272 Reports on Special Researches 



Although the mean corrections, as is seen, are too small to make it worth while to 

 correct our published land magnetic results for dechnation and horizontal intensity, they 

 have been applied in Tables A and B (p. 278), which summarize the results of the observa- 

 tory comparisons. The application of the corrections slightly improves, on the average, 

 the observatory differences on the international standards, as determined in the above 

 manner. As C. I. W. magnetometer No. 3 (corrected by +0.00015//) has been the pro- 

 visional C. I. W. standard, the application of the mean correction 0.00015//, shown in 

 Table 31 D, gives for the resultant correction of C. I. W. No. 3 on I. M. S., 0.00000//; this 

 apparently perfect accord is wholly accidental for no greater accuracy in the quantities 

 in Table 31 D than about 0.0001// can be expected. 



IMPROVEMENT OF INCLINATION STANDARD. 



We were not quite as fortunate in the selection of the provisional standard for inclina- 

 tion /, as was shown above to have been the case with respect to our provisional standards 

 for declination and horizontal intensity. Our provisional /-standard, as has already been 

 stated, was the large Wild-Edelmann earth inductor (No. 26) at the Cheltenham Magnetic 

 Observatory. Up to about 1908 all indications pointed to an agreement of the Cheltenham 

 standard with the Potsdam standard (earth inductor Schulze No. 1) within about 0'.2. 

 Thus, Schulze earth inductor No. 48 ( Wild-Eschenhagen pattern) , purchased by the Depart- 

 ment of Terrestrial Magnetism in 1907, was found, from limited comparisons (3 sets), by 

 the Potsdam Observator}- in Februarj- 1907 to have a correction of about 0'.27 on the 

 Potsdam standard; from more elaborate comparisons (10 sets), in March and April 1908 at 

 Cheltenham it was found that Schulze No. 48 required a correction of 0'..54 on the Chel- 

 tenham standard; the weighted mean of the Potsdam and Cheltenham comparisons was 

 nearly 0'.5. We accordingly adopted as the provisional C. I. W. standard for /, Schulze 

 No. 48 corrected by 0'.5 (for footscrew A south). 



The very large number of comparisons, 1907-14, with dip circles and earth inductors 

 of various design and make, gradually began to show that the C.I. W. provisional standard 

 fixed above was about 0'.5 too low on an international standard. This is borne out by the 

 following table (No. 31 E) from which it appears that we should apply a correction of -|-0'.5 

 to the provisionally adopted C. I. W. Inclination Standard to refer it to the International 

 Standai'd. Or, in other words, if the negative correction of 0'.5 which was originally applied 

 to our Schulze earth inductor No. 48 be dropped, this instrument will give inclinations 

 apparently conforming \vith an International Standard within about O'.l. /Accordingly, all 

 results of inclination comparisons for the quantity (C. I. W. Observatory), as given in the 

 tables, pp. 214-267, have been algebraically corrected by +0'.5 in referring them to I. M. S. 



Table 31 E. Correction of C. I. W. Inclination Standard. 



No. 



Based on 



A/ Weight 



I 

 II 



III 

 IV 



Me.in of 10 Dover laud dip circles compared at Washington, 1907-14 



Mean of 5 magnetometer-inclinometers (dip circles), made by Department of Terrestrial Mag- 

 netism (needles furnished Iiy Dover), and compared at Washington, 1910-14 



Mean of 8 earth inductors of various designs belonging to the Department of Terrestrial Mag- 

 netism and compared at Washington, 1907-14 



Mean of 20 observatory standards, consisting of 10 dip circles and 10 earth inductors, 1910-14 . . . , 



-1-0.55 2 



-fO.36 j 1 



+0.61 1 2 



+0.50 I 3 



Weighted mean correction of C. I. \V. Stand.ird, 1907-1914 (on Iiitcrnalionul Magnetic Staudaid).. . +0. 



Considering the various errors to which field observations for inclination are subject, 

 it will probably not be found worth wliile to apply to the inclination results of 1905-1913 

 given in \o\. I and this volume, a correction which, on the average, will possibly not be 



