OO EE 
ON COMPARISON AND REDUCTION OF MAGNETIC OBSERVATIONS, 237 
days. It would, however, be impossible to separate the two sets of cyclic 
terms by any mathematical device, without an addition of physical facts 
or a supply of theories in their place. One way of obtaining additional 
facts would be to compare for a series of years the constant coefficients in 
the harmonic analysis of the diurnal inequalities from ‘quiet’ and un- 
restricted days. The accidental features introduced by the arbitrary 
nature of the choice of ‘quiet’ days might, however, prove troublesome. 
The term in the non-cyclic effect treated as a linear function of the 
time may in its turn be composed of a series of terms, some possibly fluc- 
tuating regularly with the season of the year, others possibly of very long 
period ; its magnitude, at least in individual months, may depend in large 
measure on the accidental preference of one set of ‘ quiet’ days to another. 
Associated Phenomena. 
§ 11. It was pointed out last year (/.c., p. 213) that the elimination of 
the non-cyclic effect through a correction consisting of a linear function 
of the time was determined solely by considerations of convenience and 
mathematical simplicity. It was carefully explained (J.c., §$ 5, 6) that 
General Sabine and Dr. Lloyd had observed phenomena in magnetic 
storms so exactly the converse of those presented by the non-cyclic effect 
on ‘quiet’ days as to suggest that the two classes of phenomena were inter- 
dependent ; and the conclusion was drawn that if this interdependence 
were true the non-cyclic effect might be expected in reality to progress 
irregularly throughout the twenty-four hours. 
These conclusions may now, perhaps, be regarded as more than sur- 
mises. In the ‘ Met. Zeitschrift’ for September 1895 Dr. van Bemmelen 
has described phenomena he terms Wachstérwng, which appear to be of the 
same general character as, if not identical with, what has been termed 
here the non-cyclic effect. 
As the title he selected implies, Dr. van Bemmelen associates the 
phenomena very intimately with ‘magnetic storms. His investigations 
have included data from a variety of stations ; and whilst his theoretical 
conclusions may, perhaps, undergo modification in the future, his work 
certainly indicates that an increase of knowledge as to this outstanding 
phenomenon on ‘ quiet’ days is likely to be of service in the general theory 
of terrestrial magnetism. 
§ 12. In the meantime it might be safest not to assume that the non- 
cyclic element is an effect, and a preceding magnetic storm a cause. The 
fact that the horizontal force, for instance, tends to rise abnormally fast 
during a ‘quiet’ day may, of course, merely represent a recovery from an 
abnormal loss occasioned by a magnetic storm ; but it is at least con- 
ceivable that the abnormal fall during a magnetic storm may be partly a 
consequence of abnormal increase preceding it, or the two phenomena 
may be effects of a common cause. 
Tf ‘quiet’ days, with no appreciable disturbance, were the rule, one 
might possibly determine with ease the relationships of any given ‘quiet’ 
day to a preceding or succeeding disturbed day ; but appreciable move- 
ments will usually be found both before and after a ‘quiet’ day at no great 
interval of time. If the causes operating in large and small disturbances 
are the same, then it is not improbable & priori that a small disturbance 
within a day or two of a ‘quiet’ day may have more to do with it than a 
large disturbance a week before or after. It should also be remembered 
