for computing Magnetic Declination. 1 85 



posing to compute magnetic declination. The epochs selected 

 are the years 14'92 and 1722: the first is 357 years in the 

 past, and the latter 127 years. The epoch of 1492 refers to 

 the celebrated discovery of the magnetic line of no declination, 

 during the first voyage of Columbus; and the epoch of 1722 

 refers to Graham's excellent observations on the declination 

 of the needle at London. The epochs are as memorable as 

 any in the history of terrestrial magnetism, and are as widely 

 distant from each other, and from the present, as any which 

 can be selected for the trial we propose. 



In the first place, we will state the data on which we ground 

 the adaptation of our method to different epochs. The data 

 simply consist in discovering the annual velocity of the north 

 and south magnetic poles ; for when this is obtained, their situa- 

 tion at any given epoch is easily found, by multiplying the 

 annual velocity of each by the number of years intervening 

 between the epoch and a given year (1849), when their rela- 

 tive situation is known. After a long and almost hopeless 

 investigation, I have ascertained the velocity in longitude of 

 the north magnetic pole to be double that of the magnetic 

 pole in the southern hemisphere, and that both liave had a 

 diminishing velocity, with occasional exceptions*, since the 

 epoch of 1492. The rule for computing the velocity and 

 situation of the two poles is as follows: — Commencing with the 

 year 1460, multiply the whole interval between that year and 

 1849 (389 years) by seventeen minutes ; and for every success- 

 ive diminution of 60 years in that interval, lessen the multi- 

 plying number one minute, as in the following table. The 

 fixed epoch by which the interval is determined is the past 

 year 1849. 



Forall epochs between 1460 andl 520 multiply theinterval by 17'. 



1520 ... 1580 ... ... 16'. 



1580 ... 1640 ... ... 15'. 



1640 ... 1700 ... ... 14'. 



1700 ... 1760 ... ... 13', 



1760 ... 1820 ... ... 12'. 



1820 ... 1880 ... ... 11'. 



This table is only for the northmagnetic pole ; in all cases 

 the multiplying number for the south magnetic pole must 

 be one-half the multiplying number for the opposite pole. 

 The mean positions of the two magnetic poles for the fixed 

 epoch 1849, are as follows : — 



* Grover's memoir, Orbital Motion of the Magnetic Pole round the 

 North Pole of the Earth, read at the Nineteenth Annual Meeting of the 

 British Association held at Birmingham, 1849. 



