55 
Correctionernes Bestemmelse -udførtes ved Hjælp af 1 
Yandprøverne I og YII, hvis Egenvægter tidligere ere be- 
stemte til 1.02691 og 1.02669 ved ^5 , For det paa 
tørste Togt anvendte Araeometer erholdtes sa«a)edes gjennem 
5 Aflæsninger i I Oorfectionen —0.00023 og gjennem 12 
Aflæsninger i VII ligeledes —0.00023. Paa samme Maade 
bestemtes det andet Aræometérs Gorrectioner ved 5 Aflæs- 
ningei- i I til —0.00037 og ved 8 Aflæsninger i VII til 
— 0.00038. Under disse Aflæsninger var Vædsken altid 
bragt til 17. "5 eller en meget nærliggende Temperatur, 
hvorfra Aflæsningerne efter de forheu gjengivne Correctio- 
ner reduceredes til Normaltemperatureu. Gjennem flere 
Rækker Aflæsninger i Vandprøyen 1 ved forskjellige Tem- 
peraturer har jeg tillige forvisset mig om, at den ved Be- I 
.regning af Correctionstabellén benyttede Udvidelsescoefficient 
for Aræometrene er passende valgt. 
Hermed er givet de fornødne Data til Reduction af 
de paa den norske Nordhavsexpedition aflæste Egenvægter, 
og jeg gaar dernæst over til Bestemmelsen af Relationerne 
mellem Saltgehalten, Chlormængden og Egenvægten. 
Til Bestemmelse af Saltmængden har, saavidt jeg ved, 
tidligere kuu været benyttet den simpleste Methode, bestaa- 
ende i Afdamphing af Våndet og Residuets Tørring ved en 
passende Temperatur, som af de forskjellige Chemikere er 
bleven valgt noget jorskjelligt fra 150° — 180°. Denne 
Methode har jeg imidlertid af flere Grunde fiindet lidet I 
tilfredsstillende, livad man ogsaa paa Forhaand maatte vente. 
Efter Graham 1 og andre taber nemlig den svovlsure Mag- 
nesia. om hvis Tilstedeværelse i Søvandet der vel ikke kan 
reises Tvivl. først ved en Temperatur* af over 200° sit 
sidste Molekyl Vand, medens man paa den anden Side 
allerede ved en Temperatur af betydeligt under 200° maa 
befrygte en delvis Decomposition af den i Saltene tilstede- 
værende Chlormagnesium. Efter de Forsøg, som jeg an- 
stillede, viste det sig, at Saltene selv efterat være tørrede 
ca. 20 Timer i Luftbad ved en Temperatur fra 170° — 
180° endnu indeholdt ikke ubetydelige Mængder Vand (om- 
kring 15 Mgr. pr. Gr. Salt), medens de tørrede ved lidt 
lavere Temperatur indeholdt noget mere. Samtidig under- 
søgtes ogsaa Saltene paa fri Magnesia, hvorved jeg i Strid 
med ældre Angivelser fandt, at de bestandig selv ved Tør- 
ring ved 160" til 170" indeholdt uventet store Qvantiteter, 
saaat der for hvert Gr. tørret Salt faudtes en Magnesia- 
mængde tilstrækkelig til at neutralisere over. 20 Mgr. HC1 (ved 
Tørring ved 180" fandt jeg .endog en enkelt Gang 40 Mgr.). 
Bestemmelsen af den frie Magnesia foretoges ved Saltenes 
Opløsning i en afmaalt Mængde titreret Svovlsyre og der- 
paa følgende Retitration med forty ndet Natron lud af be- 
kjendt Styrke. Ved Anvendelse af Rosolsyre som Index 
erholdtes her en meget skarp Endereaction. 
The determination of the corrections was performed 
with the water of samples I and VII. whose specific grav- 
* 17.°5 , , , . 
Ry at had been found to be respectively 1.02691 and 
1.02669. For the areometer used on the first voyage 5 
readings with the water of sample I gave the correction 
— 0.00023, and 12 readings with the water of sample YII 
likewise — 0.00023; in the same manner, the corrections 
for the other areometer were determined, by 5 readings 
with the water of sample l. to be — 6.00037, and, by 8 
readings with the water of sample VI 1, to be —0.00038. For 
these readings, the fluid was always brought to 17.*5, or 
as near that temperature as possible, the readings having 
in the. latter case to be reduced, by means of the correc- 
tions given above, to the normal temperature. Several 
series of readings with the water of sample I, at different 
temperatures, convinced me that the coefficient of expan- 
sion for the areometer which 1 had computed for prepar- 
ing the Table of Corrections was practically correct. 
Having now specified the data necessary for reducing 
the specific gravities read on the Norwegian North- Atlantic 
Expedition. I shall proceed to determine the relation be- 
tween the specific gravity of sea-water and the amount 
of salt and chlorine it contains. 
For determining the amount of salt, the only method 
formerly resorted to was, so far as I am aware, the sim- 
plest, viz. that of evaporating the water and then drying 
the residue at a proper temperature, which has been vari- 
ously fixed by different chemists at from 150" to 180°. 
This method, however, has proved in several respects de- 
fective, as was indeed to be expected. According to Gra- 
ham 1 and others, sulphate of magnesia, the presence of 
which in sea-water can hardly admit of doubt, does not 
part with its last molecule of water till exposed to a temp- 
erature of more than 200" whereas, on the other hand, it 
is highly probable that partial decomposition of the chloride 
of magnesium contained in the salt takes place consider- 
ably below 200°. Even after the salts had been dried for 
about 20 hours in an air-bath at a temperature of 170" — 
180", they were still found to contain, according to my 
experiments, a considerable quantity of water (about l5 M ' /r 
salt per gramme); dried at a lower temperature, the 
amount was somewhat greater. L also tested the salts for 
free magnesia, and found, in direct opposition to earlier 
statements, that, even when dried at 160°— 170", they in- 
variably contained a very large amount, the quantity of 
magnesia to every gramme of dried salt being sufficient to 
neutralize more than 20 m,/ ' MCI (once, when dried at 180", 
even 40 m ' ,r ). For determining the free magnesia, tin* salts 
were dissolved in a given quantity of titrated sulphuric 
acid, and the fluid then retitrated with dilute soda-lye of 
known strength. With rosolic acid as the index, the final 
reaction was very decided. 
1 Phil. Mag. .1. <i. p. 422. 
Phil. Mag. J. (i - 4 22. 
