90 A Possible Variation of the Solar Radiation. 
by a permanent diminution of solar radiation of 10 per cent. 
It is the absorption of heat by the oceans which slows the 
terrestrial change (if any) produced by change in the solar 
constant. Since the indirect effects of the absorption of 
solar radiation are feltin the oceans to depths of 2000 metres, 
it may be shown that a change of 2° over the entire surface 
of the earth would be delayed over a year. 
Summary and Conclusion. 
A series of determinations of the solar radiation outside the 
atmosphere (the Solar Constant) * extending from October 
1902 to March 1904 has been made at the Smithsonian: 
Astrophysical Observatory under the writer’s direction. 
Care has been exercised to determine all known sources of 
error which could seriously affect the values relatively to 
each other, and principally the varying absorption of the 
earth’s atmosphere. Though uncertainty must ever remain 
as to the absorption of this atmosphere, different kinds of 
evidence agree in supporting the accuracy of the estimates 
made of it, and of the conclusions deduced from them. 
The effects due to this absorption having been allowed for, 
the inference from these observations appears to be that the 
solar radiation itself fell off by about 10 per cent., beginning 
at the close of March 1903. I do not assert this without 
qualification, but if such a change in solar radiation did 
actually occur, a decrease of temperature on the earth, which 
might be indefinitely Jess than 775 C., ought to have 
followed it. 
On comparing the observed temperatures of 89 stations, 
distributed over the North Temperate Zone, with the mean 
temperatures of the same stations for many previous years, 
itis found that an average decrease of temperature of over 
2° centigrade actually did follow the possible fall of the 
solar radiation, while the temperature continued low during 
the remainder of the year. Stations remote from the re- 
tarding influence of the oceans show a much greater variation 
than that of the general mean. 
While it is difficult to conceive what influence, not solar, 
could have produced this rapid and simultaneous reduction 
of temperatures over the whole North Temperate Zone, and 
continued operative for so long a period, the evidence of 
solar variation cannot be said to be conclusive. Nevertheless, 
such a conclusion seems not an unreasonable inference from 
the data now at hand, and a continuation of these bolographic 
* T wish to renew my caution that the absolute values of the solar 
constant thus given are more likely to be found in defect than in excess. 
