Dip-Needle Errors 371 



CONCLUSIONS. 



1. In order to furnish the data required for a proper elimination of results affected 

 by imperfect pivots, not less than four needles should be used at each station. 



2. An inconsistent value by a given needle at any station does not necessarily imply 

 that the needle will not behave satisfactorily at another station where the inclination 

 is slightly different. 



3. Although the correction for a needle may remain indefinitely constant at a given 

 value of inclination, it is not safe to assume that it will certainly do so, on account of 

 the possibility that minute defects on the pivots may appear during the progress of 

 work ; such defects are not necessarily permanent, for if they are rust masses, they may 

 possibly be removed by friction with the material used for cleaning. 



4. The stability of a needle when disturbed by a defect on the pivot depends on the 



condition that the coefficient __ . , shall be positive and as small as possible, where F 



WOT 



is the total force and A is the needle-constant . Since this constant for ordinary 



Dover needles has a range of values not far different from the range in F, needles for 

 work in a given region should be carefully selected. 



5. From comparison observations at a base-station and without a knowledge of the 

 behavior of a needle in adjacent regions, it is not possible to decide whether an unusual 

 result may be caused by a pivot defect ; nor has any practical method been found whereby 

 the existence and position of these defects may be determined, other than by actual 

 field observations. 



6. Since the magnitude of pivot defects may vary from comparatively large ones 

 as studied in this discussion to others indefinitely smaller, and since the distribution of 

 field stations can not always be sufficient to reveal the possible existence of such defects, 

 the weakness of the dip circle as an instrument of high precision is apparent. 



