10 



Land Magnetic Obsekvations, 1914-20 



the mean correction is usually within about 0'.2 in declination and about 0.0001// in 

 horizontal intensity. The tabulated corrections are to be applied algebraically, east 

 decimation being reckoned as positive and west declination as negative; horizontal in- 

 tensity is always taken as positive. 



It will be noted that for some of the instruments the //-corrections vary with time ; 

 this is because of gradual change with time during field use in the moment of inertia, K, of 

 the long magnet system. That such changes take place, particularly in the tropics and for 

 magnets sheathed with brass, and that in general they are closely linear with time, is 

 shown clearly by discussion of the results from numerous intercomparisons at Wash- 

 ington to be contained in a later special report. In some cases the final values as given in 

 Table 2, of the distribution coefficients, P and Q or P', which result from compilations of 

 available data through 1920, differ from the values used for the original constants and 

 computations ; for several of the instruments the same remark applies for the final value of 

 logarithm t 2 K as given in Table 2. The tabulated //-corrections shown by Table 3 apply 

 for the constants given in Table 2. The special report to be entitled "Discussion of 

 Magnetometer Constants and Corrections on Standards" will give in detail the reduction- 

 factors involving changes in constants and modifications of these constants with the 

 gradual accumulation of data which tend to eliminate in the mean the unavoidable ac- 

 cidental errors. 



' There is a variation in the //-correction because of change in K (moment of inertia of long magnetic system) for this 

 instrument while in field use at the rate of about 0.0003/? per year during 1915; the instrument was not in field use again 

 until June 1918, and the indicated annual effect because of change in K on //-correction is only 0.00007/? between 1915 

 and 1919. Pending the first opportunity for new determination of K and re-standardization observations upon return of 

 instrument from field, values of //-correction are adopted as given. The //-correction from comparisons at Washington, 

 referred to same constants, was 0.00001/? at 1910.15. 



