EARTH’S MAGNETISM TO THE SOLAR AND LUNAR PERIODS. 14] 
ponent which becomes the principal minimum in swmmer. For the horizontal 
component the 22" minimum is ¢#e minimum throughout the year ; for the vertical 
component the 5" maximum is ‘hé maximum throughout the year. 
10. Perhaps the most curious fact in connection with the vertical component 
is that of the annual variation of the diurnal range. It has always been ima- 
gined, I believe, that the diurnal range of all the magnetical elements increased 
from winter to summer. This has been shewn to be the case for the horizontal 
component (No. L, 6*), and will afterwards be shewn to be true for the magnetic 
declination. It is not so, however, for the vertical component, as may be seen by a 
glance at the six diurnal curves projected (Curves, No. 1, Plate VI.). The transition 
in form and range is evidently worst exhibited by the mean for February and 
November ; this, however, and other irregularities, may probably be due to dis- 
turbance. The elimination of these (which I have not at present attempted), or the 
observations of other years, must decide this. When the range for each month is 
projected (Curve No. 2, Plate VI.), it is at once evident that the diurnal range is 
least at the solstices, and greatest at the equinoxes. The mean of the ranges for 
_ January and December, and also the mean for June and July (the solstitial months), 
is about 000028, the whole vertical component being unity, while the mean for 
each of the couples of equinoctial months, namely, of March and April, and of 
September and October, is about 0:00068. 
11. It might have been expected that this curious variation of the ranges 
would shew itself more or less in the ranges for the horizontal component. If we 
refer to No. I. of these results (6), page 101 of this volume, and to the projected 
curves, No. 2 of that series (Plate III.), we shall find this actually the case, although 
in that place it was supposed that the variations in the regular increase of the 
ranges might be due to disturbances. These facts seem to point to a difference in 
the modes and causes of increase of the diurnal range for the magnetic dip, and 
for the total magnetic intensity ; the diurnal ranges of the latter seeming to obey a 
law which is equally related to the two solstices, and also to the two equinoxes, 
a circumstance in favour of the annual period previously announced (No. I., 10). 
12. The mean values of the vertical component, at 21" and 0", are nearly 
equal to the mean for the year, but no single hour, as for the horizontal component 
(No. I, 8), indicates the mean for each month. 
13. Proceeding in the order of No. I., I should now consider the annual period. 
» The results for the different years are discordant. This, it is my opinion at present, is 
due to an insufficiency in the temperature correction, which will be found alluded to 
elsewhere.t The results of the years 1842 and 1843, the latter more strongly, indi- 
enke a period similar to that found for the horizontal component, namely, maxima of 
* Reference to Results of the Makerstoun Observations, No. I., p. 99 of this volume. 
+ Introduction to the Makerstoun Observations for 1843. 
VOL. XVI., PART II. 2N 
