Jan Mayen. For 1878 Aars Vedkommende maatte Grup- 
peringen tildels blive noget forskjellig for de forskjellige 
Elementer, idet der i enkelte af de forsøgte Grupper ikke 
viste sig Spor af nogen rimelig Periode. De Hensyn, som 
ligge til Grund for de valgte Grupper, skulle paa sit Sted 
blive forklarede. 
I hver af de valgte Grupper er der kun medtaget 
bele Observationsdage, det vil sige saadanne, paa hvilke 
der er blevet observeret alle 24 Timer i Døgnet, regnet 
fra Midnat til Midnat. 
Førend de valgte Gruppers Middeltal for de enkelte 
Klokkeslet ere anvendte til Beregningen af Formlerne for 
de daglige Perioder, ere de blevne corrigerede for Virk- 
ningen af de uperiodiske Variationer og den aarlige For- 
andring samt Stedforandring af Observationspunktet (Skibet). 
Da hver Gruppe i Regelen udgjøres af flere mindre, adskilte 
fra hinanden ved de Dage, Expeditionen laa i Havn, ere disse 
Correctioner beregnede først særskilt for hver af de mindre 
sammenhængende Afdelinger, og deres Medium, Hensyn taget 
til Dagenes Antal, anbragt ved de større Gruppers Medier. 
Med Hensyn til Methodens Anvendelse henvises til H. Wild, 
Die Temperaturverhåltnisse des Russischen Reiches." Erste 
Halfte. 
givne Oplysninger. 
Side 9, og til de under hver Gruppe nedenfor 
Som ovenfor i denne Afhandling (Side 46) nævnt, ud- 
førtes de timevise Observationer efter Skibsuret, der stilledes 
paa 12 hver Middag sand Tid. Ved Skibets Sejlads vest- 
over eller østover svarede saaledes Observationstiderne ikke 
ganske til sand Tid paa Skibets paaværende Plads. Af- 
vigelserne, der voxe med Tiden, naaede sin største Værdi 
den næste Dags Middag, forudsat, som Tilfældet uden 
mange Undtagelser — Se Kartet over Vøringens Route — 
var, at 1 Mellemtiden Længden var alene voxet eller aftaget. 
At indføre nogen Correction af de beregnede Middel- 
tal for Observationsklokkeslettene for at bringe dem i Over- 
ensstemmelse med den sande Tid paa Stedet kunde vistnok 
lade sig gjøre ved Interpolation mellem Værdierne for de 
enkelte Timer. Men en saadan Fremgangsmaade vilde neppe 
lønne Umagen. I ethvert Tilfælde bliver man afhængig af 
den Variation, som selve Skibets Stedforandring medfører. 
Betragter man de daglige Perioder, som resultere af vore 
Beregninger, saa ser man, at det mest fremtrædende Træk 
er Amplitudernes Lidenhed. De absolute Værdiers Varia- 
tion med Tiden indenfor den daglige Periode bliver saa- 
ledes kun ringe, og en mindre Fejl i Klokkeslettet har liden 
eller ingen praktisk Betydning ligeoverfor Observationens 
absolute Nøjagtighed. Hertil kommer endvidere, som jeg 
nedenfor for hver Gruppe skal vise, at Fejlen i Klokke- 
slettet for de enkelte Observationer compenseres næsten 
aldeles fuldstændigt ved den Qmstændighed, at Summen af 
1 
Supplementband zum Repertorium fiir Meteorologie. St. 
Petersburg ‘1877. 
e 
other those taken within the Greenland Polar current and 
during the stay of the Expedition at the Island of Jan 
Mayen. As regards 1878, this grouping had to be varied 
for the different elements, some of the groups I tried to 
form not having exhibited even traces of an admissible 
period. The various reasons that affected me in forming 
the groups, will be explained in the sequel. 
For each of the groups selected, none but whole “ob- 
servation days” have been taken, that is to say, such on 
which we observed every hour of the four and twenty, from 
midnight to midnight. 
Before applymg the mean values of the groups for 
each individual hour to compute the formule for the diurnal 
periods, they were corrected for the effect of the non-periodie 
variation, as also for the annual change and the change in 
the place of observation (ship’s position). Each group 
comprising as a rule several smaller ones, separated one 
from the other by the interval passed by the Expedition 
in harbour, these corrections were first computed separately 
for each of the smaller continuous sections, and their mean —, 
regard being had to the number of days — applied to the 
means of the greater groups. With respect to the appli- 
cation of the method, the reader is referred to H. Wild. 
Die Temperaturverhiltnisse des Russischen Reiches,! Erste 
Hiilfte, p. 9, and likewise to the information given under 
each group in the sequel. 
As previously stated in this Memoir (p. 46), the hourly 
observations were taken by the ship’s clock, which every 
day at noon was set to apparent solar time. Hence, with 
the vessel steaming westward or eastward, the hours of ob- 
servation did not correspond exactly with apparent time on 
board. The difference, continually increasing, reached its 
maximum the next day at noon, provided, as was actually 
the case without many exceptions, — see Plate showing 
the track of the Vøringen, — that in the meantime the 
longitude had only increased or diminished. 
To introduce a correction of the computed means for 
the hours of observation, in order to make the latter corre- 
spond with the apparent time of the ship’s position, might 
indeed be effected by means of interpolation between the 
values for the different hours. But this would hardly be 
worth the trouble. In any case, the results would still be 
affected by the variation involved in the change of position 
of the ship herself. If we regard the diurnal periods resulting 
from our computations, the most salient feature is found 
to be the smallness of the amplitudes of variation. Hence, 
the hourly variation of the absolute values within the diurnal 
period will be but triflmg, and a slight error in the hour 
for the different observations is of little or no practical 
importance as compared with the absolute accuracy of the 
observations. And moreover, as will be shown below for 
each group, the error im the hour of the different ob- 
servations is almost wholly compensated by the sum of the 
1 Supplementband zum Repertorium fiir Meteorologie. St. 
Petersbure 1877. 
