For Thermometer No. I have vi N C = + 1.59, 
id, I = OI x PAS JE 259 og Sv SE Welsh, Ino 
Oi 
Correctionsformelen for No. I for Skala bliver saa- 
ledes 
C = + 00.105 — 0.0072 t. 
De efter denne Formel beregnede Værdier af Cor- 
rectionerné ere opførte 1 ovenstaaende Tabel under Ru- 
brikken B. I den sidste Rubrik staar Forskjellen mellem 
den observerede og den beregnede Correction : C—B. 
Summen af C—B er, uden Hensyn paa Fortegn, 1.27, 
og den gjennemsnitlige Værdi, 1.27: 24, er + 00.058. 
De Miller-Casella’ske Dybvandsthermometre ere con- 
struerede: saaledes, at Kuglen er beskyttet mod Sammen- 
trykning ved Vandets Tryk. Men Thermometerrøret er 
udsat for Trykket, der gjør dets Kaliber mindre og driver 
Kviksølvtraaden bort fra Kuglen, hvorved Aflæsningen giver 
en for hø) Temperatur. Paa Minimumsiden, hvor Indexen 
er nærmest Kuglen, bliver denne Virkning af Trykket 
ringe, men paa Maximumsiden bliver den merkelig stor, 
da Kviksølvtraaden, der kun er lidet sammentrykkelig, 
antager en forøget Længde 1 det indsnevrede Kaliber. De 
for Maximumsiden fra Casella opgivne Trykcorrectioner 
for vore Thermometre vare fra 00,6 til 10.1 for 2000 
Favnes Vandtryk. Ved vore Undersøgelser benyttedes kun 
Minimumsiden.  Correctionen for Trykket paa denne har 
jeg kunnet bestemme, i 1878, directe for hvert Indexther- 
mometer ved Sammenligninger af de paa større Dyb gjorte 
Registreringer med, hvad de Negretti-Zambra'ske Vende- 
thermometre udviste paa samme Sted og til samme Tid, 
idet de sidstes Registrermger 
Trykket. Ved Lodskuddene paa større Dybder lod jeg 
et eller flere Indexthermometre ledsage til Bunds af et 
Vendethermometer og erholdt derved en Række Sammen- 
ligninger til Bestemmelse af Indexthermometrenes Tryk- 
correction. Ved Beregningen af denne kan Correctionens 
Størrelse uden merkelig Fejl sættes proportional med Dyb- 
den. Den skal, som ovenfor paavist, være negativ. I den 
følgende Tabel -indeholder den første Rubrik Loddestatio- 
nens Nummer, den anden Rubrik Dybden i engelske Favne, 
den tredie Rubrik Aflæsningerne af Minimumsindexen paa 
Thermometer No. I, corrigerede for Skalaens Feil efter 
den ovenfor givne Formel, den fjerde Rubrik Aflæsningerne 
af et af Negretti & Zambras Vendethermometre No. 89 
eller No. 91, corrigerede for Skalafejl (se nedenfor), den 
femte Rubrik Forskjellen mellem/Tallene i fjerde og tredie 
Rubrik eller Indexthermometrets Trykcorrection. 
ere ganske uafhængige af 
For Thermometer No. I, we have 3 0 = + 1.59, 
md IN = O0b x 24 = I 22%, and Sw = 1988: henes 
ge a 0.0072. 
128.8 
The formula of correction tor No. I (scale) will 
accordingly be 
C = + 00.105 — 0.0072 t. 
The values computed for the corrections according 
to this formula have been given in the above Table, Col- 
umn B. In the last column, the difference between the 
observed and the computed correction is C —B. The 
sum of C —B amounts, without reference to signs, to 1.27, 
and the average value, 1.27: 24, is + 00.053. 
The Miller-Casella deep-sea thermometers are so con- 
structed as to have the bulb protected from the pressure 
of the water. But the tube of the instrument is exposed 
to such pressure, which reduces the bore and forces the 
thread of quicksilver away from the bulb, whereby the 
reading gives too high a temperature. On the minimum- 
side, where the index les nearest the bulb, the effect of 
this pressure is but shght; but on the maximum-side it 
gets remarkably strong, since the thread of mercury, which 
is but very slightly compressible, assumes an increased 
length in the contracted bore. The pressure-corrections 
given by Casella for our thermometers on the maximum- 
side, were as high as 0°6 to 1°91 for the pressure 
at a depth of 2000 fathoms. With our investigations, 
the minimum-side only was used. The pressure-correction 
of this side I was able to determine direct, in 1878, 
for each index-thermometer, by comparing the registrings 
at great depths with the results exhibited by the Negretti 
& Zambra inverting-thermometers in the same spot and 
at the same time, the registrations of the latter in- 
struments being wholly independent of pressure. With 
soundings undertaken at great depths, I sent down along 
with one or more of the index-thermometers an inverting- 
thermometer, and thus obtained a series of comparisons 
for determining the pressure-correction of the index-ther- 
mometers. When computing the latter, the correction may 
be put, without serious error, proportional to the depth. 
It should, as stated above, be negative. In the following 
Table, the first column the numbers of the 
Sounding-Stations, the second column, the depth in English 
fathoms, the third column, the reading of the minimum- 
index for Thermometer No. I, corrected for error of 
scale according to the formula given above, the fourth 
column, the reading of Negretti & Zambra's inverting- 
thermometer, No. 89 or 91, corrected for error of scale 
(see below), the fifth column, the difference between the 
figures in the fourth and third columns, or the pressure- 
correction of the index-thermometer. 
contains 
3* 
