588 MR J. A. BROUN ON THE 
In order to diminish the effects of the larger disturbance, which require longer 
periods for their repetition according to law, the means for each three hours 
have been taken; these are projected as means for the middle hours of each 
three (Plate XL.). It will. be remembered, in considering these curves, that 
the successive points are obtained from observations made at the same hour 
in diferent years. We conclude from the projections,— 
46. That the decennial variation does not follow exactly the same law at 
different hours. Thus there is some variation in the exact epochs of minimum 
near 1856°5, and of the maximum near 1860-0, according to the hour con- 
sidered. The maximum occurs earliest (1859°5) in the curve for 93 P.m., and 
latest (1860-2) in that for 94 a.m. Still more marked differences appear in the 
descending branches of the curves (after 1860). A secondary minimum appears 
in the curves for 33 to 93 A.M. inthe years 1862 and 1863, which disappears in the 
curves after noon, the minimum occurring then in the year 1861. A secondary 
maximum occurs in all the curves near 1863:0, being least marked in those for 
noon and midnight. These differences are not due to large accidental disturb- 
ances at any given hour, since they change gradually from hour to hour. 
47. When we compare the magnitude of the mean disturbance for each 
three hours at the maximum and minimum of the curves (Plate XL.), we obtain 
the following ratios :— 


llipm.to ld am, O.14t = 248 
2h a.m. to 44 ,, : : : . : : 2h = 2:58 
De 4,t to TE, : $ : ; : : ah = 2-06 
Sh ,, to lO& ,, si = 1-64 
11k ,, to ld Pm, z : : ; , : a =e ul 
2k pm. to 44 ,, : . 3 5 : 4 = = 181 
Se i ys) to We |. : : : ‘ : a = 2°05 
8 ,, to10Z ,, : ‘ ; : ; : — = 252 
All the hours, : 4 : : , ‘ : sak = 1-86 
48, In a similar way we obtain the ratios of the maximum to the minimum 
disturbances in the diurnal variation for each year. Taking the maximum and 
minimum for three successive hours, and for one hour only, we have,— 


