ON COMPARING AND REDUCING MAGNETIC OBSERVATIONS. 79 
the deviations from the normal daily path show.! In how far we are to 
conceive of the two last variations as disturbances must, in my opinion, 
be decided by experience. 
In any case we require, for the fixing and estimation of these last varia- 
tions, a distinct starting point, which the normal daily path may present. 
It is, therefore, specially important here to establish this normal dail 
path of the magnetic elements. Nowregarding the method which Sabine 
has devised for this, and also used so much, there is, in the first place, 
displayed what Lefroy, Weyprecht, and others have made so prominent, 
the arbitrary nature of the limits which are assumed for the expulsion of 
the so called disturbed data. Among the different proposals which have 
been made for a rational fixing to these limits the most worthy of notice 
is that of Buys Ballot, in which these limits are to be set where the devia- 
tions begin to show another period. Van der Stok has distinctly modified 
the Sabine method for the discovery of the normal daily path. His 
altogether very complicated method suffers, in my opinion, from the same 
wide evils as the Sabine method, viz., that it proceeds from the daily 
paths, derived, like them, from the sum of all observations without dis- 
tinction, i.e. including disturbances. Now it is evidently, as Weyprecht 
has already shown, impossible out of the so procured data to get rid alto- 
gether of the influence of the disturbances on the normal daily path, if 
these are not quite irregularly distributed over the day, but are all subject 
to a certain daily period. Lefroy again, in his working out of observations 
at Fort Simpson and Lake Athabasca, has not employed all the data for 
the deriving of the first hour’s means, but only the days and hours which, 
according to him, were not to be regarded as disturbed, i.e. where the 
amplitude of the movements does not go beyond a certain limit. The 
fact that the exclusion of these movements is not settled through the 
criterion of Buys Ballot on the one hand, as well as the consideration, on 
the other, that days with not less amplitude of movement may also be 
disturbed, because the disturbed periods might unite with the normal 
periods, so as to weaken themselves through interference (which, as we 
shall see, is partly the case), prevents the method from being satisfactory. 
In the Programme and in the Sittings of the fourth International Polar 
Conference in Vienna (April 1884) I have given out and developed a 
new method for the derivation of the normal daily path of the magnetic 
elements (see ‘Communications of the International Polar Commission,’ 
No. 94, p. 199; No. 97, pp. 208, 211; No. 98, pp. 254, 255, 257, 258), 
which is supported by the observation that in the magnetograph traces, 
even at the epoch of maximum disturbances, in every month are to be 
found a number of days in which a quite regular, and also as regards 
these days concerned a recurring periodical path is distinctly recognised. 
_ I regard these days as days with undisturbed daily paths, and the 
ourly means of all these days as representative of the normal daily path 
of the elements concerned in the month in question, according to its 
relative as well as to its absolute size. The selection of these normal days 
may from the first likewise seer very arbitrary ; in practice, however, 
this is not the case, as hardly a doubt can arise as to which days are to 
be taken, and besides the result will not be very distinctly different 
whether one chooses one or two days more or less, if from the first one 
? For the sake of simplicity I have spoken here only of the daily periods ; clearly 
for the remaining periods also, which show the variations of the earth’s magnetism, 
suitable distinctions can be made. 
