No. 6.| INTRODUCTION. VOYAGE ALONG THE COAST OF SIBERIA. LV 
The compass is, however, not a very trustworthy instrument in these 
high latitudes. 
the earth’s magnetism the local influence on board, as well as its variations, 
Owing to the feeble intensity of the horizontal component of 
attain relatively greater importance than in lower latitudes. Between the de- 
parture from Vardé 1893 July 21 and the enclosure in the ice on September 
22 Mr. Scorr-Hansen took in all 65 compass-bearings of the Sun or a star, 
giving the sum of magnetic declination and local deviation, and 4 direct deter- 
minations of the deviation by mutual settings between the compass on board 
and another compass placed at a convenient distance ashore or on the ice. 
In order to separate the declination and deviation it was necessary to examine 
the declination first. Professor Neumayer’s isogonic chart for 1895 extends 
to 75° of latitude, but by means of three determinations made by Mr. Scott- 
Hansen during the voyage, and a good many taken during the following years 
on the ice, it was possible to continue the curves and join the separated 
branches on a polar map. 
An inspection of the values of the deviation thus found showed that it 
could not be considered as constant for a given course during the whole 
voyage. On putting the deviation in the usual form 
A+Bsina + C cos a + D sin 2a + E cos Ye 
where @ is the compass-course from north through east, the constants were 
determined separately for the following three periods, containing respectively 
20, 22 and 26 observations (one of the 69 being omitted) taken between the 
limits given in the table below. 
Limits of 
Per. 
Date 1893 Lat. Long. Decl. 
ee OO eae iis 69°37’ 41°55’ E 10° E 
PNT ay ate 1a to 1G 71 20 66 44 20 
I eee i cameos 72 14 68 25 93 
; IN Wsho 6 0 0 76 54 95 2 29 
WW eek Mpa d avs 73 52 100 40 28 
; Sepie 2anetaney 78 50 137 8 7 
On solving the equations by the method of least squares the observations 
of period II proved insufficient to determine the quadrantal deviation, most 
of the observations having been taken in the first and the adjoining part of 
