157 
of Edinburgh, Session 1875 - 76 . 
period is 10-45 years ; but that it appears to undergo a variation 
between 8 and 12J years in an interval of 42 years. The small 
maximum of 1797*7, if a true result, may be expected to repeat 
itself at some future time, a result which could not fail to aid in 
the search for the cause of these variations. 
The author shows, that according to the long period of 42 years, 
a maximum should have happened in 1776 ; but that year Dr Wolf 
has concluded to be one of minimum sun-spot frequency. That 
1776 was really a year of maximum is confirmed by the observa- 
tions of Van Swinden, who, it is shown, appears to have been the 
first to obtain a variation due to the decennial period, and to have 
pointed out the appearance of a law : it is also confirmed by the 
observations of Cotte at Montmerency. 
The ratio of the ranges of the diurnal variation in the years 
when it is a maximum to that in the years of minimum, is com- 
pared for different parts of the world, and found nearly the same 
in both hemispheres. It is also found that the law of the diurnal 
variation is the same in the year of maximum as in the year of 
minimum. The author concludes that the increase of the diurnal 
variation is not due to a different cause from that which produces 
the variation at the minimum, and that this cause acts when there 
are no sun-spots in the same way as, though with less intensity 
than, when the latter have their maximum frequency and area. 
The magnetic variations are therefore not due to the sun-spots ; the 
latter appearing only when the common cause produces diurnal 
variations having at least two-thirds of the maximum amplitude. 
The results derived from the sun-spot area are compared with 
those from the magnetic observations. While a general agreement 
is found in the decennial variations from year to year, it is evident 
that the attempt to calculate the amplitude of the diurnal varia- 
tion from the sun-spot frequency (as has been done by Dr Wolf) 
must give results frequently deviating widely from the truth, as 
might be expected from the previous conclusion. 
The decennial period of disturbance of magnetic declination at 
Trevandrum, deduced from hourly observations in the eleven years, 
1854 to 1864, is next considered. It is found that the mean disturb- 
ance at each hour of the day shows the decennial period; but that 
the range of the mean value, from the minimum to the maximum 
