LVI GEELMUYDEN. ASTRONOMICAL OBSERVATIONS. [NORW. POL. EXP. 
the fourth quadrant, but only 4 in the second and none in the third. On 
putting the constants D and E for this period equal to the mean of the values 
found for periods I and III, the following values were calculated and tables 
formed from them: 
A B C D E 
Ih —1.°72 —0.° 98 + 5.°68 +3.°46 —0.°10 
Il. +1. 16 —0, & +10. 40 +2. 20 —1. 13 
Ii. +3. 51 —3. 47 +11. 13 +0. 96 —2. 15 
Another difficulty in the determination of the situation of new islands 
etc. arose from the necessity of using dead reckoning. Owing to the difficult 
navigation, frequently hindered by fog or ice and often conducted between 
unknown banks and islands and in strong currents, the dead reckoning was 
often seriously in error, especially in longitude. The difference of east longi- 
tude found by dead reckoning being almost invariably too great, it was neces- 
sary to introduce a proportional reduction, which of course gives rise to some 
uncertainty when the interval from the nearest astronomical observation was 
considerable. Apart from some days spent in harbour there were, however, 
only few days without astronomical observations. 
On one occasion the difference of east longitude deduced from the astro- 
nomical observations was considerably in excess over that of the dead recko- 
ning, viz. on 1893 August 29 when the ship had encountered dead water in the 
afternoon and arrived in the evening at the ice border near Cape Laptev. 
This would seem to indicate an easterly motion of the fresh water forming 
the upper layer, relatively to the sea below through which the bulk of the ship 
was going, thus making the speed only apparently so small as given in the 
log book (for the last hours 1 or 2 knots with calm weather and the engine 
going full speed). In sailors’ parlance it looks as if the ship was dragging 
some miles of sea with her. But other observations show that such a cur- 
rent would not suffice as a sole cause, and both Mr. Nansen and Mr. Scott- 
Hansen are of opinion that the dead reckoning under these circumstances is 
not sufficiently trustworthy to form the basis of an explanation of this curious 
phenomenon. 
