126 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1944 
The weighted divergence between stations being 0.0076 calorie, 
the weighted average departure of one station from the mean solar 
constant derived from two stations is 0.0038 calorie. 
Although it is not fair to Montezuma to suppose that the stations 
are of equal merit, yet if we make that assumption, and proceed as 
usual, we find the weighted mean percentage probable accidental error 
of a single day of observation of the solar constant at one station 
to be: 
100 X 0.0088 X 0.84 1.94=0.164, or 1 of 1 percent. 
In volume 6 of the Annals of the Astrophysical Observatory of the 
Smithsonian Institution are contained in table 24 nearly 19,000 meas- 
urements of the solar constant observed through the years 1924 to 1939. 
Several thousand earlier observations of the years 1920 to 1928 are con- 
tained in other publications. Figure 8 is a facsimile of a part of page 
133 of the Annals, which includes the work of September 1934. The 
several observing stations are distinguished by letters M, K, T, meaning 
Montezuma, St. Katherine, and Table Mountain. The solar-constant 
values in columns “S. C.” and “Pfd. 8. C.” are to be understood as pre- 
fixed with 1.9. Thus for “50” read “1.950.” Using the result of Monte- 
zuma and St. Katherine only, which are more accurate than those of 
Table Mountain, there was apparently an increase in the column “Pfd. 
S. C.” from the 1st to the 5th and from the 10th to the 14th of Sep- 
tember, and a decrease from the 14th to the 19th. These changes had 
an amplitude of the order of 0.5 to 0.9 percent, that is about 0.010 to 
0.018 calorie in the solar constant of radiation. 
SEQUENCES OF RISING AND OF FALLING SOLAR ACTIVITY 
I give in table 2 a summary of nearly 500 of the best supported in- 
stances of rise and of fall in the solar constant of radiation selected 
from table 24 of volume 6 of the Annals. The table is arranged by 
months and will readily be understood by an example. Thus, “Janu- 
ary, Rising, 24, 12” means that a case of the solar constant rising for 
a few days appeared to occur beginning January 12, 1924. 
It is of interest and importance to note that the solar variation 
increases in percentage toward shorter wave lengths. It is six times 
as great at 3500 A. in the ultraviolet as in the total solar constant. 
EFFECT OF SEQUENCES OF SOLAR CHANGE ON TERRESTRIAL 
TEMPERATURES 
Using this tabulation of the dates whereon sequences of rise and of 
fall of the solar constant apparently began, I have sought to determine 
whether such phenomena were associated with special behavior of the 
