398 



Lieut. Rokeby on Magnetic Observations, [Apr. 15, 



These observations were discontinued in June 1866, when Lieut. Rokeby 

 left the Island. 



Observations of absolute horizontal force and dip were made on the Green 

 Mountain, Ascension, once every month from July 1863 to March 18G6, 

 two months in 1865 excepted. 



Observations were not made with the portable declinometer. 



Observations of Horizontal Force and Dip. 



The horizontal, vertical, and total forces (Table No. 1) are calculated to 

 English measure ; one foot, one second of mean solar time, and one grain 

 being assumed as the limits of space, of time, and of mass. 



The vertical and total forces are obtained from the absolute measures of 

 the horizontal force and the dip. 



The observations of dip (Table No. 1) were made in every instance save 

 one with the needle marked A 2. 



For the observations of deflection and vibration taken each month for 

 absolute measure of horizontal force, the same magnet (collimator 5) has 

 always been employed. 



The moment of inertia of the magnet, with its stirrup for different de- 

 grees of temperature, and the coefficients in the corrections required for the 

 effects of temperature and of terrestrial magnetic induction on the magnetic 

 moment of the magnet, were determined in 1858 at the Kew Observatory 

 by the late Mr. Welsh. 



That these corrections held good in 1862 was proved by the agreement of 

 the horizontal force obtained at Kew with this instrument previous to 

 its departure, with the value of the force determined by the observatory 

 unifilar. 



The moment of inertia of the magnet with its stirrup is 5 '3828 at 

 60° Fahr. 



The induction-coefficient ji = 0-000252. 



The correction for error of graduation of the deflection-bar at TO foot is 

 0-00000 foot, and at 1-3 foot 0-00003 foot. 



The formula used for determining the temperature correction was 

 q(t -35°) + q' (* -35°) 2 , 

 where t is the observed temperature, 35° F. being the adopted standard 

 temperature. 



The values of q and q' for the magnet used are respectively 0*00011035 

 and 0-000000581. 



The observed times of vibration have been corrected for rate of chrono- 

 meter when it has exceeded five seconds per day. A correction has also 

 been applied for the effect of torsion on the suspending thread. 



The initial and terminal semiarcs of vibration have always been less than 

 30', consequently no correction was requisite on this account. 



The time of one vibration is derived from the mean of twelve observations 

 of the time of 100 vibrations. 



