MAGNETIC OBSERVATIONS. log 



T her., comparing equation (14) with equation (15), we find, to terms of the third 

 order inclusive, 



81 =A, 



a 



+ 5, ^ -L + iL 



IT 



2f 1 = lJ) 1 ' 



"W T hen the deviation of the compass is small, the several parts of which it is 

 composed are simply added together; these parts are, 



1. A, the constant deviation. 



2. B sin '-{- C cos ', the semicircular deviation. 



3. D sin 2' -f- " cos 2', the quadrantal deviation. 



"When the deviation is large, 21, 23, , 35, (, or the angles of which these 

 quantities are the natural sines, may still be considered as the constant and as the 

 several parts of the semicircular and the quadrantal deviation, each of these angles 

 being in fact the maximum deviation which would exist if all the other coefficients 

 were zero; but their effects are no longer combined by simple addition." 



Before submitting the observed deviations to comparison with the theory, it is 

 necessary to free them from constant errors. These errors originated in two ways. 



1. When the ship was swung, the variation of the needle at the port where she 

 was lying was seldom accurately known. Hence, in order to obtain the true 

 magnetic azimuth of the object used as an azimuth mark, it was necessary to 

 adopt, for the time being, the best value of the variation which happened to be 

 accessible. In order to facilitate the setting of the sight vanes of the Admiralty 

 Standard Compass while the ship was being swung, the value thus adopted was 

 always so taken that, when the ship's head pointed successively to each of the true 

 ^magnetic points, the reading of the sight vanes on the azimuth circle attached to 

 the cover of that compass was always either some whole degree or some quarter 

 of a degree. When the declinometer observations were reduced, the true value of 

 the variation of the compass at each port became known, and then it was discovered 



17 August, 1872. 



