DEPARTMENT OF TERRESTRIAL MAGNETISM. 321 



As the result of numerous comparisons, it would seem safe to conclude that 

 the H, or horizontal-intensity, standards of Kew, Potsdam, and Washington 

 (Department of Terrestrial Magnetism) give an absolute accuracy within 1 

 part in 10,000, except for some possible error which may be inherent in the 

 magnetometer method but not yet disclosed. In brief, to reach an accuracy 

 in the determination of the Earth's magnetic field intensity within 1 in 

 10,000 by magnetic methods requires great -pains, but it can be done, if necessary, 

 and with a comparatively simple type of magnetometer. 



For all field requirements, an absolute accuracy of 5 in 10,000 amply suffices, 

 and this can readily be secured with a properly constructed instrument. 

 For observatory purposes, where the prime purpose is the study of the mag- 

 netic variations, it is a question chiefly of relative accuracy of the magneto- 

 meter employed. Various comparisons between skilled observers would indi- 

 cate that their instrumental difference as resulting from any one set of careful 

 observations extending over about an hour and a quarter need not differ from 

 the mean of several days' observations by much more than about 1 in 10,000. 

 //, then, further improvement is sought in magnetic measures, it should be in the 

 direction of securing the desired degree of accuracy with less labor and less time. 

 The chief direction in which the present field magnetometers remain to be 

 improved is the partial or total elimination of oscillations; the transport of 

 such a delicate piece of mechanism as a chronometer, in comparatively unex- 

 plored regions, is a continual source of worry to the observer. 



When the ocean work of the Department was begun in 1905, it was necessarj' 

 either to design entirely new or to modify considerably existing devices. There 

 have resulted thus the following new instruments, for use at sea, designed as 

 well as chiefly constructed by the Department : 



C. I. W. deflector for determining the magnetic declination (D) and the horizontal intensity 



(//), designed by L. A. Bauer and J. A. Fleming (see Terr. Mag., vol. 11, 1906, pp. 



79-82: vol. 14, 1909, pp. 167-168; and vol. 18, 1913, pp. 57-62). 

 C. I. W. modified L. C. dip circle for determining the magnetic inclination and the total 



intensity, constructed by A. W. Dover, according to specifications of L. A. Bauer 



(see Terr. Mag., vol. 11, 1906, pp. 77-78, and vol. 14, 1909, pp. 164-166). 

 C. I. W. marine collimating compass, primarily for determining D, but may also, if desired, 



be adapted to measuring H, designed by W. J. Peters (see Terr. Mag., vol. 14, 1909, 



pp. 17-24). 

 C. I. W. marine earth inductor, designed bv N. E. Dorsey and J. A. Fleming (see Terr. 



Mag., vol. 18, 1913, pp. 1-48). 



With these instrumental devices, an accuracy in ocean magnetic work has 

 been secured which, under favorable conditions and devoting the same time as 

 in land work, do not fall much short of the general accuracy of land field-work. 



Electric methods.— One great disadvantage of the earth inductor, in its 

 present form, lies in the fact that it is not readily adapted to ex-meridian 

 observations, such as would be requisite when working in the vicinity of a 

 magnetic pole. Or, if an observer elsewhere accidentally makes a mis-setting, 

 there is no such simple formula, as in the case of the dip circle, for computing 

 the true dip from the ex-meridian one. Only under certain conditions can 

 ex-meridian observations with the earth inductor be utilized, as was shown in 

 the investigation "On the theory of the earth inductor as an inclinometer," 

 made for the Department by Dr. N. E. Dorsey and published in Terrestrial 

 Magnetism, vol. 18, 1913, pp. 1-38. 



The requirements for measuring the intensity of the Earth's magnetic field 

 by balancing it against that of a field produced by a current in some form of 

 coil, common to the methods thus far proposed, are: (a) accurate knowledge 

 of the current ; (6) accurate knowledge of the essential dimensions of the coil, 



