auroral manifestations have been known to deflect the needle more 

 than 20. 



The magnetic needle freely suspended on a silk fiber is known, 

 as a matter of observation, to be seldom or never at rest. The 

 needle which reouirpi the longest time to settle will doubtless be 

 he most trustworthy in indicating the algebraic sum of the magnetic 

 fortes at the place of observation. 



This total magnetic force, pulling the needle either to the righ\ 

 or to left of the true meridian, is known as the "Declination of the 

 Needle". Surveyors and mariners call it '"variation" so that the 

 scale upon which a correction may be made for this error is known 

 as the Variation Plate, in contradistinction to the Declination Arc 

 which is used in connection with solar attachments for quite a 

 different purpose. In the spring of 1913 we designed a new varia- 

 tion ring, which not only compensates for horizontal displacement 

 but for about 3 of dip as well. This construction permits us to 

 turn off by vernier, to the nearest minute, any variation up to 90. 

 This recourse has been devised to meet the personal preferences of 

 certain surveyors who are particular about this idea. If the needle 

 is not properly balanced, or the vertical component of the earth's 

 magnetism suddenly becomes excessive, the needle will dip down- 

 ward, or a wind storm may so charge the compass glass with static 

 electricity that the needle will rise and occasionally adhere to the 

 glass. The first case will be met by the graduations on the inner 

 vertical edge of the ring, or perhaps by readjusting the counterpoise 

 weight if the conditions grow severe; but the second annoyance 

 must be dispelled by breathing a coat of moisture over the glass 

 cover. 



To Turn off Variation in any required amount, loosen the 

 thumb screw in the compass rim, as shown in Fig. 48, and push the 

 circle along past the index that is just above the letter " \ " If the 

 ordinary limitations are not sufficient, remove the thumb screw and 

 insert it in the next hole. Variation to the nearest 10' can be set b\ 

 inspection, but if one wishes to exceed this limit of accuracy, the 

 rest must be done by vernier. We never furnish the variation ver- 

 nier except on special order. 



If one attempts to read a needle to the nearest 10', the 

 uncertainty of reading, with other errors, is perhaps 5' or <>', 

 which would cause a probable total error of 10' or more for 

 each observation and an eiror of ctotun . t double this 



amount. 



(See also Declinatoiro pa g.- 140.) 



N 



