224 
DR. C. CHREE: MAGNETIC DECLINATION 
Table XIII.—Difference in Minutes in Times of Occurrence of Maxima in Ordinary 
and Disturbed Days (+ denoting Later Occurrence in Ordinary Days). 
Whole 
year. 
Winter. 
Equinox. 
Summer. 
h. 
m. 
h. 
m. 
h. 
m. 
h. 
m. 
24-hour term. 
+ 1 
50 
+ 1 
44 
+ 1 
52 
+ 1 
38 
12-hour „. 
-1 
0 
- 1 
18 
-0 
20 
-1 
31 
8-hour ,,. 
-0 
49 
- 1 
13 
-1 
0 
-0 
39 
6-hour ,,. 
+ 0 
48 
+ 0 
12 
+ 1 
27 
+ 1 
21 
We see how, as it were, by filtering out disturbed days we obtain a gradual retarda¬ 
tion in the phase. Some of the Astronomer Royal’s quiet days are decidedly less 
quiet than others. Supposing we adopted a still higher standard should we observe 
even more retardation ? On the other hand, our disturbed days vary much in the 
degree of disturbance, and very considerably in its type. Is there an advance in 
phase irrespective of the type of disturbance, and is the advance greater in highly 
disturbed than in moderately disturbed days 1 Obviously there are many further 
questions arising out of the above result. Evidently a study of the phase constitutes 
a method of considerable delicacy for advancing our knowledge of disturbances ; but 
a very obvious consideration is that in order to obtain eveu approximately smooth 
diurnal inequalities from highly disturbed days one must deal with a long period of 
years. When diurnal inequalities are not smooth, merely “accidental” irregularities 
may introduce a fictitious element into the Fourier coefficients, especially those of 
shorter period. 
The results reached emphasise one difficulty in the way of an exact inter- 
comparison of stations. Even if all stations suffered equally from disturbance, their 
published data would not be strictly comparable unless the standard for omitting 
disturbed days when forming the diurnal inequality were absolutely uniform. Under 
existing conditions all that we do know is that the standard is not uniform, and that 
there is no obvious way of making it so, whilst there is every reason to believe that 
the incidence of disturbance is widely different. 
We have already referred to the different effects produced by disturbance in the 
angles and a 2 . Possibly this difference may be connected in some way with 
another, viz., that whereas a x increases as we pass from summer to winter, a 2 
diminishes. Summarising the results we have in short:— 
a x largest in disturbed days, in winter, in years of sunspot minimum, 
aj least in quiet days, in summer, in years of sunspot maximum. 
a 2 largest in quiet days, in summer, in years of sunspot minimum. 
a 2 least in disturbed days, in winter, in days of sunspot maximum. 
