78 
DR. C. CHREE: SOME PHENOMENA OF SUNSPOTS AND OF 
three groups. The distrihution expressed in percentages of the total number of days, 
whether 660 or 209, was as follows 
In group of days of- 
Largest 
Intermediate 
Smallest 
spot area. 
spot area. 
spot area. 
“Quiet’’(lays. 
31-7 
31-5 
33-8 
Highly disturbed days. . . . 
35-6 
34-0 
30-4 
Thus the group of days of largest spot area contained 2'1 per cent, fewer “quiet” 
days, and 5'2 per cent, more disturbed days than the group of days of smallest spot 
area. 
§ 5. The fact that this second investigation gives at least a suggestion of the 
relationship sought for, while the first investigation gave a purely negative result, 
was at first very puzzling ; but it began to dawn on me that the difference might 
have something to do with the fact that days of large spot area, and days of small 
spot area both tend to congregate in groups, and not to be isolated. Influenced 
partly by this idea, and partly by the desirability of testing a theory of Arrhenius, 
who suggested that the magnetic disturbance effects visible on the earth, are due to 
the discharge from the sun of electrified particles likely to take some 48 hours to 
travel to the earth, I next investigated (in the same paper) w^hether the association 
was not between magnetic phenomena on the earth and phenomena existing on the 
sun, 1, 2, 3, or 4 days previously. In one investigation the 10 days of each month 
were taken for which the absolute D range {i.e., the excess of the daily maximum over 
the daily minimum declination) was greatest. Calling any one of these selected days 
n, the Greenwich projected sunspot area was put down in separate columns for the 
day n., and three previous daysn—1, n—2, and n—3. This being done for each of 
the 10 days of largest D range in the month, we have 10 spot areas in each of the 
four columns headed w, 1, —2, and n —3. Summing up and taking means for 
each column, we get results representative of the spot area on a representative day, 
characterised by a large D range, and on each of the three previous days. 
If Arrhenius’s theory were true, then we should expect the mean spot area for day 
n — 2 to be decidedly in excess of the means from the other three columns, and also in 
excess of the mean derived from all days of the month. 
As a matter of fact, when all the months of the 11 years were treated in this 
way, and the results combined, the mean spot areas from the four columns were all in 
excess of the mean from all days, and the excess was largest for column n — o. 
As complementary to this investigation, the 10 days of least D range in each month 
ol the 11 years were got out, and the corresponding sunspot areas were put down for 
