148 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1914. 
in the infra-red as great depressions of the curve. Solar absorption 
lines are shown in the visible spectrum as smaller depressions of the 
curves. The three curves were taken at different hours of the morn- 
ing, when the path of the solar rays in air was five, four, and three 
times, respectively, that which would occur if the sun were vertically 
overhead. The reader will notice that the curves are respectively 
higher and higher, especially in the violet end of the spectrum, owing 
to the decrease of the length of path of the sun rays in the air. At 
several points a change of scale of the curves is shown. ‘This is due 
to the introduction in the beam of rotating sectors of different angular 
apertures in order to keep the record always within the limits of the 
registering photographic plate. If these changes of scale had not 
been made, the curve would run up in the edge of the red to the 
height of several feet. 
By suitable computation, by the aid of the exponential formula 
developed by Bouguer about the year 1760, it is possible to compute 
for each of the parts of this spectrum energy curve the intensity 
which would be found if one were outside the air altogether. In 
this way one might construct a curve similar to the three shown in the 
Ulustration, which would represent the intensity of radiation beyond 
the limits of the atmosphere. In such curves as these the area 
included between the curve and the axis of zero radiation is propor- 
tional to the intensity of the whole solar beam, including all wave 
lengths. This, of course, is also measurable by the pyrheliometer. 
Accordingly we multiply the reading of the pyrheliometer by the 
ratio between the area of the curve outside the atmosphere and that 
which is found at the observing station, and thereby we obtain the 
solar constant of radiation. In this process, however, we follow 
Langley’s assumption that there will be no absorption by water vapor 
or oxygen in the sun itself, and therefore draw a smooth line in our 
extra-atmospheric energy curve, where great atmospheric bands occur. 
About 700 determinations of the solar constant of radiation have 
been made by the Astrophysical Observatory of the Smithsonian 
Institution, some at Washington, at sea level; others at Mount Wilson, 
at an elevation of about 1 mile above sea level; others at Mount 
Whitney, at an elevation of nearly 3 miles; and others at Bassour, 
Algeria, at an elevation of three-quarters of a mile. No differences 
beyond the reasonable errors of measurement are found between 
observations made at two stations on the same day, whether made at 
sea level or at any of these stations, up to the elevation of Mount 
Whitney, 14,500 feet above sea level. Hence it appears that the 
method of estimating atmospheric transmission is probably sound. 
The mean value of the solar constant of radiation, as thus found from 
700 determinations, is 1.933 calories per square centimeter per 
minute. By this is meant that if the sun’s rays outside the atmos- 
