INVESTIGATION 



SOLAR-DIURNAL VARIATION, AND OF THE ANNUAL INEQUALITY OF THE 

 HORIZONTAL COMPONENT OF THE MAGNETIC FORCE. 



THE discussion of the diurnal and annual variations of the horizontal force is 

 based on the resulting montlily normal values for each observation hour as given in 

 the preceding part (IV.), in which the horizontal force has been discussed in relation 

 to the ten or eleven year period, and which also contains the investigation of the 

 disturbances; in the same part all necessary statements are given relating to the 

 instrumental data and the absolute values of the horizontal force. 



The normals, as has been shown, are referred to a uniform standard temperature ; 

 they are corrected for irregularity in the progressive change, and are necessarily 

 freed from all the larger disturbances. The use of the normals instead of the sim- 

 ple means of the readings (corrected for difference of temperature) will insure 

 greater regularity in the variations of the horizontal force, now under consideration. 



The diurnal variation requires an arrangement of the five year series of monthly 

 normals according to the months of the year and hours of the day; in general, the 

 method of interpolation for an occasional omission in either a month or hour, is the 

 same as that used in Part II. of the discussion of the Girard College observations ; 

 there is, however, this difference in the tabulation of the monthly values, that in 

 the present case the results are consolidated in a five years' arrangement, and in 

 consequence the year commences with the month of July. This arrangement was 

 preferred, particularly since it was found desirable to make no use of the observa- 

 tions in the first month of the series. 



Tabulation of monthly normals for each observing hour and each observing year, 

 beginning and ending with July. The individual values are taken from Table 

 VII. of the preceding Part IV. 



After applying the corrections of 19 scale divisions to the normals for January, 

 1841, and of +8 scale divisions to those of June, 1841, to allow for defective num- 

 ber of observations in these months, a further correction of + 68 scale divisions was 

 applied to all values between July, 1840, and May, 1841, inclusive, and of + 60 to 

 all values between July, 1840, and December, 1840, inclusive, to allow for defects 

 in the regularity of the progressive change, thus making the total correction for 

 the latter months = 128 scale divisions. The above corrections, when divided by 5, 



(35) 



