22 



DISCUSSION OF THE HORIZONTAL COMPONENT 



subject to the same inequality of about eleven years as the declination, and tin- 

 means of investigation will be analogous to those used in Part I of this discussion. 

 For greater convenience, the preceding monthly normals were united into annual 

 means and the results put into an analytical form, using Bessel's function applicable 

 to periodical phenomena, and determining the numerical quantity by the application 

 of the method of least squares. 



In the following table of the regular solar diurnal variation of the horizontal force 

 the means for 1842-43 depend only on nine months of observation; the correction 

 given to refer them to twelve months of observation depends on the mean difference 

 between the results of the same nine months and twelve months of the preceding 

 and following year ; this correction is nearly constant and the same within one scale 

 division for the adjacent years. In the second corrected column for 1842-43 the 

 effect of the annual inequality is thus eliminated. In the year 1843-44 the results 

 from nine months of observation at the odd hours were reduced to twelve months 

 by means of corresponding differences in the series of even hours ; thus (omitting 

 the minutes) at hour 2, mean of 12 months = 1006, mean of 9 months = 1028; at 

 hour 3 for the same 9 months, mean = 1026, or 2 divisions less; at hour 3 for 

 12 months the mean is therefore 1004, and the same result is found by comparing 

 with the following hour 4 ; the mean is given in case of a difference in the two 

 results. 



