368 



Special Reports 



were of the slightly larger type used with the universal magnetometer of the Department 

 (seep. 6). The resulting weights, magnetic moments, and constants for each of the 10 

 needles are shown in Table 47. Equation (8) assumes that and /J are of the same sign, 

 that is, that (F A) is positive when there is equilibrium. Table 47 shows that values 

 of A for different needles he within the range of values of F and thus under certain con- 

 ditions - will be negative, signifying that for such cases equilibrium will be impossible 



1 Using g = 980 cm. and r = 0.025 cm. 



for any values of /9 between +6 and B when the needle will rest upon the tangents 

 TG or T'G (Fig. 8a). Hence, all needles are more unsteady when used in regions of low 

 total intensity. This deduction agrees with common experience, and observers will 

 recognize the condition of instability as one where the needle "bumps" and will take up 

 either of two quite different positions equally well. 



P 



Fig. 9a. Fia. 9b. 



Theory of Minute Pivot-Defects on Dip-Needle Axles. 



In equation (11) the angle 8 represents one-half the range of inclination over which 

 the effect of the rust particle extends, so that a comparison between the range and the 

 amplitude may be made for any assumed case. Thus as an average condition for 

 A =0.480 and F = 0.600 from (12) we have Aln= 0.10. Reference to the correction- 

 curves given in Figures 4, 5, and 6 shows 8 will be about 2 to 2. 5 for an average case, 

 so that the maximum correction to be expected under the above assumptions would 

 be of the order 15'. This order of correction is approximated only in one case examined, 

 viz, that shown by Figure 5; hence it is improbable that the short-period deviations are 

 caused by rust particles of this simple form. 



