MAGNETIC DECLINATION, AT DIFFERENT PERIODS OF THE DAY. 
637 
December 31, 1846; and the mean or middle year at Hobarton is from July 2, 1845 
to July 1, 1846; the mean of the results in the same month and at the same hour in 
each year of the series (three years in the one case, and five years in the other) being 
taken as the declination-values corresponding to the mean year; which may thus be 
regarded as a type of the actual changes taking place in the declination in the course 
of a single year, at the period when the observations were made ; but with the ad- 
vantage that the various declination-values in the mean or typical year are derived 
in the one case from three, and in the other case from five years of continuous and 
intercomparable observations, instead of from one. 
The dark vertical lines in Plate XXVII. exhibit the actual annual range of the de- 
clination at each of the observation hours. The figures on the left of the lines (on the 
left in the Toronto figure, and accidentally placed on the right of the lines in the 
Hobarton figure) show the positions in the annual range at each hour of the mean 
declination in the different months as derived from observations at that hour only; 
the months being distinguished by numbers, according to their order of succession, 
from 1 to 12, commencing with January. The broken line MM shows the mean de- 
clination of the whole year, viz. the mean declination of all the months and all the 
hours. The dotted line DD represents the mean diurnal variation in the year, or 
the mean variation of the declination at the different hours of the day and night: 
it is drawn through and connects the points of mean declination in the year at each 
of the hours respectively. The scale of declination-value in this Plate is reduced for 
convenience to half the dimensions of that of Plate XXVI. ; in Plate XXVII. an inch 
is the equivalent of two minutes of declination, and in Plate XXVL of one minute. 
The declination in both plates is that of the north end of the magnet, or that end 
which in the middle latitudes points towards the geographical north. 
On an intercomparison of the observations represented in their mean monthly values 
at each hour in Plate XXVII., it is at once obvious that they are affected by three 
distinct and easily distinguishable variations, two of w^hich are periodical, and one is 
secular. 1°. If the mean declination of the year at the different hours of the day and 
night had an uniform value, the line DD would be a straight line, and there would 
be no mean Diurnal Variation, or variation whose period is a day : — but DD is very 
far from being a straight line. 2°. If the mean declination at any particular hour of 
the day were of uniform value in all months of the year, the dark vertical line cor- 
responding to that hour Mmuld be reduced to a point instead of a line, and there 
would be no Annual Variation, or variation whose period is a year, at that hour : — but 
the dark vertical lines are very far from being points at any hour of the day or night. 
3°. If, on comparing the mean declination observed in any particular month and at 
any particular hour in one year with the mean declination in the same month and 
hour in preceding or succeeding years, the values were found uniform, or presenting 
at most such differences only as might reasonably be ascribed to observation errors, 
or to what are usually called magnetic disturbances, we might infer that the line 
4 N 
MDCCCLI. 
