VOYAGES IN THE POLAR SEA 



west that are mentioned, and no intermediate points; for 

 one course alone, from Bergen to Hvarf in Greenland, the 

 direction "due west" is given, which must be true west.^ 

 Langanes is said to lie on the north side of Iceland instead of 

 on the north-east, from Reykjanes to Ireland the course was 

 south, instead of south-east, etc. The points of the compass 

 are evidently used in the same way as is still common in 

 Norway ; " in the north of the valley " may be used even if 

 the valley bends almost to the west. The " Landnama's " state- 

 ment [Sturlubok] that it is four " doegrs' sea " from Snaefellsnes 

 " west " to Greenland (i.e., Hvarf) then agrees entirely with 

 the common mode of expression that I have found among 

 the arctic sailors of our day in Denmark Strait, where they 

 never talk of anything but sailing east or west along the edge 

 of the ice, even though it is north-east and south-west; we 

 sail westward from Faerder to Christiansand, or we travel 

 south from Christiania to Christiansand. Consequently " on 

 the north in Hafsbotn" means the same as when we say 

 north in Finmark (cf. Ottar's directions. Vol. I, p. 171), or even 

 north in the White Sea, and speak of sailing north to Jan 



1 True north of Langanes there is no land: Jan Mayen lies nearest, N.N.E., 

 and Greenland W.N.W. As the " leidar-stein " (compass) was known in Ice- 

 land when Hauk's Landnamabok was written (cf. Vol. I, p. 248), magnetic di- 

 rections might be meant here, and the variation of the compass may at that 

 time have been great enough to make Greenland lie north (magnetic) of Lan- 

 ganes. In that case it is perhaps strange that Langanes should be mentioned 

 as the starting-point, and not some place that lay nearer; but it might be sup- 

 posed that this was because one had first to sail far to the east to avoid the 

 ice, when making for the northern east coast of Greenland. A large eastern 

 variation would also agree with Jolldulaup in Ireland lying south of Reykjanes, 

 the uninhabited parts of Greenland lying north of Kolbeinsey (Mevenklint, see 

 Vol. I, p. 286), and the statement in the Sturlubok that from Snsfellsnes it was 

 "four 'doegrs" sea west to Greenland" (i.e., Hvarf). But it does not agree 

 with this that from Bergen (or Henno) the course was " due west " to Hvarf 

 in Greenland; and still less does it agree with its being, according to the Stur- 

 lubok, " seven ' dcEgrs' ' sail west from Stad in Norway to Horn in East Ice- 

 land." If these are courses by compass, we must then suppose a large eastern 

 variation between Norway and Iceland, which indeed is not impossible, but 

 which will not accord with a large western variation between Reykjanes and 

 Ireland. The probability is, therefore, that magnetic courses are not intended. 



169 



