NO. 6.] INTRODUCTION. SLEDGE EXPEDITION. LIX 



in the clock error. It was first believed that the glass horizon had got out 

 of adjustment after levelling; a comparison with the result of a series of 

 5 altitudes taken in the morning of April 4 seems, however, to indicate another 

 explanation, viz. a constant error in the altitudes measured from the natural 

 horizon, evidently due to irregular terrestrial refraction causing the correction 

 for dip to be nearly as large positive as it should be negative under ordinary 

 circumstances for the given height of the eye. On both days the sky was 

 clear, the Sun above the horizon all day long, and the weather mostly calm, 

 but the temperature of the air below 30 C. There is some probability that 

 a similar anomaly may have taken place also an other occasions under the 

 same meteorological conditions; but the assumption that the horizon has on 

 all occasions been elevated to the same amount must necessarily be affected 

 by a considerable uncertainty. The same phenomenon (to a smaller extent) 

 made itself manifest later on at the winter hut though the temperature was 

 then much higher. 



Considering that in the high latitudes reached during this expedition an 

 error of 1' in an altitude measured near the prime vertical, that is to say 

 under the most favorable conditions, gives an error of a minute of time in 

 the clock correction, it will be understood that the determination of local time 

 by the means at hand was no easy task. 



On two occasions Mr. Nansen took Lunar Distances, one on the ice, the 

 other at the winter hut. After the stopping of the watches he was often on 

 the look out for the Moon during the periods of her visibility, but could not 

 perceive her with the naked eye on the pale sky with the strong reflection of 

 light from the immense white surface of the ice, till August 10; and even 

 then the Moon disappeared in the haze after the measuring of a single dis- 

 tance. The cutting out of the tables of Lunar Distances from the English 

 Nautical Almanac having been forgotten he had no other data for the Moon 

 than the mean time and the declination for upper culmination at Greenwich, 

 by which means the computation on the spot was of course rather difficult. 



The uncertainty of the Greenwich Time deduced from the Lunar Dis- 

 tances is not much greater than that of the Local Time. An approximate 

 determination of the longitude of the winter hut at Franz Joseph Land may 

 also be obtained by a combination of Mr. Nansen's observations in 1896 on 

 the way to Mr. JACKSON'S station at Cape Flora. The writer does not know 



