17 
til i Middel 34.96 og for det i 1878 undersøgte melleni 
70de og 80de Breddegrad beliggende Strøg til 35.64 -°/ 0< - 
Fuldstøendigt tilsvarende er- det' af Buchanan fuudet at 
være paa den sydlige Halvkugle. idet Syrstofprocenten i 
Overfladen ' der varierer fra omkring 33 i Æ(jvator egnene 
til ca. .35 omkring den ’sydlige Polarcirkel. 
Dette lod med temmelig stor Bestemthod formode, 
at de for destilleret Vand gjældende; af Bunsen opstillede, 
Absorbtionscoefficienter ikke skulde være gyldige for Søvand, 
idet det. naar Oveyflådévandsluftens Sammensætning fandj.es 
at variere med Blinklen, maatte være det Naturligste at 
skrive disse Variationer paa Temperaturforandringen^*. Det 
vil imidlertid ikke føre til ligget rimeligt Resultat, om man 
foreliggende Obsefvationer forsøgo at .udlede en Lov for 
Absorbtionscoefficienternes Variationer med Temperaturen. 
Man vil da finde, at de enkelte Observationer staa ikke 
ubetydeligt i. Strid med hinande.n, idet der. for Surstof- 
mængdernes Vedkommende overalt optræder meget større 
Afvigelser, end man kan -antage begrundede i Observations- 
feil. D.et kunde dog ikke synes tilraadeligt at lade det 
bero hermed og forsøge at di^cutere- de foreliggende Ob- 
servationer udeir nøiere Kjendskab fil Absorbtionscoeftici- 
enternes Afhængighed af Temperaturen, og jeg besluttede 
derfor at bestemipe saavel Sammensætning som Mængde af 
.den i Søvandet ved forskjellige Temperaturer opløste Luft. 
Først gjordes en Del Forsøg, hvorved Søvandet i et 
Bad af constant Temperatur søgtes mættet ved flere Timers 
Gjennemledning. af Luft, (saaledes som Bunsen har gaaet 
frem ved sine Besjemmelser 1 ), hvorefter den opløste Luft 
uddrcves og analyseredes paa den før beskrevne Maade. 
De paa denne Maade mættede Vandprøver afgave bestan- 
dig Luftmamgder, som uden Hensyn til .den Temperatur, 
hvorved Våndet var .mættet. viste nogenlunde nær den 
samme Sammensætning (med 34.9 °/o Surstof mod' (55.1 °/ 0 
Kvælstof), medeus de ofte temmelig stærkfc afvigende Tal. 
som udtrykte de absolute Mængder af opløste (laser, tyde- 
ligt viste, at . der paa denne Maade ikke var opnaaet fuld- 
stændig Mætning. 
Professor Waage foreslog mig derfor at gjentage disse * 
Forsøg med nbgen Variation i den Maade. hvorpaa Ma*t- 
ningen ivæfksattes, og hår jeg som Følge déraf ved de 
senere Forsøg benyttet følgende Fremgangsmaade. En pas- 
sende Portion Søvand af nogenlunde høi Egenvægt rystedcs 
med Luft i en rummefig Kolbe i et Tidsrum fra 1 til 2 
Timer under stadig Vexlen af den i Kolben værende Luft 
og hensattes derpaa i nogle ' Timer ganske rolig, idet Tem- 
peraturen saavel under Kystningen som senere holdtes fuld- 
stændig constant. Forat -overbevise mig om. at jeg har 
opnaaet fuldstændig Mætning, bar jeg nærmet mig Mæt- 
Norwegian Expedition in 1876 and 1877. amounts to 34.96. 
and for that lying between the 70th and 80th parallels of 
latitude, to 35.64 per cent. Buchanan observed precisely 
tlu* same phenomenon in thy southern hemisphere, the 
proportion of oxygen varying from about 33 per cent in 
the Equatorial Seas to about 35 per cent in tlu* vicinity of 
the Antarctic Circle. 
Reasoning on these data, there -were strong grounds 
to- assume, that tlu* coefficients of absorption given by 
Bunsen for distilled water could not apply to sea-water ; 
for, the composition of the air in surface-water having been 
found to. vary with the latitude, the most probable cause 
of this' phenomenon would seem to he temperature. Mean- 
while. we shall not arrive at a satisfactory result by regard- 
ing temperature as the variable factor, and by seek- 
ing from the observations here sot forth to discover a 
law. according to which the coefficients of absorption 
vary with the temperature. The individual observations 
would in that case be found ' to clash, inasmuch as the 
variation with regard to oxygen is invariably greater 
than can he assumed to arise from errors of oliservation. 
However,, it did not seem advisable to leave the question 
as it stood, and proceed to the discussion of the results 
without having further investigated the relation of the coeffi- 
cients of absorption to the temperature : and 1 resolved, there- 
fore. on determining alike the composition and the amount • 
Of the air absorbed by sea-water at different temperatures. 
A series of experiments were first instituted with 
a .view to saturate sea-water wjtli air. viz. by placing 
it in a bath of constant temperature, and for the space 
of several hours uninterruptedly conducting through it a 
current of air,’ — tlu* mode of operation adopted by 
Bunsen for his determinations, 1 — after which the air ab- 
sorbed in the water was driven off, and analysed by the 
process previously ■ described. The samples of water satu- 
rated in this manner invariably yielded quantities of air 
which, irrespective of the temperature at which the water 
had been saturated, were found to be very nearly uni- 
form in composition, viz. 34.9 per cent oxygen and 65. 1 per 
cent nitrogen, whereas’ the figures, often widely divergent, 
■expressing tlu* absolute quantities of the gaseous bodies 
absorbed* gave sufficient proof that by this method complete 
saturation had not been attained. * 
At Professor Waago’s suggestion, 1 repeated these 
experiments, varying slightly the means by which satura- 
tion was sought to he effected, and have since adopted tlu* 
following mode of operation. A quantity of sea-water, ol 
considerable specific gravity, is shaken, along with air, in 
a roomy matrass for one or two hours, the air in the 
matrass being frequently renewed, and then left perfectly 
still for a few hours, at the precise temperature preserved 
during its continual agitation. To be quite sure that 1 have 
really succeeded in saturating the water, I approach- the 
point of saturation as it were from opposite, directions: 
1 Bunsen. Gasoiu. Metkoden — 1 
‘ Bunsen, Gnsom. MethodCn. |>. Bin. 
