64 



DISCUSSION OF THE INFLUENCE OF THE MOON 



deviations arising from other sources must first be eliminated. In the method, as 

 indicated above, the magnetic disturbances (as far as they could be recognized as 

 such), the diurnal and annual solar variation, as well as the eleven (or ten) year 

 inequality and secular change, are all eliminated, leaving numbers fitted for the 

 lunar research. 



The readings taken in the month of June, 1840, have not been used in this 

 discussion (these had likewise been rejected in the two preceding parts), on account 

 of the imperfect manner in which the allowance for the progressive change could 

 only be made at that time. For the lunar hour 21 in July, 1840, the number of 

 differences is so small that the mean had necessarily to be reduced; one-fourth of 

 its amount was set down in the table. In January, February, and March, 1843, 

 the observations were discontinued, excepting a single dady reading. These months, 

 therefore, do not occur in the lunar discussion. 



The number of observations used are distributed over the several months and 

 years, as shown in the following table. 



Tables 111., IV., V., VI. and VII. contain the monthly and annual means of the 

 lunar diurnal variation for the years 1840 to 184o. The numbers are expressed in 

 scale divisions. 



'■ f Commencement of the hourly scries. 



