104 



KIGHTH REPORT 1838. 



Table XXXIII. 

 Mr. Lloyd's Observations in Dublin in 1836. 



The observations of Captain Ross were made in October and 

 November, 1838, with the needles designated as R, 4, R. 5, 

 R. 6, R. 7, L. 3, L. 4, in the preceding pages. The stations 

 of observation being sufficiently numerous, as well as uniformly 

 distributed, it has been thought advisable to combine them in a 

 separate determination. The observations are contained in 

 Tables XXXIX. and XL. 



We have now to consider the actual errors of the instru- 

 ments employed in the preceding observations. 



The errors of dipping needles may be ascribed to one or 

 other of the three following causes: namely, 1, the friction of 

 the axle on its supports ; 2, the imperfect curvature of the 

 axle itself; 3, magiietism in the limb. 



It is owing to the first-mentioned cause that a dipping nee- 

 dle assumes, in general, a new position of equilibrium after it 

 has been disturbed, the limit of error being the angle at which 

 the directive force, increasing as the sine of the deviation, 

 becomes equal to the friction. This limit varies, for a given 

 state of polish of the axle and of its supports, with the radius 

 of the cylindrical axle, the weight of the needle, and its direct- 

 ive fores'^. In all the earlier dipping needles constructed in 

 this country, this limit of error is considerable, owing to the 

 unnecessary size of the axle. 



The errors arising from the two latter causes are, however, 

 of a very different nature. The positive and negative errors 

 due to friction are equally probable, and the effect of the dis- 



• Trans. Royal Irish Academy. Vol. xrii. p. 1 66. 



