Opvarmning trækker Kviksølvet sig tilbage 1 Røret; og 
Efter 
en saadan Operation gjør man vel i at sammenligne In- 
man bringer med Magneten Indexen ned 1 dette. 
strumentet med Normalthermometret. 
En anden Ulempe ved Index-Thermometrene er føl- 
gende. Naar et eller flere saadanne benyttedes, sammen 
med Kviksølvpiezometret eller Negretti & Zambra's Ven- 
dethermometer, til Bestemmelse af Bundtemperaturen, eller 
som Led til Bestemmelsen af en Temperaturrække, fandt 
jeg undertiden, efterat have anbragt de behorige Correc- 
tioner ved de aflæste Temperaturer, at et af Indexther- 
mometrene angav en Temperatur, der var merkelig for- 
skjellig fra og højere end hvad de øvrige Thermometre 
angav. Naar jeg strax derpaa sammenlignede med Nor- 
malthermometret, fandt jeg, at vedkommende Thermometer 
havde forandret sin Correction og, naar den sidst fundne 
Correction anbragtes, var der i Regelen god Overensstem- 
melse med de andre Thermometre. De saaledes indtraadte 
Ændringer kunde beløbe sig til en hel og en halv Grad. 
Samtlige vore Indexthermometre, ældre og nyere, viste 
Det mest 
paafaldende var imidlertid, at den pludselig indtraadte For- 
saadanne Uregelmæssigheder en eller flere Gange. 
andring ikke varede mere end en, højst to Dage; efter 
Tid gjenvandt Thermometrene sine tidligere havte 
Correctioner. Det synes, som om Aarsagen til Fæno- 
menet nærmest skulde ligge deri, at den lille Udvidelse, som 
den 
Thermometerrøret har lige under Kuglen, under et større 
Vandtryk i Dybet er bleven sammentrykket og derved har 
bevirket en højere Stand (Aflæsning) af Kviksølvtoppen, 
men at denne Virkning har tabt sig i Løbet af den føl- 
gende Dag under almindeligt Lufttryk, idet Glasset har 
gjenvundet sin oprindelige Ligevægtsform. Det har ikke 
lykkets mig at paavise ydre Omstændigheder, som 1 det 
enkelte Tilfælde skulde kunne forklare denne Virkning. 
Man maa derfor stadig være paa sin Post og have Control 
ved andre Thermometre. Det var først i 1877, at jeg blev 
opmerksom paa dette Fænomen. Det tør nok hænde, at 
flere af de 1 1876 iagttagne Uregelmæssigheder havde en 
lignende Aarsag, navnlig ved Thermometer No. VII; men 
jeg har, ved at anvende skiftende Correctioner ved dette 
Instrument, dog ikke opnaaet nogen gjennemgaaende bedre 
Overensstemmelse end med faste Correctioner. 
Negretti & Zambra’s Vendethermometer, der er inde- 
sluttet 1 et ydre tilsmeltet Glasror, som optager Vandtryk- 
ket, saaledes at dette ikke virker paa selve Thermometret, 
benyttedes af os i 1878 med den Vendemekanisme, som 
Fabrikanterne da havde givet det, nemlig en Trækasse med 
en Rende, hvori en Del Blyhagel kunde rulle og saaledes 
holde Tyngdepunktet til den ene Side. Under Udfiringen 
holdtes ved Lodlinens Fart Thermometerkuglen nederst. Ved 
Havbunden eller, ved Temperaturrækker, i en bestemt Dybde, 
standsede Bevægelsen, Thermometret gaves Tid til at an- 
tage det omgivende Vands Temperatur, og derpaa haledes 
14 
a less height and greater diameter, leaving the index suffic- 
iently free to be extracted by the magnet. The subsequent 
heating causes the mercury to draw back within the tube, 
and the index is then brought down by means of the magnet 
into that part. After such an operation, it is advisable to 
compare the instrument with the Standard-Thermometer. 
Another drawback attaching to the index-thermometers 
is the following. When one or more such were used, 
either along with the mercury-piezometer or Negretti and 
Zambra’s inverting-thermometer, for measuring the bottom- 
temperature or as a link for determining serial tempera- 
tures, I sometimes found, after having applied the neces- 
sary corrections to the temperatures read off, that one of 
the index-thermometers indicated a temperature remarkably 
different from, and higher than, that mdicated by the other 
thermometers. When, immediately after, I made a com- 
parison with the Standard-Thermometer, I found the ther- 
mometer in question to have changed its correction, and, 
on applying the new correction, there was, as a rule, very 
The changes 
thus occasioned half a 
All our index-thermometers, both old and new, 
fair agreement with the other thermometers. 
might amount to a whole or 
degree, 
gave evidence of the ‘same irregularity on one or several 
occasions. The most striking instance however, was that 
of this sudden change not having lasted more than one — 
at most two days; after that period the thermometers 
recovered their original corrections. It would seem as 
though the cause of the phenomenon lay principally in the 
expansion which the bore of the thermometer has directly 
under the bulb way exposed in the 
deep to a greater pressure, thus occasioning a higher 
reading of the top of the mercury, but in such influence 
ceasing to operate during the course of the following 
day under ordinary atmospheric pressure, the glass haying 
recovered’ its original form for equilibrium. I have not 
succeeded in finding the outward circumstances which in each 
case might explain this influence. Hence the observer must be 
constantly on his guard, and control the readings by means 
of other thermometers. It was not till 1877 this phe- 
nomenon attracted my attention. Very possibly, several 
of the irregularities observed in 1876 had a similar cause, 
in particular as regards thermometer No. “VII; but I 
have nevertheless failed, by applying an interchange of 
corrections to this instrument, to attain better general 
giving when 
agreement than with unchanged corrections. 
Negretti and Zambra’s inverting-thermometer, enclosed 
in an outer sealed glass tube, which resists the pres- 
sure of the water, that it does not 
act directly on the instrument, we used in. 1878, with 
the inverting-mechanism which the maker had then de- 
vised, viz., å wooden frame containing a quantity of shot, 
end to end, and which kept the 
When veering out the 
in such manner 
free to move from 
centre of gravity on one side. 
instrument, the sounding-line pulls the bulb of the ther- 
mometer downmost. On the sea-bed, or, when taking 
serial temperatures, at a given depth, the motion ceased, 
the thermometer was given time to accomodate, and then 
