Horizontal Component of Magnetic Force. 231 



24 observations minus 9 observations. 



Jan. 





Feb. 



March. 



April. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Bee. 



0-0000 1 



-36 



-61 



-69 



+ 24 



-24 



+ 13 



+ 29 



-34 



+ 68 



+ 46 



-06 



-28 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



2769 



2682 



2753 



3044 



3206 



3784 



4179 



These quantities being applied as corrections to the monthly means, foot of Table VIII., we obtain t]ie 

 following : — 



Jan. Feb. March. April. May. 



0-00 I 1708 1697 1502 1686 2404 



From these it appears that the horizontal component diminishes from January to March, increases from 

 March to June, diminishes again slightly from June to between July and August, and then increases again 

 till December ; the whole range being 0-002677, and the increase of force from January to December being 

 0" 002471- The annual period can only be rendered distinct when this secular change is eliminated. It has 

 been found by a comparison of the monthly means of 1843 with those of 1842 and of 1844, that the mean 

 annual change during the 12 months of 1843 is 0-002826. When proportional parts of this are subtracted 

 from the means after January, we have — 



Jan. Feb. March. April. 



0-00 I 1708 1462 1031 0980 



We have here a well marked annual period, consisting of maxima near the solstices, and minima near the 

 equinoxes. The maximum is rather greater in the winter than in summer, but the difference is less than the 

 effect of half a degree Fahrenheit on the bifilar magnet. 



The range of the annual period is 0-000728, or about half the mean diurnal range for the year. 



It may be desirable to examine the annual period for the year 1842 in a similar manner. If we reduce 

 the monthly means of four observations, Table IX., Abstracts for 1841-2, to parts of force by the formula 



/= {n- 530) 00001021 + 0-003000 



where n is the mean in scale divisions in the Table referred to, we obtain the following values of/, which may 

 be rendered comparable with the means, Table VIII., by subtracting 0-002000. The second line below con- 

 tains the corrections for each month of the 4 observations to the 24 as deduced from the observations for 1844, 

 and the third line contains the corrected means. 



May. 



.June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



1462 



1592 



1269 



1105 



1160 



1087 



1429 



1589 



000 

 000 

 000 



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



0960 0994 1188 0840 1646 1803 1743 1836 2016 2488 2772 3270 



-0057 -0045 -0003 +0294 +0245 +0309 +0352 +0390 +0318 +0187 +0068 -0020 



0903 0949 1185 1134 1891 2112 2095 2226 2334 2675 2840 3250 



If the secular change be assumed the same for 1842 as for 1843, and if this be eliminated, we have — 



Jan. 



Feb. 



March. 



April. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



0-000 1 903 



714 



714 



428 



949 



935 



682 



578 



450 



556 



485 



660 



These means shew the same law as the means for 1843, with more irregularity, it is true, but this may 

 be accounted for by the fewness of the observations, and the possible inaccuracy of the corrections. 



