424 Prof. Kreil's Deductions from thejir.st Year's Observations 



creased force in increased temperature, and vice versa ; but 

 there is a correspondence between these numbers and the 

 monthly means of the inclination; the inclination having de- 

 creased and the times of oscillation increased from October 

 to March, whilst subsequently to March the inclination in- 

 creased, and the times of oscillation decreased. Such obser- 

 vations, continued with instruments somewhat differently con- 

 structed, but having the same purpose in view, will soon show 

 what part of this apparent connexion in the variations of the 

 intensity and inclination is due to the instrument, and what 

 to the phaenomena themselves. 



5. The influence of the moon on the magnetic condition of 

 the earth was examined in the same manner as was done 

 in the case of the Milan observations. Having corrected 

 approximately the observed horizontal intensities for tempera- 

 ture and the loss of magnetism sustained by the bar, there 

 appeared a confirmation of the results previously obtained, 

 viz. that the magnetism of the earth is stronger at the time of 

 the new moon than when the moon is full. The mean of the 

 whole body of the observations showed an intensity 



During the last quarter = 549'99 

 At the new moon = 548'79 



During the first quarter = 542'62 

 At the full moon = 54 I'll 



6. The difficulty referred to in par. 2. of freeing the ob- 

 servations of the horizontal intensity from the influence of 

 temperature, and the uncertainty of the correction for the de- 

 crease of magnetism in the bar, induced me to adopt a mode 

 of arranging the observations in reference to the inquiry re- 

 latively to the moon, by which the errors arising from both 

 these causes might be avoided. If the moon has an influence 

 on the magnetic condition of the earth, it must produce a 

 daily variation, masked by the greater effect produced by 

 the sun, but recognisable when the latter is eliminated. I con- 

 structed, therefore, tables, having for their argument the day 

 of the month, and for the titles of the several columns the 

 different distances of the moon from the magnetic meridian, 

 i. e. her magnetic horary angles. To simplify the calcula- 

 tion, I assumed that the moon passed the magnetic meridian 

 an hour earlier than the true meridian. I deducted from 

 each observation the monthly mean corresponding to the 

 solar time of the observation ; thus eliminating the influence 

 of the sun, considered as the cause of the regular diurnal 

 variation. This deduction being made, the remainders repre- 

 sent the sum of all other influences ; but as these numbers 

 were sometimes + and sometimes 1 augmented each by a 



