ON COMPARISON AND REDUCTION OF MAGNETIC OBSERVATIONS, 299 
Taste III.—WMean Non-cyclic Effect per ‘ Quiet’ Day for each Year. 
| (Horizontal Force) | (Vertical Force) 
6 
Year Declination x 10 % 108 Inclination 
‘ 
1890 —0'36 +23 +15} ~0'10! 
1891 + °29 +23 —12 — 18 
1892 + ‘14 +53 — 20 | — ‘41 
1893 + °26 +40 — 26 | — 35 
1894 + ‘15 +33 +12 — 18 
1895s — 05 +44 —15 | — °33 
Relation of Non-cyclic Effects to Annual Changes. 
§ 3. To see the full significance of these data regard must be had to 
the magnitudes of the annual changes of the several elements. Table IV. 
gives these for the period considered, along with the number of average 
‘quiet ’ days, which, according to Table I., would have sufficed to produce 
changes numerically equal to the annual changes observed. 
Tasie IV. 
\Declinatt Horizontal Vertical F eblivalk 
— } eciination Force ertica orce nelination 
Mean annual change, 1890-95. | —6”8 21x10° | —19x10°% —1'9 
Number of ‘quiet’ days pro- | | 952 6 23 72 
ducing equal change . As = 
The figures relating to the horizontal force and inclination are so 
significant that comment in their case seems unnecessary, 
As regards the declination and vertical force in individual months, 
notably January, the non-cyclic effects have been as large and consistent 
as with the other two elements, but in general this has been far from the 
case. As Table III. shows, in two years out of the six both declination 
and vertical force have exhibited a mean non-cyclic effect opposite in sign 
to that supplied by the six years as a whole. 
In considering such a phenomenon one ought of course to remember 
that it is contrary to probability that any sixty arbitrarily selected days— 
the number on which an annual mean is based—will produce a diurnal 
Variation truly cyclic after allowance is made for the normal annual 
change ; and thus part of the irregularity exhibited by Table III. may 
reasonably be attributed to pure chance. When, however, one looks at the 
uniformity of sign in the non-cyclic effects in the horizontal force and 
inclination exhibited in Table I., and remembers that the monthly means 
in that table are based on only thirty days, one must, I think, conclude 
that the variability of sign in the declination * and vertical force? results, 
at least in Table III., has probably a true physical basis. 
1 In 1890 the means of vertical force and inclination are based on the results of 
only ten months, in one of which (March) an abnormally large positive non-cyclic 
effect was recorded in vertical force. 
* In the case of declination, and it alone, the non-cyclic effect is opposite in sign 
to the secular variation. 
° In 1890 a positive non-cyclic effect appeared in only one month (January); in 
1891 a negative effect in only one month (November). 
* In both 1890 and 1894, however, a slight majority of individual months 
exhibited a xegative non-cyclic effect as usual. 
