TERRESTEIAL MAGNETISM AT KEW OBSERVATORY. 
95 
§ 17. Table VIII. gives for comparison with Table VI. the cliaracter data for 
individual years from which Table VII. was derived. The entries are the arithmetic 
means of the character figures, heavy type being used when the mean value for the 
year derived from the whole 31 columns is exceeded. 
If entries in Tables VI. and VIII., which exceed the yearly mean, be regarded as 
affected by a positive sign, and those which fall short of the mean as affected by a 
negative sign, the parallelism between the two tables can be roughly gauged by 
comparing the number of agreements and differences in the signs of the corresponding 
entries. There are in all 11x31, or 341 entries in each table. One of these—the 
entry in column n + 6 of Table VI. for year 1895—exactly equals the mean value. 
Omitting this, there are no fewer than 286 agreements in sign as against 54 difterences. 
Many of the differences of sign occur in cases where the departure from the mean 
value is trifling. The natural inference is that disturbance plays a large part in the 
phenomena exhibited even by the H ranges. 
Of the 99 entries in the columns headed n —9 to n—1 only 32 are above the mean 
in Table VI., and only 31 in Table VIII. ; whereas of the 99 entries in the columns 
headed n +1 to n + 9, those above the mean number 59 in Table VI. and 62 in Table VIII. 
§ 18. The difference between the days which follow and which precede the 
representative day n of large sunspot area is brought out, perhaps even more clearly, 
by considering the algebraical excess of the percentage value for day n + s over that for 
day n—s, for the values 1 to 15 of s included in Tables IV., V., and VII. Table IX. 
gives the results thus obtained for the 11 years combined. It was derived from data 
going one decimal place beyond the data given in Tables IV., V., and VII. 
Table IX.—Excess of Percentage Figure for Day n + s over that for Day n—s 
in Tables IV., V., and VII. 
s. 
1. 2. 3. 4. 5. 
6. 7. 8. 9. 10. 
11. 12. 13. 14. 15. 
Spot area . . . 
H ranges . . . 
Character figures . 
+ 2 + 6-l-4 + 3-f-l 
+ 8 +12 +17 +18 +16 
+ 6 +11 +14 +18 +15 
- 1 - 1 -2 -2 -1 
+ 13 + 7+4+1 -1 
+ 15 +12 +4 +1 -3 
0 +2 +3 +4 +4 
-3 -5 -3 -2 -1 
-2 -5 0 +4 +4 
The first line shows how trifling the a-symmetry was in the sunspot areas, considering 
how greatly the area on day n exceeded the mean. 
The differences appearing in the last two lines of Table IX. for values of s exceeding 
9 are presumably dependent in part on sunspot phenomena prior to day n—15. The 
fact that the data from H ranges and character figures accord so closely, not merely 
in sign but in absolute size, must be regarded as largely fortuitous, because the scale 
of the character figures is a wholly arbitrary one. 
§ 19. The application of the method of Table IX. to the data for the group of years 
