Variation of the Solar Radiation. ; 85. 
As for the causes of the increased opacity of the atmo- 
sphere, nothing positive is certainly known, though some 
have surmised that the many and long-continued volcanic 
eruptions of 1902 may have distributed more fine dust than 
usual in the air. It can, on the other hand, be more de- 
finitely stated that the defect of transparency was not caused 
by increased water-vapour; for the inferior transparency was 
observed on several of the driest days on record, as shown 
by the form of the great water-vapour absorption-bands in 
the intra-red spectrum. 
(b) Solar Radiation outside the Atmosphere.—The following 
Table (p. 86) includes the best determinations of the solar 
radiation outside the atmosphere, made at the Smithsonian 
Observatory from October 1902 to March1904. A part of these 
were included in Mr. Abbot’s paper above referred to, but as 
additional computations and data are given for several of the 
days, they are here repeated*. As was stated at an earlier 
page, the values are not regarded as absolute measures of 
the solar constant, but only as forming a series comparable 
among themselves, from which all known terrestrial variations 
are excluded, and which may furnish evidence of a suspected 
variability of the solar radiation. 
Column 9, giving the radiation outside the atmosphere, is 
derived by multiplying the radiation at the earth’s surface, 
- given in Column 7, by the ratio of the area of the eneryy- 
curve outside the atmosphere to that at the earth’s surface, 
as given in Columns 6 and 5, and correcting to mean solar 
distance by Column 8. As already said, the agreement, both 
in Column 6 and Column 9, of the two computations of the 
same day, furnishes two checks on the goodness of the day’s 
work. Different instrumental conditions may, however, alter 
the values in Column 6 as between different days. 
Tt will be seen that the days of observation during 1903 
fall in two groups, yielding high values and low values of 
solar radiation respectively, the former including all days 
prior to March 26, and the latter all days subsequent to that 
date. A possible return to the higher values is indicated by 
the observations of February 11, 1904, but we must await 
subsequent measures to confirm this. : 
* The values of March 3, April 17, and April 28, included in Mr, 
Abbot’s paper, are here omitted because the observations of those days 
are of little weight. 
