rimelfgst bliver at tilskrive ObseVvationsfeil. Som Chlor- 
coefficient kan lieraf opstillcs 
1.809 + 0,00076 
med en sandsynlig Feil af en enkelt Bestemmelse al + 0.002 
og som Egenvægtscoefficient 
131.9 4 - 0:058 
and. hence the variation in the results should most prob- 
ably be ascribed to errors of observation. The .coefficient 
of chlorine may accordingly be taken at — 
1.809 + 0.00076 
with a probable error in a single determination of +. 0.002, 
and the coefficient of specific gravity, at — 
131.9 + 0.058 
med en sandsynlig Feil af en enkelt Bestemmelse.al + 0.15. 
Disse Værdier stemme især for Chlorcoefficientens 
Vedkommende ganske vel overens med tidligere fundne 
Værdier, saaledes' har baade Forchhammer og Ekman i 
Middel fundet 1.811. mcdéns de af Andre opstiUede Egen- 
vægtscoefficienter overalt ere noget mindre end den af mig 
fundne.. 
Ved Hjælp af disse Coefficienter har jeg af de paa 
Expeditionens Togter udførte Chlor- og Egenvægtsbestem- 
melser beregnet Vandprøvernes Saltgehalt og sammen med 
Originalobservationerne opført dem i nedenstaaende Ta bel. 
Egenvægterne ere i Regelen kun aflæste med 4 Deci- 
maler, det 5te er kun opført, hvor det havde en Værdi nær 
5, saa at det kunde være Tvivl underkastet, om der ved 
Afrunding skulde formindskes eller forhøies. I de reduce- 
rede Egenvægter findes ligeledes kun opført 4 Decimaler, 
hvor det uden Tvivl kunde algjøres, til hvilken Side 
Afrundingfen skulde finde Sted, hvorimod der i modsat 
Fald ogsaa der er tilføiet et 5te. De mod * betegnede 
Egenvægter ere aflæste paa Aræometre, hvis Correction 
ikke er bleven bestemt. Til Optagelse af de til Under- 
søgelse af Saltholdigheden bestemte Vandprøver er foruden 
det tidligere beskrevne, af Wille construerede, Apparat og- 
saa paa grundere Vand ofte benyttet den af Ekman an- 
givne fortrinlige Vandhenter, som imidlertid ifølge den Frem- 
gangsmaade, hvorefter Dyblodninger paa den norske Ex- 
pedition foretoges, ikke egnede sig til Brug ved større Dyb. 
Ved Velvillie af Professor Mohn har jeg faaet op- 
givet de undersøgté Vandprøvers Temperatur i Havet, hvor- 
ved det er bleven muligt ogsaa at tilløie en Rubrik for 
deres Egenvægter ved denne Temperatur i Forhold til 
rent Vand af 4°. 1 Af de i Tabellen gjengivne Observa- 
tioner ere . alle indtil No. 149 udførte paa . Iste Togt af 
Svendsen,, alle fra 149 til 225 paa 2det Togt af mig, de 
øvrige ere udførte paa sidste Togt af Schiuelck og mig i 
Fællesskab, .saaledes at det største Antal skyldes Sclmielck. 
der dette Apr medfulgte Expeditionen. 
1 Rent Vand af 4° er ved denno Reduction valgt, som Enhed, 
fordi «len allerede tidligere er anvendt af .J. V. Huchanan (Proc. 
Roy. Soc. 2-1 Ved Beregning af Egcnvægternes Værdi 
ved Havets Temperatur i Forhold til Vand af 4° er Forhol- 
det Jncllem Volumet af rent Vand ved 4° og 17. 0 .‘> sat = 0.098708. 
with a probable error in a single determination of +0.15. 
These values agree closely, in particular as regards 
the coefficient of chlorine, with those previously found. 
Thus, for instance, the mean value found both by Forcli- 
hanmier and Ekman was 1.811. whereas the coefficient of 
specific gravity given by . former observers is somewhat 
lower than mine. 
By means of these coefficients I have computed from 
[ the determinations of chlorine and specific gravity the 
amount of salt in the samples of water collected on the 
Expedition, and have set down the observations and their 
results in the accompanying Table, which calls for a brief 
explanation. 
The specific gravities are as a rule not read to more 
than 4 decimals, a fifth being added only in the event of 
its having a value of 5. in which case it is often doubtful 
whether, when rounding off the figures, there should be 
increase or diminishment. The reduced specific gravities, too. 
are expresed with 4 decimals only, wherever it was mani- 
fest in which direction the rounding off had to be made; 
when such is not the case, a fifth has been added. An 
asterisk at the side of a specific gravity denotes that the 
latter was determined with an areometer • for which no 
correction had been found. Besides the instrument devised 
by Wille, of which a description has been given, Ekman's 
excellent apparatus was likewise made use of, in compara- 
tively shallow localities, for collecting the samples of sea- 
water in which to determine the amount of salt: the mode 
of sounding practised on the Norwegian North-Atlautic 
Expedition would not admit of its adoption for greater depths. 
Professor Mohn has kindly furnished me with the 
temperatures of the samples of water in situ, which lias 
enabled me to give an aditional column for the specific 
gravities at those temperatures as compared with pure 
water of 4". 1 Of the observations given in the Table, 
those extending from No. 1 to No. 149 were performed 
on the first voyage, by Mr. Svendsen; those extending from 
No. 149 to No. 225, on the second voyage, by myself; the 
remainder were taken on the last voyage, by Mr. Schmelck 
and myself conjointly, the greater number, however, by Mr. 
i Sclpnelck, who that year accompanied the Expedition. 
1 Pure water of 4° was chosen as the unit of reduction, J. Y. Bu- 
chanan having previously adopted ‘it as such (Proc. of Roy. 
Sne. 24, ]). When computing the specifio gravities at the 
temperature of the sea, as compared with water of 4°, the 
ratio existing between the volume of pure water at 4° and 17.°.'» 
was assumed to be 0.99870S. 
