356 Results of Makerstoun Observations, 1844. 



Table XXII. has been formed from the daily means in scale divisions by the following formula : — 



/=(n-500) 0-000140, 



where /is the mean in the Table, n is the mean in scale divisions, 500 scale divisions being taken as the zero ; 

 the corresponding zero for the same adjustment in the end of 1843 was 484*93 ; in order, therefore, to render 

 the means for 1844 comparable with those for 1843, 15-07 x 0-00014 = 0-002110 must be added to the 

 former, 0-001000 being subtracted from the latter, and the differences multiplied by 1-316. See Introduction, 

 Bifilar Magnetometer. 



The mean value of the horizontal component from Table XXII. = 0*003605. 



Secular Change. — When the monthly means at the foot of Table XXII., and the monthly means for 

 1843 (as corrected, p. 231, line 6, 1843), are rendered comparable, as indicated after Table XXII., we have 

 the following monthly means for the two years, with the corresponding secular changes : — 



1844, 0-00 



1843, 0-00 



Secular | Q 

 change, J 



Jan. Feb. March. April. May. June. July. Aug. Sept. Oct. Nov. Dec. 



4773 4955 4771 5086 5789 6535 6694 6486 6470 6454 6850 7322 



0932 0917 0661 0903 1848 2328 2214 2307 2690 2903 3664 4183 



3841 4038 4110 4183 3941 4207 4480 4179 3780 3551 3186 3139 



The value of the secular change, as deduced from the monthly comparisons, does not vary greatly till 

 October, after which it diminishes considerably. 



The mean secular change corresponding to 1st July 1843, till 1st July 1844, . = +0*003886 



A comparison of the mean value of the component for 1844, with that for 1845, shews that the amount 

 of the secular change has continued to diminish, the mean corresponding to 1st July 1844, till 1st July 1845, 

 being only +0-001436. As it is desirable, in some case, to eliminate the secular change in order to exhibit 

 other laws with more distinctness, the mean of these two determinations has been taken as the secular change 

 for 1844, or, 



The mean secular change corresponding to 1st January 1844, till 1st January 1845, . = +0*002661 



This quantity is nearly the same as that obtained from a comparison of November and December 1843, 

 January and February 1844 with November and December 1844, January and February 1845. 



Annual Period. — From the means at the foot of Table XXII., the horizontal component does not vary 

 greatly in value from January till April, but it increases rapidly from April till June, after which it is nearly 

 constant again .till October, increasing again in November and December. It appears, therefore, that the secular 

 increase is neutralized for three months after the solstices, and augmented in the three months after the 

 equinoxes. In order to shew the law of the annual variation independent of the secular change, the latter has 

 been eliminated from the means at the foot of Table XXII. by the application of the correction — n 0-000222, 

 where n is the number of the month after January : the means, thus corrected, are — 



Jan. Feb. March. April. May. June. July. Aug. Sept. Oct. Nov. Dec. 



0*00 I 2663 2623 2217 2310 2791 3315 3252 2822 2584 2346 2520 2770 



These means exhibit the law already concluded from the observations for 1842 and 1843 — namely, that 

 the horizontal component is a minimum near the equinoxes, and a maximum near the solstices. The secular 

 increase is diminishing in value, and therefore the application of a constant correction from month to month 

 cannot be quite accurate ; any other mode of elimination, however, could only have affected the values of the 

 maxima and minima ; the epochs would remain as stated above. 



It should be remarked, with respect to this law, that the variations of the magnetic dip are best repre- 

 sented in this latitude by the bifilar magnetometer : this is not the case, however, in low latitudes ; in these, 

 the variations of intensity are best represented in the horizontal component ; if, therefore, the magnetic dip 

 and magnetic intensity have each an equally well-marked annual period with different epochs, the same 



