LVIII GEELMUYDEN. ASTRONOMICAL OBSERVATIONS. [NORW. POL. EXP. 



tangent to one of them, and when the edge used is apparently the same (e. g. 

 the upper) in both positions, they are really different; thus the horizontal 

 point of the circle will be different in the two positions, unless the semidia- 

 meter is given a constant correction. In the mean the effect will of course 

 be eliminated. 



Both the travellers carried pocket chronometers which will be designated 

 in the following by I and II. The first, carried by Nansen, was marked 

 "Johannsen 6455", the other, carried by Johansen, was denoted in the jour- 

 nals of the From as ,,No. 19787"; it was Nansen's watch from the Green- 

 land expedition. Both were carefully compared with chronometer Hohwti by 

 Lieut. Scott-Hansen during several months before starting, and under varying 

 conditions which differed, however, considerably from those of the sledge 

 expedition. The mean daily rate of I on Mean Time during periods of a 

 week or more varied between 3. 8 9 and 5. 8 2 fast, of II between 1. 8 2 and 3. 8 3 

 slow. During the days February 26 March 6, including the first trial expedi- 

 tion, the rate of II was 2. 8 42, and during the next eight days which were 

 spent on board, it was 2. 8 35. Watch I was not compared with Hw on 

 returning to the ship (Nansen returned on March 3, Johansen on March 4) 

 but the mean rate during Febr. 26 March 14 was +5. 8 0. The relative rate 

 of I II was thus +7. 8 4. During the expedition it was more irregular, but 

 on the average greater, about 10 13 seconds. Both watches appear to have 

 been going faster, but I more than II. 



It happened several times that one of the watches ran down. Unfortu- 

 nately it happened also once, when the working day of the men had been 

 longer than that of the watches, that both ran down. The astronomical obser- 

 vations between which the stopping occured (1895 April 12) were 5 days apart. 

 As the weather was clear and the ice good, the dead reckoning for this inter- 

 val will probably not be much in error; but the drift of the ice is of course 

 unknown. 



There is, however, another difficulty which for a certain period of the 

 expedition is more serious than the stopping of the watches. During the 

 months of April and May all altitudes were measured with the sextant, and, 

 with one exception, from the natural horizon. In the afternoon of April 2 

 a series of 6 altitudes was taken, the first three with glass horizon, the rest 

 with natural horizon. The two sets give a difference of nearly ten minutes 



