NO. 6.] 



INTRODUCTION. CHRONOMETERS. 



XLIII 



Table b (concluded). 



Sat. IV D. x = 0.49. 



3.97 

 -3.82 

 -3.72 



1.73 

 1.61 

 1.61 



Sat. IV R. x = 0.54. 



1895 March 8 

 March 8 

 April 10 



Jena 



Uccle 



Windsor 



16. 

 15. 

 20. 



-4.23 

 4.12 



-4.70 



1.61 

 1.61 

 1.41 



-0.010 

 -0.010 

 -0.0136 



-19m 58s 

 -18 33: 

 -13 17 



15m 36s 

 14 17: 

 - 9 11 



An inspection of the last column of Table b shows that it is no easy 

 matter to deduce corrections to the predicted times. It is evident that the 

 correction cannot be considered as constant for any length of time; it was 

 therefore necessary to make some combination of the results surrounding the 

 Fram-observations, but in making such combinations some arbitrariness is 

 scarcely avoidable. In the few cases where the eclipses observed on board 

 had also been observed elsewhere the deduced correction could, for some of 

 them, be applied without alteration, but for others a combination was pre- 

 ferred when sufficient surrounding material was at hand. The corrections 

 deduced from such combinations were often rounded to the nearest 5 or 10 

 seconds. 



