470 



Special Reports 



able error of a single determination of the difference (I. M. S. C. I. W. sine galvanometer 

 No. 1) certainly indicates that well designed magnetometers with carefully determined 

 constants are capable of an absolute precision of 0.000 15H. 



III. Summary of Mean Correction's on I. M. S. 



Table 20G summarizes the mean corrections on I. M. S. for various instruments and 

 observatories as determined from all the data considered in this and in the first report 

 (see Vol. II, pp. 211-278) and obtained during the 15 years from 1907 to 1921. The 

 weighted mean corrections resulting are practically zero for AD, &H/H, aDd Al. Thus 

 with the additional evidence afforded by the results of the comparisons between I. M. S. 

 and absolute values derived from the sine galvanometer (see Table 20F) there seems 

 little doubt that, for all practical theoretical investigations, the provisional international 

 standards designated as "I. M. S." in Volume II are correct well within the errors of 

 observation and of the determination of constants. 



'See p. 6; cf. also Res. Dep. Terr. Mag., Vol. I, pp. 2-12, and Vol. II, pp. 5-15. 



-Correction applied to computed values of H account of bending of deflection bar, O.OOOOStf. 



-Correction applied to computed values of H account of bending of deflection bar, 0.00012H. 



