M. KreiPs Magnetic Observations* 247 
than they had done in 1837, as is shown by the following 
comparison : 
Jan. 
Feb. 
Mar. 
April. 
June. 
July. 
Aug. 
Sept. 
Oct. 
Dec. 
1837. 
25 
13 18 
22 
6 22 27 
2 
2 28 
25 
14 16 
18 
14 19 
1838. 
28 
16 21 
17 
11 29 30 
4 
12 34 
23 
14 15 16 
17 
20 28 
These coincidences would show a new fact, namely, the 
periodical nature of these phaenomena, if equally considerable 
disturbances had not occurred, (ex. gr. those of the 12th, 14th, 
and 15th November 1837, and of the 17th January 1838,) 
without any corresponding disturbances being traceable in 
other years. The subject requires to be further elucidated 
by continued observation. 
Another point to which the attention of the observer should 
also be directed is, the symmetrical arrangement of these dis- 
turbances in the same year, many of them being nearly six 
months apart, for example : 
In the year 1836 and 
1837 the disturbances on the 22nd April and 18th Oct. 
1838 -- — 17th Jan. and 12th July, 
do. — — 21st Feb. and 23rd August, 
do. — — 29th April and 31st October, 
do. ~ — 4th June and 5th December. 
18. All hours of the day do not appear to be equally fa- 
vourable to the development of this phaenomenon, at least its 
first indications occur much more frequently in the evening 
than in the morning hours. If we take from the 105 dis- 
turbances marked in our day-book, those which were ob- 
viously continuations of perturbations which had begun on 
preceding days, as well as those which were first remarked 
during the intervals between the regular hours of observa- 
tion, then the first indications of 
19 were 
at 
20^ 
0' 
1 
at 
22 
30 
9 — - 
at 
1 
0 
21 — 
at 
4 
30 
16 — 
at 
7 
30 
11 
at 
11 
0 
Here it is to be remarked, that the observation hour, 11, 
precedes the longest interval namely, the night, and that the 
greater number of the more considerable disturbances ex* 
tend over several hours or even days; therefore perturbations 
which may have begun in the early hours of the night ought 
