Ganske anderledes stiller Forholdet sig, naar Tempe- 
raturen stiger raskt med Dybet. En vastrukken Træ- 
kasse kan da komme til at lade Thermometret registrere 
Et slaaende Exempel her- 
I 170 Favnes Dyb 
gav Kviksølvpiezometret en Temperatur af 2".40. Ved 
Bunden, 1 204 Favnes Dyb, altsaa kun 34 Favne dybere, 
gav Kviksølvpiezometret — 09.51 og Miller-Casella No. I 
— 09,83, begge reducerede for Skalafejl og for Tryk. Ven- 
ganske fejlagtige Temperaturer. 
paa havde vi paa Station No. 375. 
dethermometret Negretti & Zambra No. 89 registrerede 
+. 0,49, altsaa næsten en hel Grad for højt. Lignende 
Forhold vise Stationerne No. 327, 328 og 338. Allerede 
i de første Dage af vor Rejse 1 Østhavet blev jeg opmerk- 
som paa disse Uregelmæssigheder, og lod derfor altid et 
Indexthermometer følge med et Vendethermometer ved Lod- 
skuddene. Eftersom Trækasserne ved gjentagen Brug bleve 
vastrukne, lod jeg Tømmermanden gjøre nye Kasser, der 
benyttedes til de, mindre Dybder, saalænge de, efter an- 
stillet Forsog 1 Havvand, bevarede sin Flydeevne. De 
Det er ikke 
muligt ved Maling eller Fernissering at beskytte Trækasserne 
Heller 
ikke er det muligt at gjengive vastrukne Kasser denne ved at 
Paabundne Kork- 
stykker kunne hjelpe, men kun for en Tid. Sterke til- 
smeltede Glasrør, indsatte i Kassen, kunde raade Bod paa 
vastrukne Kasser brugtes kun paa større Dyb. 
saaledes, at de hindres fra at tabe sin Flydeevne. 
tørre dem og give dem noget Overdrag. 
disse Ulemper. Nu for Tiden vil man dog foretrække de 
senere 1 Brug komne Vendemekanismer, ved hvilke Ven- 
dingen besørges under Begyndelsen af Ophalingen af en 
Skruepropeller. og det vendte Thermometer ikke senere faar 
Dette sidste er af 
Vigtighed ved Temperaturrækker, naar man har flere Ther- 
mometre paa 
Anledning til at vende sig om igjen. 
Lodlinen 1 forskjellige Dybder, og der er 
nogen Segang. Under saadanne Omstendigheder kan man, 
naar man bruger Trækasser, frygte for, at et Vendethermo- 
meter, der allerede har registreret sin Temperatur, vender 
sig om igjen paa et højere Trim, idet Fartojet sænker sig 
i Søen. 
For at fmde den Tid, de forskjelligartede Dybvands- 
thermometre behøve for at antage det omgivende Vands 
Temperatur, gjorde jeg en Række Forsøg. Thermometret 
stilledes i Vand eller en Blanding af Sne og Vand, efter 
først at være bleven bragt paa en højere Temperatur. Et 
Normalthermometer angav Vandets Temperatur. Begge 
Thermometre stode rolige 1 Vandet, og dette rortes ikke 
om. Til visse Mellemrum aflæstes Dybvandsthermome- 
tret og samtidig dermed af en Assistent Tiden efter et 
Sekundur; umiddelbart ovenpaa aflæstes Normalthermome- 
tret. For hver Aflæsning beregnedes Forskjellen mellem 
begge Thermometres Udvisende. Hermed fortsattes, indtil 
at denne Forskjel var bleven constant. Med Argument: 
Aflæste Tidsminuter og Sekunder og Ordinater: Bereg- 
net Forskjel mellem Thermometeraflæsningerne opconstru- 
16 
The case is very different when the temperature in- 
creases rapidly with the depth. A water-soaked wooden 
frame can then cause the thermometer to register utterly 
erroneous temperatures. We had a striking example of 
this fact at Station No. 375. At a depth of 170 fathoms 
the mercury-piezometer indicated a temperature of 2°.40. 
At the bottom, in a depth of 204 fathoms, accordingly 
not more than 34 fathoms deeper, the mereury-piezometer 
indicated —O°.51 and the Miller-Casella No. I —0°.33, 
after reducing both instruments for error of scale and 
for pressure. The inverting-thermometer, Negretti and Zam- 
bra No. 89, registered + 0°.49, or almost a whole degree 
too high. ‘The same phenomenon was observed at Stations 
327, 328, and 338. As early as the first days of our 
cruise in the Barents Sea, these irregularities attracted my 
attention, and therefore I had an index-thermometer inva- 
riably sent down with Negretti & Zambra’s instrument. 
No sooner had the wooden cases got soaked through with 
water than I ordered the carpenter to make new ones, 
which were used for the minor depths, so long as, after 
careful testing im sea-water, they had been found to retain 
their buoyancy. The water-soaked cases were made use 
of for greater depths only. It is not possible by means of 
paint or varnish to protect the wooden cases, and thus pre- 
vent them from losing their buoyancy. 
to 
or 
Nor is it possible 
restore buoyancy to the water-soaked cases by drying 
giving them a covering of some kind or other. Pieces 
Strong glass 
tubes, sealed at the ends, may serve, when put in the 
cases, to counteract these drawbacks. 
of cork are of use, but only for a time. 
At present, however, 
preference will generally be given some one of the va- 
rious kinds of inverting-mechanism, recently devised, which 
cause the instrument to turn over on hauling in the line, 
by means of a propelling screw, and then prevent it from 
turning back agam. This is a matter of importance 
with series of temperatures, if several thermometers are 
attached to the sounding-line at different depths and a 
Under such circumstances there is 
danger, with the wooden cases, of an inverting-thermome- 
ter that has already registered the temperature turning 
sea 18 running. 
over again, on a higher level, when the vessel dips in 
the sea. 
To find how long a time the various kinds of deep- 
sea thermometers require to take the temperature of the 
surrounding water, I instituted a series of experiments. The 
thermometer to be tested was placed either in water or a 
mixture of snow and water, after having first been given 
A standard-thermometer indicated 
Both 
in 
a higher temperature. 
thermometers were 
the flud, which I 
the temperature of the water. 
allowed to remain undisturbed 
did not stir. At given intervals the deep-sea thermom- 
eter was read off, an assistant noting simultaneously 
the time by a watch, and immediately after the stand- 
For every reading, the difference was 
of the two thermom- 
said differ- 
ard-thermometer. 
computed between the indications 
This operation I continued till the 
With argument: minutes 
eters. 
ence proved to be constant. 
