det Sted løselig er ble ven antydet. 
Som bekjendt herskede der i ældre Tider den An- 
skuelse. at de i Søvandet i de større Dyb indeholdte Luft- 
mængder paa Grtind af det der herskende Tryk maatte 
by a decrease in the amount of salt, is almost compensated 
by the rise resulting from the low temperatures prevailing 
there. 
It appears, therefore, that the differences between the 
amount of salt in the warm upper strata and that in the 
cold water at the bottom, are, on an average, exceedingly 
small, though more striking perhaps in localities where the 
water of the deeper strata to a very great extent is Polar 
in origin. These differences — about 0.02 per cent — are, 
however, as will be seen from the Table, more than suffi- 
cient in the deepest and coldest strata of Atlantic water 
to occasion an appreciable, though a low. maximum of 
specific gravity, which is explained by the fact, that sea- 
water below 0" has very nearly reached its maximum of 
density, and the increase in volume then resulting from 
trifling variations in temperature, is a well nigh inappreci- 
able magnitude: hence, under such circumstances, the in- 
fluence of a very slight addition to the amount of salt 
with but one degree’s difference in temperature will be ex- 
ceedingly great. 
Such being the case, there is nothing in the specific 
gravities observed in the different strata of water to dis- 
favour the assumption that the comparatively warm At- 
lantic water should sink through the cold water of Polar 
origin; nay, from these specific gravities we may infer its 
correctness, provided only that such descending motion 
be not counteracted by the effect of ocean currents. To 
give an illustration. Let us imagine two columns of 
water, 2000 fathoms (3658 metres) deep, in both of 
which, for convenience' sake, the variation in temperature 
with the depth is assumed to be equal ; the amount of salt 
on the other hand, being put at 3.52 per cent throughout 
the whole of the first, but in the second, at 3.7)2 per cent 
from the surface to a depth of 500 fathoms (914 metres), 
at 3.50 per cent from 500 to 2000 fathoms (914—3658 
metres). — a ratio of distribution actually observed in some 
localities. This given, it is obvious that such a distribution 
must cause the water in the first column to sink, and spread 
itself over the bottom, displacing as it does so the speci- 
fically lighter. The rapidity of this downward motion will 
of course be proportionate with the difference in pressure 
at the same level in the two columns, a difference which, 
at the depth of 2000 fathoms (3658 metres), has been 
computed equal to that of a column of mercury 32 wm in 
•height. 
With a view to furnish additional confirmation of 
the hypothesis brought forward above, which assumes the 
cold bottom -water in the red-coloured sections of the 
Plate to be of Atlantic origin, 1 shall have recourse to 
my observations on the amount of nitrogen in sea-water, 
published in a former paper, 1 where their application to 
such a purpose was briefly alluded to. 
The opinion formerly entertained, that the quantity 
of air contained in sea-water at great depths must be ex- 
ceedingly great, by reason of the immense pressure pre- 
’ "Om Luften i Søvarittet/’ 
1 “On the Air in Sea-Water." 
