550 MR. W. ELLIS ON THE RELATION BETWEEN MAGNETIC DIURNAL 
The order of magnitude of the shorter intervals (minimum to maximum) and of the 
longer intervals (maximum to minimum) is the same for both sets of numbers, as before. 
The mean, as before, of the three intervals from minimum to maximum of magnetic 
© 
. y y 
effect is 4'27, and of sun-spot frequency is 4'07; the mean of the two intervals from 
. . ' y y 
maximum to minimum of magnetic effect is 7'23, and of sun-spot frequency is 7*40. 
y y 
Whole period of magnetic effect 1T50, of sun-spot frequency 11'47. 
These results generally agree closely with those deduced from Table III.* 
It will be noticed (Tables III. and IV.) that in two instances in which the declina¬ 
tion epoch of minimum was retarded, that of horizontal force was accelerated, giving 
a mean magnetic epoch according well with the sun-spot epoch. 
It has already been pointed out how closely the intervals between successive 
magnetic epochs agree with those between the corresponding sun-spot epochs, not¬ 
withstanding the difference in magnitude of the different intervals. As related to 
this it may be here further mentioned that if we add together the successive values 
of the numbers immediately following Table IV., to form complete periods, we get 
y y y y 
For magnetic periods.12 - 50 11'90 10'95 10'55 
For sun-spot periods.12 '30 12'00 11’20 10'50 
These numbers, whilst showing that the duration of the period has, for several 
periods, been steadily decreasing, exhibit in a yet more striking manner the corre¬ 
spondence between the two phenomena. 
One other circumstance may be mentioned, which is that according to the numbers 
of Tables III. and IV. (last column), the epochs of magnetic minimum and maximum 
appear, on the whole, to follow slightly the corresponding solar epochs. Further 
allusion will, however, be made to this point. 
The general circumstance that the diurnal ranges of magnetic elements are subject 
to an eleven year period, concomitant with that of sun-spot frequency, being thus, by 
the comparison of the smoothed curves of these phenomena, considered to be sufficiently 
well established, it seems now desirable to ascertain whether, by comparison of the 
actual monthly indications, the more fitful changes of the phenomena in any way also 
correspond. Before proceeding to explain how this has been done it is necessary to 
premise that whilst (as has been previously mentioned) the magnetic diurnal ranges 
are subject to an inequality of annual period, of considerable amount and large in 
comparison with the other changes to which they are subject, the solar spot energy 
* The mean periods deduced from. Tables III. and IY. are exhibited simply for the purpose of showing 
the accordance between the mean magnetic period and the mean sun-spot period as given by the series of 
observations discussed in the present paper, without at all implying any correction of the generally received 
mean value of the sun-spot period, or indeed stipulating for any definite length of period. 
