IN VEST I (i ATI ON 



OK THE 



INFLUENCE OF THE MOON ON THE MAGNETIC HORIZONTAL FOB( I 



The method pursued in the investigation of the lunar effect on the horizontal 

 force is, in general, the same as that explained in Tart III. of the discussions of 

 the Girard College observations. The process may be briefly recapitulated as fol- 

 lows: Each horizontal force observation, after it had been corrected for the effect 

 of difference from the standard temperature and for progressive change, the dis- 

 turbed readings being omitted (as fully explained in Part IV.), was marked with 

 its corresponding lunar hour; the observation nearest to the time of the moon's 

 upper transit over the true meridian of the observatory was marked 0\ that nearest 

 to the lower transit was marked 12\ and the observations between, for western and 

 eastern hour angles of the moon, were marked with the proper lunar hour by inter- 

 polation. In the hourly series where thirteen observations arc recorded in twelve 

 lunar hours, that observation which is nearest midway between any two consecutive 

 lunar hours was omitted. Each observation and reduced reading thus marked with 

 its corresponding lunar hour was subtracted from the monthly normal belonging to 

 its respective hour, and these differences were set down in tabular form, arranged 

 according to lunar hours and keeping each monthly result separate for future com- 

 bination. Let n = any normal belonging to any reduced reading r, the following 

 tables contain the mean monthly values of the differences n — r; a positive sign. 

 therefore, indicates greater force, a negative sign less force than the normal. It 

 need hardly be repeated that in the original record of the horizontal force increasing 

 numbers denote a decrease of the force. The greatest possible difference is 33, the 

 number of scale divisions, which, according to the criterion, separates a disturbed 

 from an undisturbed observation. For the formation of these differences which 

 amount to more than 22,000, the manuscript tables of the reduced record were 

 used: these tables have already been referred to in the preceding Part IV. 



The units in which the differences n — /• are expressed are scale divisions, one 

 division being equal 0.0000365 parts of the horizontal force, or equal to 0.000152 

 in absolute measure, the mean X being = 4.178 (in units of grains and feet). 



The lunar effect on terrestrial magnetism being exceedingly minute, the process 

 required for its elucidation is proportional!) delicate; all the regular and irregular 



( 63 ) 



