242 
DR. C. CHREE: DISCUSSION OF KEW MAGNETIC DATA 
G.M.T. was employed, as already stated, in the diurnal inequalities, and thus the 
values deduced in the first instance for the Fourier coefficients referred to G.M.T. 
Local mean time is, however, more generally instructive, and accordingly the necessary 
transference to local time has been made in Tables XIX. and XX., and all the 
subsequent tables relating to Fourier coefficients. The transference does not affect 
the amplitudes. The corrections it requires to the phase angles are 
+ 19' in op, +38' in a 2 , +57' in a 3 and +1° 16' in a 4 , 
Kew local time being the equivalent of 19' after Greenwich. 
The relation with sunspots will be discussed more fully later, but the significance of 
the tables will be better understood if regard is paid to the following facts. 1890 was 
a year of sunspot minimum with a frequency of 7'1. The frequency rose rapidly to a 
maximum of 84'9 in 1893, and fell more slowly to 9'5 in 1900. In 1898 there was a 
slight arrest in the fall, the mean frequency being 267 as compared with 26'2 in 
1897. The year 1893, though the year of sunspot maximum, was on the whole a 
quiet year free from large disturbances, much quieter than 1892, 1894 or 1895. 
In Table XIX. c 1} c 2 , c 3 and c 4 all rise to a maximum in 1893, and the general 
parallelism with the sunspot variation is readily seen, especially in c 4 and c 2 . Table 
XX. shows a similar state of matters, except that the value of c 1 in 1893 fell short of 
that in several adjacent years. 
The variation in the phase angle is by no means so decisive. In Table XIX. the 
largest values of op and a 2 are associated with the years of sunspot minimum, but the 
smallest values occur in 1897, an intermediate year as regards sunspots. Accident 
seems to play a sensible part in the relation between phase angles in successive years, 
especially in the case of a 3 and a 4 . The same remark applies in even greater measure 
to Table XX. 
In fact the influence of sunspot frequency on the phase is so small that for its study 
it is desirable to combine the years in groups representative respectively of large and 
small sunspot frequency, in hopes of eliminating accidental features. 
§ 20. Tables XXI. and XXII. show the variation in amplitude and phase angle 
throughout the year in the Fourier waves representing the diurnal inequality in H, and 
Tables XXIII. and XXIV. do the same for V. Results are given for the whole 11 years, 
also for 1890, 1899 and 1900 representing few sunspots, and 1892 to 1895 representing 
many sunspots. Arithmetic means from the 12 months are given in Tables XXI. and 
XXIII. When there is considerable variation of phase throughout the year, contribu¬ 
tions from different months to the seasonal and yearly diurnal inequalities to some 
extent neutralise one another. Thus frequently a better idea of the average activity 
of the forces to which any particular Fourier wave is due is derivable from the 
arithmetic mean of the c’s than from the corresponding c in the mean diurnal 
inequality for the year. 
