145 



for all observations made with ship's head between 



N.E. and S.E., 



N.W. S.W. 

 In these formulae 



f represents the direction of ship's head as shown by the compass. 



f represents the magnetic direction of the ship's head. 



ff represents the uncorrected Inclination as observed. 



B, C, and D represent the coefficients of deviation of the compass. 



N represents the natural tangent of the Inclination value after correction for errors due to the 

 direction of ship's head. 



s represents the maximum effect on the Inclination, which would be caused by the induced 

 magnetism in a horizontal soft-iron rod in the fore and aft direction, one end of which is 

 immediately below the dip circle (rod g of the ' Admiralty Manual of Deviations '). 



By means of these formulae, values of s cos f + N, for positions (1), (2), (3), and (4), were calculated to be 



o / 



(1) tan" 1 18-365 = 86 53 Inclination S. 



(2) tan" 1 15-821 =86 23 



(3) tan" 1 16-198 = 86 28 



(4) tan" 1 17-793 = 86 47 



Mean = 86 38' 



Observations at Spithead in August, 1901, showed that s was zero, and therefore the above values 

 represent N (the natural tangent of the Inclination, corrected for errors due to the direction of ship's head). 



The values of the Inclination so reduced should agree for all directions of the ship's head ; the differences 

 of the above results from the mean are, however, considerable, viz. : 



(1) +16 



(2) -15 



(3) -10 



(4) +9 



These differences are consistent with an error which has a maximum when the ship's head is N.W. or 

 S.E., varying as the cosine of the azimuth of the ship's head measured from these points. 



Such an error might be caused by an elongated mass of horizontal soft iron situated at an angle of 45 

 to the keel of the ship, and having one end directly under the position of the dip circle. The error is 

 similar to that which is indicated in the formulae, by s cos f, but having a maximum value when the ship's 

 head is N.W. or S.E. instead of N. or S. 



The differences as found were plotted and a curve drawn, from which the correction for any direction 

 of ship's head could be measured, and these corrections were applied as a constant for the particular 

 direction of the ship's head during each observation. 



The corrections were applied to the observed readings after the correction for instrumental differences. 



From the value of Inclination so deduced, the value of N was calculated by means of the formulae 



N = (cos f + sin B) sec f tan ff 

 and 



N = {(1-2 sin D) sin f - sin C} cosec f tan B'. 



To ascertain the correct values of the coefficients of compass deviation for different Inclinations and thus 

 determine the necessary corrections to compass readings, it was assumed that over the area under 

 consideration the Total Force might be treated as constant. 



This assumption is supported by the Total Force results during Mr. BERNACCHI'S S.E. sledge journey and 



U 



