282 ON EARTH-CURRENTS, AND THE DIURNAL 



dividing the area of the diurnal curve unequally, and being nearer 

 to the night observations than to those of the day. From the last 

 line in Tables I. and III., we find that the mean of all the daily 

 observations of the galvanometers is neyatire on both lines, or 

 below the true zero. On the other hand, the mean of the niyhi 

 observations is positive, or above the zero ; and the same thing is 

 true of that portion of the night (from 1 A. M. to 5 A. M.) during 

 which the magnetic changes are smallest. From these facts it 

 would seem to follow that the true zero of the magnetometric 

 observations lies between the mean of the day and the mean of 

 the hours of magnetic repose. This conclusion, however, it must 

 be remembered, is derived from the observations of a single fort- 

 night only; and, in the absence of fuller knowledge, we should 

 not be justified in changing the origin usually employed. 



Thus magnetometric observations furnish merely differential 

 results; and we are ignorant even of the relative values of the 

 effects, and therefore unable to compare them accurately with their 

 physical causes, whether real or supposed. It is true that, if the 

 galvanometricandthe magnetometric results were completely iden- 

 tified, the zero of the latter could readily be obtained by their com- 

 parison. For if rj and be the values of rj and ,* corresponding 

 to the true zero, measured from the mean of the day, it is plain 

 that, for the hours at which the current changes sign on any line, 



ijo cos i - sin E = rj cos e - % sin t. 



If therefore these hours be known for two lines, which differ con- 

 siderably in azimuth, and also the corresponding values of j and , 

 we should have two equations for the determination of the two 

 unknown quantities, i and . It is obvious, however, that this 

 process cannot be legitimately employed in the comparison and 

 identification itself. 



12. In order to see how far the correspondence, which we 

 have found in the mean results, is traceable in separate days, I 

 projected in curves the observed and calculated values of the 

 intensity of the currents on the Derby and Birmingham line, as 

 given in Tables II. and IV., for the five days commencing May 25. 

 The accordance of the two curves is very remarkable, every alter- 

 nation in direction in one of them having its counterpart in the 

 other. 



13. It remains to say a few words of the manner in which the 



