Magnetic Variations at different Places, $c. 207 



would any attempt to discuss movements of a mixed or involved 

 nature, and it is found that motions of the declination and horizontal 

 force magnets for the instances contained in Table I are, without 

 exception, in direction such as would be produced by the action of a 

 current passing through a wire under the magnet, similar to that of 

 the actual accompanying earth current. 



But then comes a further question : Do earth current and mag- 

 netic times really agree ? For in the discussion of the Dartford and 

 Croydon earth current records, the earth current on some occasions, 

 according to the instrumental time scales, seemed undoubtedly to 

 follow the magnetic movement, instead of being coincident with, or 

 preceding, it. But the system on which the time scales were in those 

 days laid down was not arranged with a view to the accurate measure- 

 ment of small intervals of time, and in this respect did not permit of 

 an accuracy equal to that to be obtained from, later registers, on 

 which a time indication is made at every individual hour. Thus 

 we are now able to effect a much more certain comparison of the 

 times of magnetic and earth current movement than was before 

 possible. This has been done for a considerable number of instances 

 of sudden initial movement, including all those of Table I (excepting 

 1882, April 16, on which day the earth current register was defective), 

 and also a number of others, the whole particulars being given in 

 Table IV. 



When it is considered that each separate difference in this table 

 includes the sum of the errors of two time measures from photo- 

 graphs leaving a time scale in which one minute corresponds to about 

 0*01 inch, the agreement of results, it will be seen, is very good. 

 Every case selected for measurement is given ; nothing has been 

 rejected. The variation between individual differences does not 

 exceed in amount that which may well be expected, with a time scale 

 so contracted, and wo aid appear to be simply due to unavoidable 

 errors of measurement. The mean of the 31 differences is 0'14 m., 

 equivalent to 8 sec., or, dividing the 31 differences into three groups 

 of 11, 10, and 10 values respectively, the means of these groups are 

 found to be O19 m., 0'12 m., and O'lO m., equivalent to 11 sec., 

 7 sec., and 6 sec. respectively, thus persistently showing the earth 

 current to precede by a few seconds the magnetic movement. Not 

 that these results should be taken too strictly, but simply as showing, 

 when definite comparison can be made, that the coincidence of time 

 is really very close, possibly more so than present existing means 

 will permit us to determine. This so far removes the before-men- 

 tioned time anomaly, the observed facts being now seen to be alto- 

 gether consistent with the supposition that earth currents may pro- 

 duce the magnetic movements, whether or not, as regards cause and 

 effect, this be the true relation. For it has to be remembered that in 



VOL. LII. p 



