138 MR BROUN ON THE RELATION OF THE VARIATIONS OF THE 
or positive, there will be sudden and considerable daily decrements or increments 
of the vertical component. If the effect of these disturbances be equally positive 
and negative, it may be expected that, in the mean of a sufficient number of ob- 
servations, they will destroy each other, or, if almost wholly negative, that they 
will be distributed equally over any period for which the,law is desired. If, 
however, we expect consistent and regular results, from short series of obser- 
vations, especially when the range of the variations for the resulting law is small, 
it is evidently desirable to eliminate disturbances as far as possible. To do so com- 
pletely, requires a knowledge of the observations which should be classed as dis- 
turbed, a matter of difficult attainment, as disturbances have various magnitudes, 
and we have to distinguish between variations which are periodic, and those 
which are not so; the latter constituting what are distinctively termed disturbances. 
Without doubt, disturbances seem to obey an zrregular periodic law, that is to 
say, they occur more frequently at certain hours than at others; but there are 
whole months in which they are not at all apparent, and in which, each day’s 
observations represent equally, in range and periods of maxima and minima, the 
diurnal law; thus, in the months of June and July 1844, each day’s observations 
of the horizontal component when projected, exhibit, with a few slight excep- 
tions, a complete representation of the monthly curve, affected, however, to some 
extent by the varying hour angle of the moon. 
4. As it is only the larger disturbances which affect the periodic result to a 
marked extent, the complete elimination of disturbances is fortunately not of much 
moment; the methods which I have adopted for the elimination of these will be 
be stated afterwards; it will be found, however, that I have depended as little as 
possible on these eliminations. 
5. The results as yet obtained for the variations of the vertical component 
of the earth’s magnetic intensity are confined, as far as | am aware, to determi- 
nations of the mean period of diurnal maxima and minima. These may be obtained 
from a good instrument, with moderate exactness, in spite of considerable imac- 
curacy in the temperature correction, as the mean diurnal variation of temperature, 
in well closed rooms and boxes, will not exceed a few degrees of Fahrenheit. The 
Toronto observations for 1841 and 1842, give consistent results, indicating 
A principal maximum at 6" Toronto mean time. 
A principal minimum at 14° 
A secondary maximum at 20° 
A secondary minimum at 22” 
_ There are only three months of the twenty-four indicating a single maximum and 
asingle minimum, namely, the months of November 1842 and September of each 

