232 
MAJOR-GENERAL SABINE ON THE DISTURBANCES OE THE 
element at one station and another element at another station, — all confirmatory of the 
conclusions arrived at by M. Gauss, and of the opinions of those who, antecedently to 
the establishment of the British Colonial Observatories, had anticipated that the distinctive 
characters of the disturbances at individual stations would require to be studied , as the 
first step in a systematic inquiry into their causes , sources , and mutual relations. 
The hourly observations made at the Colonial observatories were received at Wool- 
wich in the form of monthly tables, in which the days of the month were arranged in 
successive horizontal lines, and the hourly observations in twenty-four vertical columns; 
an additional column at the side showed the mean of each day, and an additional line 
at the bottom of the Table the mean of each hour in the month. Even a very super- 
ficial examination of these Tables at any one station sufficed to show that certain hours 
were more affected by disturbance than others. These hours were not the same at 
different stations ; and no distinct relation could be traced at any station between the 
hours of principal disturbance and those of the well-recognized horary fluctuation due 
to the regular solar-diurnal variation. It was obvious therefore that the horizontal line 
at the bottom of each monthly table, which showed the mean values at the several hours 
(or what might be termed the diurnal inequality), did in fact represent two variations, 
viz. 1st, the regular solar-diurnal variation, and 2nd, a diurnal variation due to the dis- 
turbances ; the two having every appearance of proceeding either from distinct causes or 
from distinct actions of the same original cause. The means of separating them per- 
fectly from each other did not readily present themselves, but to do so approximately , 
and with an approximation quite sufficient for many practical purposes, was merely a 
work of labour. The very feature which marked certain of the observations as disturbed, 
viz. the magnitude of their discordance with the other records standing in the same 
column with themselves, or (as more readily seen) the magnitude of their differences from 
the mean value at the same month and hour at the foot of the page, appeared to supply 
a ready means (in the absence of any more exact criterion) of distinguishing the observa- 
tions which were most affected by disturbance. It was soon found that by assuming for 
each element and for each station a certain amount of difference from the monthly mean 
at the same hour as the indication of disturbance , the records in each month might be 
separated into two portions, of which the smaller, containing the disturbed observations, 
might be set apart for an examination of the laws of disturbance; whilst the larger 
portion, from which the disturbances had been thus eliminated, would become more 
available for obtaining a correct knowledge and analysis of the progressive and regular 
variations than when they were mixed up with the casual and transitory affections. 
By maintaining these assumed discriminating or separating values (forming the cri- 
teria whereby each observation was assigned either to the disturbed or to the undisturbed 
category), constant at each station , the laws of disturbance in different months and 
different years, if such laws existed, might be studied with convenience and security ; 
and by so adjusting the values adopted at the different stations as to cause the number 
of the disturbed observations at each station to bear nearly an equal proportion to the 
