Chap. 8] MAGNETIC METHOD 367 



or have low sensitivity. The same is true for recording instruments when 

 they are compensated or used in rooms of constant temperature. 



2. Even when a temperature correction is apphed, magnets and magne- 

 tometers are usually well insulated in order to cut down temperature fluctu- 

 ations and temperature hysteresis. An instrument finish of high reflecting 

 power gives best protection against excessive temperature changes. The 

 procedure in determining temperature coefficients of Schmidt balances was 

 described on page 331. Other magnetometers are treated in a similar 

 manner. 



B. Magnetic Variations 



Magnetic variations may be divided into periodic (daily or diurnal varia- 

 tion), nonperiodic (magnetic storms), secular variation, and artificial 

 variations due to stray currents, and the like. 



1. The diurnal variation is closely related to the earth's rotation. It is 

 greater and more irregular during the day than at night; its amplitude and 

 phase varies with season and latitude (see Figs. 8-40 and 8-41). Pro- 

 cedures for elimination or correction are discussed below in ascending order 

 of accuracy. 



When large anomalies are encountered or when the sensitivity of the 

 instrument is low, the variation is usually disregarded. Variations may 

 be eliminated by simultaneous observations with two instruments on suc- 

 cessive stations. In stationary instruments magnetic variations may be 

 eliminated by the use of astatic systems and by shielding where applicable. 



Moderate accuracy in the elimination of the variation is attained by 

 checking back at a base station as often as possible and by considering the 

 daily variation as part of the variation of the base value. If measure- 

 ments at the base are not possible at regular intervals, it is advisable to 

 select times between which the change is approximately linear. 



Curves as shown in Figs. 8-40 and 8-41 can give reliable data for daily 

 variation corrections, particularly when measurements are made fairly 

 near a magnetic observatory for which such curves are available. Allow- 

 ance must be made for differences in mean local time as shown in the 

 next paragraph. 



Greater accuracy is attainable by arranging with magnetic observatories 

 for tracings or photostatic reproductions of their records for specific 

 days. The curve-amplitudes must be multiplied by scale value and 

 must be corrected for temperature (unless the instruments are com- 

 pensated or are set up in constant temperature rooms) and for difference 

 in local mean time between observatory and area under survey (AX of 



" For curves on magnetometers of different makes see J. B. Ostermeier, Zeit. 

 Erdoel. Bergbau, etc., June 14, 1930. 



