420 Prof. Kreil's Deductions from thejirst Year's Observations 



= 36*01 ; in April 33*01 ; and least in December, = 8-41, if 

 the observations at 8 a.m. and 1 p.m. are compared, or 13*06 

 if those at 1 p.m. and 10 p.m. are compared. The secular 

 decrease appears to have been unusually great this year ; if 

 the apparatus showed correctly, it amounted from July 1839 

 to June 1840, to no less than 17*70 divisions of the scale, 

 = 8' 01"*9; but as in October, 1839, the thread by which the 

 needle was suspended broke, this result may be inexact. The 

 observations of the following month, however, show a still 

 greater decrease. If we take the mean of all the observations 

 from 5 a.m. to 10 p.m., we obtain the following declinations : 



1839. November. Declination = 417*80 

 December. = 416*88 



1840. January. = 415*80 



February. = 414*74 



March. = 410*35 



April. = 407*50 



May. = 402-87 



June. = 399-30 



The declination, therefore, decreased in these eight months 

 18-50 scale divisions, or 8' 23"*7. 



2. The horizontal intensity observed with the bifilar had 

 in the winter months its 



Minimum • = 544*42* at 11 a.m. 



Maximum = 564*83 at 10 p.m. 



Difference = 20-41 



In summer, its 



Minimum =443*63 at 10 a.m. 



Maximum =488*90 at 8 p.m. 



Difference, = 45*27 



The minimum was earliest in August, when it took place 

 at 9 a.m., and in June, when it occurred at 9 h 30 m a.m. ; it was 

 latest in December, at 1 p.m.; and in January was at ll h 45 m 

 a.m. The maximum appears to fall in winter, i. e. in December, 

 January and February, during the hours of the night ; it was 

 earliest, at 8 p.m. in July, November, April and June. The 

 following are the means of all the observations between 5 a.m. 

 and 10 p.m. 



* A scale-division in this apparatus corresponds to 18*5757 seconds 

 of arc, or 1 -- S ^ TT of the whole horizontal intensity; but it should be re- 

 marked, that since the instrument was set up in May, 1839, the magnetism 

 of the bar has not been examined, in order that the series might not be 

 broken. 





