Magnetic Declination. 221 
The quantities in brackets are the means of the three preceding and of the three succeeding numbers ; they 
are, therefore, the weekly means, and they may be considered as the approximate westerly declinations on the 
Sundays, whose place they occupy, and on which no observations were made. 
The means for the first week in January are means of the four daily observations in that week corrected by 
— 0 ,- 84. The means in the second week are the means of the eight daily observations in that week corrected by 
— 0'-14. These corrections were obtained by comparing the observations at the four and eight corresponding hours 
in 1844 with the observations at the nine hours corresponding to the observing hours in 1843. 
Annual Period of Declination. 
Differences of the mean westerly declination, deduced from the observations made in 1844 at the hours 18, 
20, . . . 10, from that deduced from the 24 hourly observations made on each day (excepting Sundays) in 
that year. 
24 Observations minus 9 Observations : — 
Jan. Feb. March. April. May. June. July. Aug. Sept. Oct. Nov. Dec. 
-0'-33 -0'-47 -0'-75 -0'-49 -0'-85 -0'-58 -0'-56 -0'-68 - 0'-75 -0'-76 -0'-73 -0-35 
The variation of these differences is not sufficiently great to affect the character of the annual period, as de- 
duced from the monthly means at the foot of the previous Table. When these quantities are subtracted from the 
monthly means of the nine observations in 1843 we obtain the following : — 
Jan. Feb. March. April. May. June. July. Aug. Sept. Oct. Nov. Dec. 
25'-67 24'-99 24'-26 23'-95 23'-27 25'-22 23'-63 22'-35 20-'92 2L-77 18'-88 19'-24 
From these we see that the westerly declination diminishes very regularly till May, increases considerably 
from May to June, diminishes again regularly to November, with the exception of an increase in October, which, 
however, is nearly made up for by the more rapid diminution in November ; in December there is a slight increase. 
It will be found, Table III., that the retrogressive or secular change of westerly declination has a value of 
about 6' in one year ; if we add proportional parts of this to the different monthly means, the annual period will 
be more distinct. The monthly means will then be, when the secular change is thus eliminated : — 
Jan. Feb. March. April. May. June. July. Aug. Sept. Oct. Nov. Dec. 
25-67, 25-44, 25-26, 25-45, 25-27, 27-72, 26-63, 25-85, 24-92, 26-27, 23-88, 24-74. 
These means are still affected by the varying torsion force of the suspension-thread, — the error due to this 
cause cannot be eliminated ; it is probably greatest in the mean for June, in which month the suspension-thread 
broke. An examination of the variations of the plane of detorsion in several of the other months (see the notes to 
the Daily Observations of Magnetometers) will shew that the error from this cause is generally small ; keeping this 
in view, if a curve be run freely through the projected means for the 12 months, there seems to be a minimum of 
westerly declination about March, and a maximum perhaps in July or August. Having an eye to the result for 
1844, and also to the result for the horizontal and vertical components of intensity, I am inclined to believe that 
the annual period of magnetic declination will be found to consist of a minimum at the vernal and a maximum 
at the autumnal equinox. 
The means for the four seasons are — 
Spring. Feb., March, and April, mean ...... 25 - 38 
Summer. May, June, and July, mean . . . . . . 26 54 
Autumn. Aug., Sept., and Oct., mean . . . . . 25 - 68 
Winter. Jan., Nov., and Dec, mean ...... 24- 76 
These means indicate the maximum in Summer and the minimum in Winter ; but there is little doubt that 
the mean for June is too high, and those for September, November, and December - , too low. 
After correcting the monthly means of the nine observations to the monthly means of twenty-four observations, 
the mean westerly declination for the year 1843 is found to be 25° 22''85. 
MAG. AND MET. OBS. 1843. 
