188 
DR. C. CHKEE: DISCUSSION OF KEW MAGNETIC DATA 
confined to the year, winter (October to March) and summer (April to September). 
Vertical force (V) and inclination (I) suffer more than D and H from artificial electric 
currents, and no inequalities have been published for them since 1901. 
It was decided to begin by analysing the measurements that had already been 
made of the “ quiet ” day curves for an eleven-year period, 1890 to 1900. The 
necessary labour was almost entirely arithmetical, and it was completed with the aid 
of the Observatory staff without extraneous financial assistance. The results were 
embodied in a paper # published in 1903. 
It had gradually been recognised that diurnal variations derived from quiet days 
are not identical with those derived from all days, or from all days but those of large 
disturbance. It became increasingly obvious that the Kew records would not be 
fully utilised until the study was extended to other than quiet days. The 
magnitude of the task was not at first fully realised, and the original programme 
seems to have embraced all the accumulated data, limiting the enquiry however in 
the first instance to the declination. At all events the list of Government Grants for 
1903 to 1904 includes one of £82 10s. “to work up declination (magnetic) results 
obtained at Kew from 1857 to 1900. ' The work was practically confined to the 
11 years 1890 to 1900, and the grant was exhausted before it was completed. 
The curves were divided into “ordinary ” and “disturbed. ’ An ordinary day was 
one in which the general trend of the diurnal variation was clearly recognisable, so 
that when the trace was oscillatory it could be fairly replaced by a freehand pencil 
curve of moderate curvature. 
It had been the practice, when sensible oscillations occurred on a selected quiet 
day, to smooth the curve, replacing it by a pencil trace, so the procedure adopted 
with the ordinary day curves was no innovation. Smoothing was, however, done much 
more extensively than had been the case with quiet curves, and in some instances it 
called for considerable exercise of judgment. To secure uniformity, it was always 
done by myself. Disturbed days were those in which there was so much irregularity 
that smoothing appeared too arbitrary a process. To a certain extent, no doubt, the 
allotment of a day to the disturbed list depended on the judge’s condition, both 
physical and mental, at the moment. The attributes of a disturbed day were 
practically those of “ character ” 2 days under the international scheme 0 (quiet), 
1 (moderately disturbed), and 2 (highly disturbed); and it is only necessary to consult 
the returns from similarly situated stations to recognise the importance of the 
personal element in the selection. The choice, in the present case, it should be 
remembered, was based entirely on the D curves. The total number of days assigned 
to the disturbed list in the 11 years was 209, or an average of 19 a year. The 
number varied, however, from 6 in 1890 to 39 in 1896. The results from the ordinary 
day D curves, excluding a few that were imperfect, were discussed in a paper! 
* ‘ Phil. Trans.,’ A, vol. 202, p. 335. 
f ‘ Phil. Trans.,’ A, vol. 208, p. 205. 
