Kulsyrebestemmelser pleiede lian derfor altid for Opera- 
tionens Begyndelse at udfælde Svovlsyren raed concentreret 
Ohlorbariumopløsning, forat Kulsyren lettere skulde und- 
vige, men anvendte forresten den af Dr. Jacobsen angivne 
Methode, hvorved han har bestemt Kulsyren i Søvandet i 
de sydlige Have til i Middel 43.26 Mgr. per Litre 1 . 
Da jeg Vaaren 1877 opfordredes til at gaa ud som 
Chemiker paa den norske Nordhavsexpeditions 2det Togt, 
var der kun levnet mig nogle faa Dage til Forberedelser, 
og det følger derfor af sig selv, at jeg ikke paa nogen 
Maade dengang kunde kave befattet mig med vidtløftigere 
Forundersøgelser, og jeg maatte saaledes uden selv at 
kunne prøve optage de tidligere Methoder uforandrede. 
Paa Togtet i 1877 anvendtes derfor den af Dr. Jacobsen 
angivne Methode. og bestemtes efter denne gjennem en 
Række omhyggelig udførté Observationer Kulsyregehalten 
i det da undersøgte Hav til omkring 100 Mgr. per Litre* 
Der viste sig imidlertid ved Gjentagelse af samme Observa- 
tion bestandig Uoverensstemmelser, som ofte vare ikke ube- 
tydelige og engang endog løb op til hele 12 Mgr. per Litre. 
Dels herved dels ved andre Omstændigheder vaktes 
min Mistanke om Tilforladeligheden af den af Dr. Jacob- 
sen i Forslag bragte Methode. 
Det syntes mig paa Forhaand overnlaade urimeligt, 
at der hos Søvandet skulde findes en saadan mærkelig Evne 
til rent mekanisk at tilbageholde den ene Gasart, medens 
den ingensomhelst Virkning skulde udøve paa de Andre. 
Heller ikke var der nogensinde gjort noget Forsøg paa at 
sætte dette Phænomen i Forbindelse med bekjendte chemi- 
ske Egenskaber hos nogen af de i Søvandet indeholdte Stoffe. 
Ved et Tilfælde kom jeg en Dag -til at forsøge Sø- 
vandets Reaktion paa Lakmus og Rosolsyre og fandt til 
min store Forundring, at det reagerede bestemt og tyde- 
ligt alkalisk, livad jeg siden har bragt i Erfaring, at alle- 
rede v. Bibra 2 og senere E. Guignet og A. Telles 3 har 
observeret. 
Efter mine Forsøg viser to ligestore Prøver af en 
efter Gottliebs 1 Fremgangmaade frisk tilberedt -Lakmus- 
opløsuing. hvoraf den Ene tilsættes en tilstrækkélig Mængde 
Søvand og den Anden et ligestort Volum rent destilleret 
Vand, ikke ubetydelige Farvedifferentser. Ligeledes an- 
tager en med meget fortyndet Oxalsyre svagt udsyret pas- 
sende Portion rent Vand, hvori paa Forhaand er opløst 
en Draabe Rosolsyre, ved Tilsætning af Søvand strax den 
bekjendte rødlig-violette Farve. 
Paa denne Maade undersøgtes paa Expeditionens sidste 
Togt, hvor der var fuld Anledning til at erholde Vand- 
prøverne ganske friske,., et meget stort Antal af disse og 
uden Undtagelse med det samme ovenbeskrevne Resultat. 
1 Ber. Berl. chem. Ges. 11 — 410. 
* Ann Chem. Pharm. 77 — 00. 
8 Compt. rend. S3 — 019. 
4 Journ. fiir pract. Chem. 107 — 488. 
acid, before commencing the operation, by adding to the 
water a saturated solution of chloride of barium, in order 
to facilitate the liberation of the carbonic acid, but. with 
this exception, adopted the method devised by Dr. Jacobsen, 
and determined the mean amount of. carbonic acid present 
in the water of the Southern Seas to hi* 43.26 m t> r per 
litre. 1 • 
When invited, in the spring of 1877, to go out as 
• chemist to the Norwegian North- Atlantic Expedition, on the 
second cruise, I had but a few days in which to make the 
necessary preparations, and consequently no time being left 
me for preliminary experiments, I was compelled to adopt 
unchanged the earlier methods, without testing the accuracy 
of their results. On the cruise in 1877. I therefore ap- 
plied Dr. Jacobsen’s method, and determined by a series of 
careful observations the amount of carbonic acid present 
in the water of the tract then investigated to be about 
100 m ^per litre. But, on repeating the. operation with the 
same sample of water, the results were always found to 
vary, aud frequently indeed considerably ; nay, on one occa- 
sion the difference amounted to as much as 1 2 m ^ r per litre. 
Partly for this reason, and partly from other circum- 
stances. I was led to question the trustworthiness of 
Jacobsen’s method. 
Now it struck me at once as highly improbable that 
sea-water should possess so remarkable a power of retain- 
ing mechanically one gas, and yet. in this respect, exert no 
influence whatever on others. Nor had any attempt been 
made to connect this phenomenon with known chemical 
properties distinguishing the substances contained* in sea- 
water. 
Quite accidentally, I was one day Ted to investigate 
the effect of sea-water as a reagent on litmus and rosolic 
acid, and found its reaction, to my great surprise, dist- 
inctly alkaline, which, indeed, as I subsequently learnt, 
had been already observed, first by von Bibra 8 and later 
by E. Guignet and A. Telles. 3 
According to my experiments, two equal measures of 
a solution of litmus, freshly prepared by Gottlieb’s method, 4 
one of which has added to it a sufficient quantity of sea- 
water and the other an equal volume of pure distilled 
water, exhibit considerable difference in colour. Moreover, 
a proportionate mixture of highly dilute oxalic acid and 
pure water, the latter having been previously treated with 
a drop of rosolic acid, will, on the addition of sea-water, 
immediately assume the well known reddish-violet hue. 
In this manner were examined on the last cruise of 
the Expedition, which afforded excellent opportunities of 
obtaining the water quite fresh, a very large number of 
samples, and invariably with the results described above. 
1 Ber. Berl. chem. Ges. 11, p. 410. 
8 Ann. Chem. Pharm. 77, p. 00. 
8 Compt. rend. *3, p. 010. 
* Journ. fiir pract. Chem. 107, p. 488. 
1 * 
