468 FIRST REPORT UPON MAGNETIC WORK

right to left and back again from left to right, been observed. At
some stations, however, a little over half a swing has been obtained.
Comparing the time interval between the two extreme positions, i. e.,
half a swing, at various stations we are brought face to face with
another remarkable fact, that the time intervals between the extreme
positions of the needle are of different lengths in different parts of
the earth. To illustrate: At London, Paris and Rome the time in-
terval between dates of extreme positions of the needle is about 230-
240 years, while for stations in the Eastern states of this country it is
on the average about 150 years. If we take into consideration all the
facts at present known to us with regard to the secular variation, we
find that it is not possible to explain all those facts on the assumption
that there is a secular variation period common to all parts of the
earth of about 300-500 years in length. The indications are that for
a common secular variation period we must have a much longer one
than 300-500 years. But if this is so, it means that the secular varia-
tion is a far more complicated matter than generally supposed. Be-
sides the main swing as described above, there are a number of minor
swings whose periods are not as yet definitely known. These minor
swings have the effect of slightly altering the annual change due to the
main secular variation.

Fig. No. 4 illustrates graphically the change in the magnetic
declination for various points in the northern hemisphere. I have
selected such stations as would be typical of the regions represented
by them. It will be seen that the stations encircle the globe. This
one diagram exhibits at a glance all the characteristic features of the
secular variation of the magnetic declination in the northern hemi-
sphere as at present known. It is presented here for the first time,
the data having been collected from -various sources. With the aid
of the table (No. VI) the meaning of the curves will be readily
understood. Thus, for example, selecting the date 1800 and running
the eye along the horizontal line marked 1800 until it intersects the
London curve, let us say, casting the eye now upward from this point
of intersection along the vertical line, we find that the declination of
the needle was a trifle over 24° west. For Paris, the observations