xviii Introduction to the Makerstoun Observations, 1843. 



ever, differed slightly from that which it was conceived the ring occupied previously ; 

 and it should be remarked that when any cause (such as elimination of the torsion 

 of the suspension thread) rendered it necessary to remove the ring, it was replaced, 

 by means of marks on the marble slab, very nearly in the same position. After 

 the above series of observations, the ring was moved till its north end was 5° to the 

 east, and its south end was 5° to the west of its usual position, when the effect of 

 the ring was found = + 4*2 long scale divisions. 



Oct. 15. 1844. Observations were made with the ring in its usual position, and 

 in the position occupied during the series of observations in Table 5, when the effect 

 in the latter position, compared with that in the usual position, was found = — 0*8 

 long scale division, so that the effect of the copper ring in its usual position would 

 = — 0'9 scale division. A single comparison with the ring in its usual position 

 and away, gave the effect = — 1'4. 



13. The adopted effect of the copper ring during 1843 = — 1*0 long scale 

 division. 



As the effect of the ring varied with its position, if its position remained con- 

 stant it would also vary with the changing declination. When the westerly declina- 

 tion diminished, the negative effect of the copper ring would be increased, and vice 

 versa ; but this variation would be inconsiderable for the ordinary diurnal changes. 



14. Correction for the effect of the non-parallelism of the plates of glass in the 

 declinometer boxes. 



The mag-net being caused to rest on wooden blocks, the scale readings were 

 observed with the glass in its usual position, reversed, and away, but no difference 

 in the readings could be detected ; the effect is therefore zero. 



15. Correction for the torsion force of the suspension thread, and the prin- 

 cipal facts relating to the latter. 



The errors due to the torsion force of the suspension thread are produced, first, 

 by the magnet moving out of the plane of detorsion ; secondly, by the variation of 

 this plane (due generally to the varying humidity of the atmosphere).* The error 

 due to the former, even in the most marked case, is less than the probable error of 

 observations, and it is altogether inappreciable compared with the error due to the 

 second. 



If the plane of detorsion be that of the magnetic meridian, and the magnet be 

 deflected through the arc u by turning the arms of the torsion circle w, the torsion 

 is w — u, and the ratio 



u _ torsion force for an arc = radius _ Ht 



w — u earth's mag. force x mag. moment of the bar F 



* This source of error was removed to a considerable extent in January 1844, by placing a thick 

 cotton cover over the whole declinometer. 



\ The symbols used in the " Report of the Committee of Physics and Meteorology of the Royal 

 Society," and by Dr H. Lloyd, are generally adopted in this volume. 



