MAGNETIC SURVEY OF GREAT BRITAIN. 107 



was greatest in the graduated limb, the very part in which, from 

 its proximity to the needle, it must operate most powerfully. 



I had next to consider the painful question, — How far the nu- 

 merous results obtained with this instrument were vitiated by 

 this newly-discovered source of error? Whether they were 

 entitled to any confidence ; and if so, what were the probable 

 limits of error ? It is manifest that if the ferruginous matter 

 were uniformly distributed throughout the limb, it could pro- 

 duce no disturbance in the position of a needle which (Uke the 

 dipping needle) divides the hmb symmetrically. It is only by 

 an irregularity in its distribution that the magnetic matter of 

 the Hmb can operate as a disturbing cause ; and then it is ma- 

 nifestly only by the difference of the attractions, on the two 

 sides of each pole, that the needle is actually disturbed. Hence, 

 though the magnetism of the limb may produce very decided 

 effects upon a test needle, in a position at right angles to its 

 plane, the effect upon a dipping needle may be comparatively 

 trifling. 



In order to estimate the amount of these effects, I separated 

 the divided circle from the apparatus, and placed it on a hori- 

 zontal support of wood. Three strong pins in contact with the 

 inner edge of the limb, and dividing it equally, were then driven 

 into the support, so as to prevent the limb from having any 

 motion, except one of rotation in its own plane. A magnetic 

 bar, wh9se length was nearly equal to the diameter of the circle, 

 was then supported delicately within it, and the deviation of 

 the bar from its undisturbed position was observed in the 

 different positions of the limb with respect to it. It was thus 

 found that most parts of the limb exerted a sensible disturbing 

 effect upon the needle ; and that this effect was not only con- 

 siderable in the neighbourhood of the two zero points of the limb 

 (the part where the anomalies had been first observed), but that 

 it also varied there very rapidly. A detailed examination of 

 the effects in this position showed that there was a disturbing 

 centre of ferruginous matter in the neighbourhood of each of 

 these points, and that it was to the action of these centres 

 that the anomalies in the observations above alluded to were 

 owing. 



In the neighbourhood of the divisions of 70° the disturbance 

 of the needle was likewise considerable, and its direction was 

 such as to diminish the apparent dip. Here, then, we have the 

 cause of the large negative error of the results obtained with 

 this instrument. But this deflection did not vary rapidly on 

 either side of these positions, so that for small changes of dip 



