628 [June 18, 



to Sept. 23, and 24" to the west in the thirteen fortnights from 

 Sept. 24 to March 25. This is in accordance with the conclusions 

 at all the other stations at which the phenomena have been sub- 

 jected to a suitable investigation. The antagonistic horizontal com- 

 ponent of the earth's magnetism is approximately 4*5. 



Eighth Section. In the eighth section the author examines the 

 evidence which the irionthly determinations of the dip and of the hori- 

 zontal component of the magnetic force at Kew afford of the existence 

 of a semiannual inequality in the absolute values of the dip and of the 

 total magnetic force. The results of the monthly determinations from 

 April 1857 to March 1863 are exhibited in two Tables, one appro- 

 priated to the dip, and the other to the horizontal force. The whole 

 series of the determinations of the horizontal force were made with the 

 same unifilar magnetometer and the same collimator magnet through- 

 out, and also by the same observer, Mr. Chambers, one of the assist- 

 ants at the Kew Observatory. In the monthly determinations of the 

 dip from April 1857 to September 1860, twelve different circles, and 

 their twenty-four needles were occasionally employed, the mean of all 

 the observations in each month being taken as the mean result in that 

 month. There were also several observers in this part of the series, 

 chiefly four. In the months from October 1860 to March 1863, 

 one circle with its two needles were the sole instruments, and Mr. 

 Chambers the sole observer. The probable error of a single monthly 

 determination of the dip in the first part of the series, when several 

 instruments and several observers were employed, is stated to be 

 + 0'*69 ; and in the second part of the series, obtained by a single 

 circle and the same observer throughout, the probable error is + 0'' 75 ; 

 whence it is inferred that the greater number of partial results which 

 contributed to produce the monthly mean in the earlier period more 

 than counterbalanced the diversities which might have been occa- 

 sioned by the peculiarities of the different observers and of the differ- 

 ent instruments. The probable error of a single monthly determina- 

 tion of the dip, after the application of the corrections for secular 

 change and annual variation, is stated to be + 0''71, and of a single 

 monthly determination of the horizontal force derived from the 72 

 monthly determinations + '0024. 



The results of the monthly determinations at Kew, as bearing 



