MAGNETIC DECLINATION NEAR THE MAGNETIC EQUATOR. 677 



The curves resembling that for January occur chiefly towards the beginning 

 of the month, and those resembling the curve for June in the latter part of the 

 month. In some years (as in 1856) the curve preserves a maximum about 10 

 a.m., before passing to the maximum at 7 a.m. There is no evidence of a gradual 

 shift or inversal (excepting when the mean movement deduced from several 

 days' observations is considered). The law is sometimes partially or wholly 

 inverted in one day— as in March 22, 23, 1854 : March 11, 12 ; 13, 14 ; 29, 31 ; 

 1856 : March 19, 21 ; 1859 : and March 5, 7 ; 8, 9 ; 18, 19 ; 24, 25, 26; 28, 29 ; 

 1864. 



The curves are so irregular and variable in their forms as to prove that the 

 electric currents producing these motions are in a state of disturbance, and flow in 

 various different directions, from day to day, in the month of March. 



It appeared to me of importance to examine to what extent the form of the 

 mean curves, in the months from November to February, represented the usual 

 diurnal law, and whether the inflexion or slight minimum apparent in these 

 curves near noon was observable in the ordinary daily movement, or was an 

 arithmetical result like the mean curve for March, and due to a combination of 

 different forms. For this end I projected the curves for each day in January and 

 February 1859 and 1864 (see Plates XLII. and XLI.) 



Variations for each Day in January and February 1859 and 1864. 



Commencing with 1864, a year of small disturbance (see Plate XLI.), the mini- 

 mum of easterly declination occurs generally in the months of January and 

 February at 6 h 30 m or 7 h 30 m a.m. ; rarely as early as 5 h 30 ra or as late as 8 h 30 m . 

 This regularity of the hour of minimum does not hold for the hour of maximum, 

 which varies betwixt 10 h 30 m a.m. and 5 h 30 m p.m. On eight days the maximum 

 occurred near 10 a.m. ; on nine days, near noon ; on fourteen days, near 2 p.m. ; 

 on eight days, near 4 p.m ; and on nine days, after 5 p.m. In the mean curve 

 for the month of January 1864, the inflexion near noon is scarcely visible ; the 

 maximum is near 2 p.m. In the mean curve for February of the same year, the 

 slight minimum before noon is better marked. 



An examination of the several daily curves shows that one-half at least have 

 a minimum or marked inflexion near noon ; and such inflexion or minimum does 

 not happen so frequently at any other hour, not an hour of minimum. Although, 

 then, the mean curve after the 7 a.m. minimum is a result of different variations, the 

 inflexion or slight minimum observable in the mean curve near noon is probably 

 the representative of a real and persistent phenomenon; and this will appear 

 more evident hereafter. 



If we now examine the curves for the same months in 1 859, a year of con- 

 siderable disturbance (see Plate XLII.), and in the mean curves for which there is 

 a secondary minimum at ll h 30 m a.m. (see 5th division of Plate XLIIL), we shall 



VOL. XXIV. PART III. 8 X 



