De sandsynligste Værdier for Coefficienterne blive 
a=—+0.""467, b= — 0.11157, c=+0.06638, d=-+ 0.00292 
og altsaa Formelen: 
\ 
N=A- 0.47 — 0.1116 T+ 0.0664 (760 — A) + 0.0029¢. 
De efter denne Formel beregnede Verdier af NW — A 
ere opforte i ovenstaaende Tabel. Beregningen efter For- 
melen er gjort ved Hjelp af en grafisk Tabel, paa hvilken 
der med Argument Å langs de horizontale Linier og Argu- 
ment 7 langs de verticale, ere optrukne Linier, skraat- 
løbende, for Værdierne af a+ b T + c (760 — A) for hver 
Tiendedel Millimeter fra + 1.”4 til — 3.76. Af denne 
Tabel kan saaledes den samlede Sum af disse Correctioner 
udtages, og hertil lægges Correctionen dt, der er regnet 
lig O fra 15de til og med 30te Juni, lig +0.””1 fra Iste 
Juli til 5te August, og lig + 0.772 fra 6te August til 16de 
Denne sidste Correction udtrykker Forandringen 
under 
August. 
af Aneroidbarometrets constante Correction, der, 
Rejserne i 1877, ikke naaer op til 0.”"2. Vi have nemlig 
0.00292 x 62 =0.18. Af Rubrikken Obs. — Ber. 1 ovenstaa- 
ende Tabel beregnes Middelfejlen af en Sammenligning 
mellem Søbarometret og Aneroidbarometret, begge reduce- 
rede til sand Barometerhøjde efter de angivne Methoder, 
till > OPPLAG 
Den midlere Fejl bliver ¢ = + 0.””256. 
Den sandsynlige Fejl , 0=+ 0.77170. 
Den sandsynlige Fejl af en Sammenligning mellem 
Sobarometret Adie 1493 og det meteorologiske Instituts 
Normalbarometer er fundet lig + 0.7063. Herefter skulde 
den sandsynlige Fejl af en reduceret Observation af Ane- 
=+0.7"158. Jeg tror, at man maa være tilfreds med at 
kunne opnaa denne Nøjagtighed af Barometerobservationer 
tilsøs. 
opnaa under gjennemsnitlige Vejrforhold med et Søbarometer, 
da dette enten er trægt, eller, 1 uroligt Vejr, pumper. 
For at undersøge, om der mellem lIagttagernes (Styr- 
mændenes) Aflæsninger af Aneroidbarometret og mine Af- 
læsninger er nogen constant Forskjel eller personlig Lig- 
ning, har jeg gaaet frem paa følgende Maade. De Obser- 
vationer af Aneroidbarometret, jeg har noteret i det oven- 
beskrevne Øjemed, — for at controllere Aneroidbarometret — 
paa de Dage, der gjordes Observationer ombord af Styrmæn- 
dene, har jeg reduceret paa ovenangivne Maade og sam- 
menstillet med Styrmændenes, i de meteorologiske Tabeller 
nedenfor for hver Time givne. Efter disse timevise Ob- 
servationer var det ikke vanskeligt at interpolere Styrmæn- 
denes Observation til det Klokkeslet, da jeg tog min Ob- 
servation. Resultatet af 40 saadanne Sammenstillinger er, 
at Styrmændene i Gjennemsnit have noteret Aneroidbaro- 
metret 0.””045 lavere end jeg. Altsaa en umerkelig For- 
skjel. Middeltallet af Afvigelse mellem Styrmændenes og 
mine Aflæsninger er +0.””15, en Størrelse, der er næsten 
ganske den samme som den ovenfor fundne sandsynlige Fejl 
Den er visselig lige saa stor som den, man vilde 
36 
The most probable values for the coeffleients, are 
a= + 0.7"467, b= — 0.11157, ¢ = + 0.06638, d=0.00292; 
and hence we have the formula: — 
N=A - 0.747 — 0.1116 T’+ 0.0664 (760 — A) + 0.0029. 
The values of 1 — A computed from this formula are 
set forth in the above Table. The computation according 
to the said formula was made by means of a diagrammatic 
table, which, with the argument A along the horizontal 
lines, and the argument T along the vertical, gives a series 
of lines — oblique — representing the values of a+ b T+ 
c (760 — A) for every tenth of a millimetre, from + 1.””"4 
to —3.””6. Hence, from this Table the whole sum of the 
said corrections can be found, to which must be added the 
correction dt, put =0 from the 15th to (and including) the 
30th of June; =+ 0.""; from the Ist of July to the 5th 
of August; and = +0.””2 from the 6th to the 16th of 
August. The latter correction expresses the change in the 
constant correction of the Aneroid barometer, which, on 
the cruises in 1877, did not reach 0.””2. For we have 
0.00292 <62=0.18. From the Column “Obs. — Comp.” in 
the above Table, the mean error of a comparison between the 
marine barometer and the Aneroid barometer, both reduced 
to true barometric pressure by the methods previously 
described, is computed at + 0.”"177. 
The mean error becomes ¢ = + 0.”"256. 
OQ S OL? 1/7, 
The probable error of a comparison between the 
marine barometer Adie No. 1493 and the standard baro- 
meter of the Meteorological Institute, has been found = 
+ 0.°"068. Hence, the probable error of a reduced observa- 
tion of the Aneroid barometer on board, should be about 
+ V (0.170)? — (0.063)? = + 0.158. This degree of ac- 
curacy may, I think, be regarded as satisfactory. It is, at 
least, quite equal to that attamable during average weather 
with a marine barometer, which is anything but sensitive, 
or, in boisterous weather, is given to pump. 
With a view to investigate whether a constant dit- 
ference, or personal equation, actually exists between the 
observers’ (the mates) readings of the Aneroid barometer 
and my own, I proceeded in the following manner: — The 
observations of the Aneroid barometer — taken with the 
object specified above, viz. to control that instrument — 
which I have noted down on the days when observations 
were made on board by the mates, I have reduced, as 
described above, and compared with those of the mates, 
entered in the Meteorological Tables given below for every 
hour. From these hourly observations, it was not difficult 
to interpolate the observations of the mates for the hour 
when I took my observation. The result of 40 such com- 
parisons is, that the mates, on an average, have noted the 
readings of the Aneroid barometer 0.””045 lower than I 
have done — a well-nigh imappreciable difference. The 
mean difference between the indications read off by the 
mates and by myself, is + 0.”"15, a value almost exactly 
The probable error ,, 
