Survey of Northeast Greenland. 97 
greasing the cut glass rim and replacing the cover, before the bowl 
was taken back to the heated room. By putting a hygroscopic sub- 
stance e. g. chloride of calcium into the glass bowl one might pre- 
sumably also have prevented the formation of rime, but it was not 
possible to make experiments in this direction on the expedition. 
The watches were made to run for 30 to 36 hours at a stretch. 
Both on sledge trips and at home they were from prudential motives 
wound twice during the twenty-four hours. 
The Karrusel watches were good; in particular it was surprising 
how well they withstood the rather rough treatment оп sledge 
trips. 
As observation barometer we used the standard barometer of 
the expedition, a mercurial barometer which was hung in the “Villa”, 
and the cup of which was 7.6m lower than the telescope. 
The reading of the barometer was to be corrected by means of 
The index error of the barometer ......'.. — 0.5 mm 
Correction for the height of the telescope above 
О ое на dus D para + 0.7 — 
Reduction do 455 latitude." . dn ies os + 1.8 — 
Sum total of constant corrections + 0.6 mm 
To measure the temperature a swinging thermometer was used, 
the correction of which was zero. As the use of the swinging ther- 
mometer requires space, one must go outside the Observatory when- 
ever it has to be swung, and consequently one could not measure 
the temperature at the objective of the telescope, as one strictly 
speaking ought to do. Such a proceeding is by no means permissible 
in the case of observations requiring due consideration of the re- 
fraction, inasmuch as there may be such a great difference between 
the temperature inside the observatory and outside it, that it comes 
to play even a rather great part in proportion to the accuracy of the 
observation. Even if the 0.7m wide meridian aperture of the Ob- 
servatory were open throughout its length from south to north, the 
difference of temperature between the air inside and outside the 
Observatory at noon, when the sun was visible might, however, rise 
to a couple of degrees; at night and when electric light was used 
the difference was on the contrary imperceptible. Nocturnal experi- 
ments of this kind are, however, not made at a temperature below 
- 15° (Centigrade). At a still lower temperature there will presumably 
also at night turn out to be a difference, because the observer and 
his assistant will then proportionally give off more heat to the sur- 
rounding air. 
7% 
