28 AKSEL S. STEEN. [2nd arc. exp. fram 



DETERMINATION OF THE CONSTANTS. 



As already mentioned in the Introduction, a series of determinations 

 of the horizontal intensity were made with the apparatus at Wilhelms- 

 haven before the departure of the Expedition by Captains Baumann and 

 IsACHSEN, from the 19th to the 28th April, 1898, a corresponding series 

 of observations having been made with the apparatus by Captain Amund- 

 sen, after the return of the Expedition, at Potsdam, on the 29th and 

 30th November and 1st December, 1902. 



It is these two sets of observations that form the basis of my cal- 

 culations of the values that the constants a, /.i and G may be assumed 

 to have had in the observations made during the Expedition. 



DETERMINATION OF THE TEMPERATURE COEFFICIENT a. 



The result of the determinations of the temperature coefficient made 

 during the observations at Wilhelmshaven in 1898 was, 



a 

 for magnet V, 0.000204 

 for magnet VI, 0.000497 

 Amundsen's determinations at Potsdam in 1902 made it, 



a 

 for magnet V, 0.000243 

 for magnet VI, 0.000625 

 At the time that the constants of the instruments were first deter- 

 mined in Hamburg in 1893, shortly before the departure of the 1st Fram 

 Expedition, Dr. Neumayer found the following values: 



a 

 For magnet V, 0.000307 

 For magnet VI, 0.000638 

 It will be seen that these values tally better with Amundsen's deter- 

 minations than with those found in 1898 at Wilhelmshaven. Nor can 

 any very great degree of certainty be attributed to the last-named, as 

 the two observers were then less practised in the taking of magnetic 



H= [1 + fi't^ -^/Stg, + a (ij, -U] 



-^y sin 95 -^ -^ 



and also that /? and /?", by a printer's error (the omission of a 0), have bieen 



given vakies ten times greater than they should be. These mistakes, however, 



have fortunately had no influence upon the results as the latter were calculated 



from the above formulae (1) & (2), with the correct values of /? and /3'. 



