20 
84 , 85 og 8,7. fjerne sig om mere end 1 Procent fra dot al- 
mindeligc Resultat, medens man af samtlige Observationer 
kan bestemme en enkelt Observations sandsynlige Afvigelse 
fra den efter doiino Tabel optrukne Curve til + 0.52 °/ 0 , 
on Afvigelse saa liden, at eii ikke ringe Del af deii kan 
skrives paa Observationsfeil. 
De største Uoverensstemmelser optræde talrigst i et 
Dyb fra 300—600 Favne (540 1097 Meter) men brides, 
ogsaa enkeltvis i større Dyb. 
Fra Buhden liidrører i de større Dyb kun to Lult- 
prøvef mod væsentlig for hoi Surstofprocent nemlig No. 
68 og 87, optagne fra. to Puncter, som mærkeligt nok 
begge ligge paa on Linje paralel med og' tøt ved Grænd- 
sen mellem den nordover strygendo varme • Strøm og den 
sydover forbi Jan Mayen gaaende ' Polarstrøm. Bortser 
man imidlertid fra disse Vie wæsentligste Foverensstemmelser, 
ioin bidrage til at give Curyen et om end meget svagt 
M inimum i 300— 400 Favnes (549 — 732 Meters) Dyb, vil 
man i Kortlied kunne udtale Regelen for Surstofprocentens. 
Aftagen med Dybet saaledes: Surstofprocenteri er i Over- 
laden gjennemsnitlig 35.3 og aftager derpaa først liur- 
tigt senere langsommere til hfenimod 32.5 i 300 Favnes 
(549 Meters) Dyb, hvorira den med stigende Dyb holder 
sig paa det Nærmeste constant. Det kan bemærkes, at af 
de her undersøgte Vandprøver 40 ere øste lige ved Hav- 
lmiiden. Man. vil imidlertid forgjajves bestradie sig for at 
opdage nogen Forskjellighed i Egeiiskaber mellem disse og' 
de fra ligestore interriiediære .Dyb optagne. 
Hvoy det gjælder at studere Variationerne af den 
absolute Luftmængde, maa det synes naturligt som Maal 
lor denne at benytte den opløste lvvælstof,' idet den ob- 
serverede Luftmængde paa Grund af det vedvarende For- 
brug af Surstof i de dybere liggende Lag bestandig kan 
forudsættes at være mere eller mindre forskjellig fra den 
Mængde, som vilde absorberes i Overfladen under direote 
Paavirkning af Atmospluvren. K vælstofmængden kan der- 
imod paa Grund af denpe Gasarts størkt udpr'rogede In- 
differentisme ligeoverfor andre Legemer udjeir. synderlig Feil 
antages.ua fbængig af locale Tilfældigbeder. , 
Anvendes saaledes Kvælstofmængden som Maal * for 
den i Søvandet opløste- Luft, viser der sig i Fordelingen 
ogsaa her en udprægøt Lovimessigbed, naar undtages, at 
der i de af Svendsen paa det første Togt udførte Obser- 
vationer- overalt or tumlet en mindre Kvadstofmængde, end 
man efter alle øvrige foreliggende Observationer skulde 
vente. Bortser man imidlertid fra disse paa første Togt 
udførte 14 Observationer, vil man se, at alle de Øvrige 
paa fan Undtagelser nær meget, vel stemipe overens med 
de Tal. man kan beregne efter den ved de oveuciterede 
Forsøg bestemte Formel 
85, and 87* exhibiting a difference of more than 1 per cent 
as compared with the t general result, whereas the probable 
deviation of a single observation from the curve drawn 
according to this Table may be computed at + 0.52 per 
cent, a deviation so small as to arise, probably,, in no 
slight degree from errors of observation. 
The greatest discrepancies refer chiefly to a depth of 
300 — ooo fathoms (549 — 1097 metres); now and again, 
• however, they were met with in water obtained from greater 
depths. 
In only two of the samples of air expelled froip 
bottom-water drawn where the depth was great, did the 
percentage of oxygen prove much too high, viz. in Nos. 68 ' 
and 87, the samples of water yielding them having beei^ 
obtained from two spots which, strange to say, are in a 
line running parallel * and . in close proximity to the bound- 
ary, between the warm current Hewing north .and the cold 
Arctic current Howing south past the Island of Jail Mayen. 
Now, if we exclude from these differences the chief of those 
that contribute towards giving the curve a very slight 
but appreciable minimum at a. depth of from 300 to 400 
fathoms *(540 — 732 metres), the. yule according to which 
the proportion of oxygen is found to diminish with the 
depth may be expressed as follows: — The proportion of 
oxygen, which at the surface is 35.3 per cent, begins at 
i once and continues to diminish, at first rapidly and after- 
wards at a slower rate, till it hqs reached 32.5 per cent, 
j at tin 1 depth of 300 fathoms (549 metres), from whence it 
1 keeps almost constant. I will not omit to observe, that of 
the samples of water examined 40 had been drawn from 
the bottom: is was, however, impossible to detect any dif- 
ference in composition between these and the samples. ob- 
tained from equal intermediate depths. 
When investigating the variation in the absolute amount 
of air, it will ' obviously be advisable to make use <jf the 
nitrogen absorbed, since the quantity of air observed in 
the deeper strata may, by reason of the steady consump- 
tion of oxygen, he assumed to differ more or less from 
that which would he absorbed . at the surface .under the 
direct inHuence of the atmosphere, whereas nitrogen, from 
the very slight affinity evinced bv that gas for other bodies, 
may, without involving appreciable error, be regarded as 
proof against 'the accidents of locality. 
If. therefore, the amount of nitrogen be adopted as 
•the standard- of measurement for the air absorbed in sea- 
water. a marked uniformity will here, too, be found to cha- 
racterise the distribution, as determined by the observa- 
tions described, with the exception however of Svendsen s, 
on the first voyage, by which the amount of nitrogen was 
found to be less than all subsequently instituted observations 
gave reason to expect. Excluding, then, the 14 observa- 
tions from the first voyage, all of the others, with hut few 
| • exceptions, agree closely with the figures which may he 
' found by the formula stated above — 
N= 1.4.4 7 — 0.23 t. 
N= 14.4 — 0.23 1. 
