148 



observing cabin had, in the interval, undergone very great change; possibly this change was due to 

 alterations in the stowage of stores in the ship. The difference is so large that it has been considered 

 advisable to reduce the observations made in 1904, by means of data obtained in the same year only. 



In addition to the swing for compass error already mentioned, the vessel was also swung at Auckland 

 Islands, Lyttleton, Falkland Islands, and Spithead. Kelative observations for horizontal and vertical 

 force on board were made at Lyttleton, Falkland Islands, and Spithead. 



The results are given in the following Table XIV., the notation being that of the 'Admiralty Manual 

 of Deviations of the Compass,' and of the ' Antarctic Manual ' : 



TABLE XIV. 



The results of the observations at Lyttleton, on investigation, show certain inconsistencies, and are on 

 some points doubtful ; they have therefore not been used in the following reductions, except in the case 

 of X, the mean values of the constant coefficients used being 



X = 0-973, 

 s = -0-0055, 



AX = +0-991, 

 D = +1 7'. 



From the data in Table XIV. (omitting the Lyttleton results), by separate combinations, the values of 

 the parameters P, Q, c, and / (' Admiralty Manual of Deviations of the Compass '), were calculated, with 

 the following results : 



From these mean values, curves were drawn showing the values of 



c tan f tan 



~T~' 



for all values of 



and -, 



A A AH' 



(the Inclination) and H (the Horizontal Force), 



the values of H being calculated on the assumption that over the area under consideration the Total Force 

 did not vary (see Section IX.), and the value of 6 was taken from N, the natural tangent of the Inclination 

 as calculated from each observation. 



By means of the data obtained from the curves and employing the formulae 



and 



1 / P\ 



B = / c tan 6 + == ) (' Admiralty Manual of Deviations '), 

 *\ ti 



