MAGNETIC DECLINATION AT KEW AND NERTSCHINSK. 
233 
whole body, an advance might be made towards an approximate estimate of the degree 
in which the disturbing action prevailed in different parts of the globe. In the prac- 
tical application of this scheme of first or primary analysis it was found that, provided 
the selected separating value at each station were such as to place in the category of 
disturbed observations a proportion equivalent to between one and two tenths of the 
whole body of the observations, small alterations within these limits occasioned no 
significant alterations in the derived diurnal progression either of the disturbed or of 
the (for the most part) undisturbed observations. 
The monthly records of a single year at any one of the observatories sufficed to mani- 
fest an order and sequence in the ratios of the aggregate amounts of disturbance in each 
of the twenty-four hours to the mean amount in the twenty-four hours taken as unity, 
which placed beyond a doubt the fact that, casual and irregular as the disturbances 
might appear in respect to the particular times of their occurrence when viewed in single 
days, they were in their mean effects strictly periodical phenomena ; exhibiting, by the 
character of their periodical variations, a dependence on the sun as their primary source. 
To this important fact the disturbances of each of the magnetic elements, the Declina- 
tion, the Inclination, and the Intensity of the magnetic force, bore concurrent testimony, 
although the hours of maximum and minimum of their respective diurnal progressions 
were dissimilar ; confirming in that particular the inference of the existence of distinct 
periodical laws in the disturbances of each of the elements. 
The bearing of this result upon the methods by which magnetical investigations could 
most successfully be prosecuted was important. It had been remarked at a very early 
date, viz. in the 1st volume of the Toronto Observations, published in 1845, p. xv, that 
“ if the disturbances took place without any systematic prevalence at certain hours 
rather than at others, and with no systematic inequality in regard to direction and 
amount, their influence would be limited to a lengthening of the time required for 
obtaining accurate mean values of the solar-diurnal variation ; but that if systematic 
inequalities were found to prevail in those respects, it was obvious that no duration of 
the observations would eliminate their influence; and the diurnal inequality obtained 
from the whole body of the observations, whatever might be the duration it repre- 
sented, must include the effects of two distinct phenomena, viz. of the disturbances, 
and of the diurnal variation properly so called ; these two phenomena having possibly 
distinct causes, or at least distinct laws.” The conclusion could no longer be doubted, 
therefore, that the first step in the systematic treatment of a body of observations, 
whether for the purpose of studying the laws of the disturbances, or for obtaining a 
correct knowledge of the more regular periodical variations, must be to separate the 
observations into two portions, one of which should include the more significant disturb- 
ances, and the other should contain the remainder of the observations, from which the 
disturbances had been for the most part eliminated. Our present concern is with the 
treatment of the disturbed portion only ; the periodical variations of more regular occur- 
rence are discussed elsewhere. 
