414 
THE ASTRONOMER ROYAL ON THE DIURNAL AND ANNUAL 
The estimate of the amount of irregularity which rendered it necessary to withdraw 
a sheet has been made throughout by the same person (Mr. Glaisher) ; and, through 
the years collected in the two last of the columns above, the records are of the same kind 
(photographic). I think therefore that, without going through a formal process claiming 
numerical accuracy for comparison of the tendency to magnetic disturbance in different 
years, these numbers may be presented as giving very important information. If they 
point to any cycle at all, it is one of 6 or 6^ years. 
These sheets being withdrawn, a pencil-curve was drawn by eye through the slightly 
disturbed photographic records ; and the ordinates, by means of which the diagrams 
now presented to the Society were constructed, were measured to the pencil-curve. The 
number used for forming the ordinates of the diagrams was the excess of the curve- 
ordinate for each special hour above the mean for the twenty-four hours ; and these are 
not here exhibited for every day, but their monthly means are taken, which monthly 
means themselves are the basis of further operations of grouping ; in one operation the 
same nominal months of different years are combined and their means taken, and in 
another operation all the months for each year are combined and their means taken. 
In forming from these means the ordinates of the diagrams, in the two first plates, exhi- 
biting the inequalities in the horizontal plane, it will be remembered that the vertical 
ordinate represents the change of northerly horizontal magnetic force (supposed to act 
on the northerly or marked end of the magnet), the measure upwards corresponding to 
increase of force ; and that the horizontal ordinate represents the change of westerly 
declination, exhibited as a westerly disturbing force on the same scale as the northerly 
force, by considering a deviation of 1' to be produced by a westerly force equal to 3 ^ 58 " 
of the whole northerly force. 
ANNUAL INEQUALITY. 
Following generally the order of the former paper, I will here advert to the question 
of annual inequality. The mean westerly declination for the six months called January 
in the six years was taken ; in like manner, the mean for the six months called February ; 
and so for all the months. These means are not comparable for the purpose of ascer- 
taining annual inequality, because they are subject to the influence of secular inequality. 
The mean annual change produced by secular inequality was found to be — 9'*2. The 
proportional part of this, corresponding to the interval of each month from January, was 
applied with sign changed to the monthly results. After smoothing the numbers by 
taking a second mean of adjacent numbers, the following values of Westerly Declination 
were found, applying, as regards secular values, to the month January in the mean of 
the six years : — 
January 21 12*9 
February 12*9 
March 13*4 
April 12*5 
May 11*8 
* 
