56 HANDBOOK OF MAGNETIC COMPASS 



(2) More often it is desirable to begin adjustment immedi- 

 ately, eliminating the original swing for deviations and the esti- 

 mate of approximate coefficients. In this case the above prob- 

 lem would be solved by tabulating data and anticipating devia- 

 tion changes as the corrections are made. Figure 29 illustrates 

 such procedure. It will be noted that a new column of values 

 is started after each change is made. This method of tabulation 

 enables the adjuster to calculate the new residual deviations each 

 time a corrector is changed, so that a record of deviations is avail- 

 able at all times during the swing. Arrows are used to indicate 

 where each change is made. 



Figure. 29. — Tabulating anticipated deviations — One-swing method. 



Since the B error is generally greatest, it is corrected first; 

 hence on a 090° heading the 11.5° E. deviation is corrected to 

 approximately zero by using fore-and-aft B magnets. A lot of 

 time need not be spent trying to reduce this deviation to exactly 

 zero since the B coefficient may not be exactly 11.5° E., and some 

 splitting might be desirable later. After correcting on the 090° 

 heading, the swing would then be continued to 135° where a 9.2° 

 W. error would be observed. This deviation is recorded, but 

 no correction is made because the quadrantal error is best cor- 

 rected after the deviations on all four cardinal headings have 

 been corrected. The deviation on the 180° heading would be ob- 

 served as 5.5° W. Since this deviation is not too large and split- 

 ting may be necessary later, it need not be corrected at this time. 

 Continuing the swing to 225°, 0.0° deviation would be observed 

 and recorded. On the 270° heading the observed error would 

 be 1.0° W., which is compared with 0.0° deviation on the opposite 

 090° heading. This could be split, leaving 0.5° W. deviation on 

 both 090° and 270°, but since this is so small it may be left. On 

 315° the observed deviation would be 1.2° E. At 000°, a deviation 



