kelte Thermometres Angivelser og dette Middel, faar man 
den følgende Tabel, der viser Gjennemsnits-Afvigelsen for 
hvert enkelt Thermometer: 
No. I. + 00.041 
IE 44 
IO, 55 
IV, IO 
VI. AA 
VIII, 15 
IDX, 15 
B. 43 
46. 36 
48. 33 
49. Be 
89. 2 
Ne O .037 
and compute the deviation between the indications of the 
several thermometers and that mean, we get the following 
Table, which shows the average deviations for each ther- 
mometer: 
31 Observationer. 
12 
56 
4 
Middel (Mean) + 09.040; 
Efter denne Sammenstilling bliver saaledes Middel- 
fejlen af en enkelt Temperaturbestemmelse i Dybet under 
00.05 1 1878. Da saavel i 1876 som 1 1877 
Omhu er anvendt paa Aflæsningen af Thermometrene, idet 
denne altid foretoges af 2 Iagttagere, der confererede sine 
Noteringer, tor (Gjennemsnit Dybtemperaturer 
være sikkre paa en Tiendedel Grad Celsius. En lignende 
Beregning som for 1878 giver for 1877 Middelafvigelse fra 
Medium af 3 Observationer med 3 forskjellige Thermometre 
+ 09034 og for 1876 fra Medium af 2 Observationer 
sr OLOS3 
Ovenfor (Side 11) er paavist, at Buchanans Kviksølv- 
plezometer registrerer Temperaturen i Dybet rigtig, selv 
om denne skulde stige lidt med Dybden. 
den samme 
1 vore 
De øvrige Index- 
thermometre derimod registrere rigtigt med Minimumsindexen 
kun under den Betingelse, at Temperaturen aftager med 
Dybden i et sterkere Forhold end Trykkets Virkning til 
at skyde Indexen At dette 
er Tilfældet i de større Dybder godtgjøres af følgende Ta- 
bel, i hvilken jeg har sammenstillet alle de Observationer, 
der kunne tjene til en nærmere Bestemmelse af Tempera- 
turens Variation med Dybden i de dybere Lag. 
nedover voxer med Dybden. 
170 Obs. 
According to this comparison, the mean error of a 
single determination of temperature in the deep was less 
than 09.05 in 1878. As equal care was invariably taken 
both in 1876 and 1877 with the reading of the thermome- 
ters, this having in every case been done by 2 observers, 
who compared notes, our deep-sea temperatures may on 
an average be taken as correct within one-tenth of a degree 
centigrade. A computation similar to that for 1878 gives 
for 1877 the mean deviation from the mean of 3 observa- 
3 different thermometers + 09.034, and for 1876 
from the mean of 2 observations + 0°.033. 
It has been shown above (p. 11), that Buchanan’s 
Mercury-Piezometer registers the temperature accurately in 
the deep strata, even should it rise slowly with the depth. 
tions with 
The other index-thermometers, however, register correctly 
with the minimum-index solely in the event of the tempe- 
rature diminishing with the depth in a proportion greater 
than that in which the force of the pressure exerted in 
pushing the index, increases with the depth. That such 
is the case in the great depths will be seen from the fol- 
lowing Table, where I have set down all the observa- 
tions that can serve for more closely determining the varia- 
tion of temperature with depth in the deeper strata. 
| Lodskud | Dybde Dybde- Temp.- Temp.-Var. | Trykcorr. 
Aar. | No. Haynes therm: Tempe- | Forskjel. Forskjel. pr. 1000 Fv.) pr. 1000 Fv. Diff. 
(Year.) (Sounding, | (Depth.) | No. ratur. (Dif. (Dif. (Var. of Temp. (Pressure-corr. | 
No.) (Fathoms.) | | of Depth.)| of Temp.) | for 1000 Fms.)| for LOOU Fms.) 
1876 40 1029 | V — 10.03 Bø — 09.17 væ 09.91 — 00.11 — 07.80 
1215 | V — 1 .20 
51 JER | III —0.04 | 205 0016 —0.78 | —0.07 — 071 
26 | IV — 1 .1O | | 
51 926 IV — 1.10 237 | —0.04 | —0.18 | —o.11 — 0.07 
| | 1163 W — I .14 | | 
