320 ANNUAL KEPORT SMITHSONIAN INSTITUTION, 1910. 



the thickness of the medium traversed. This formula is variously 

 given, but may be used in this form: E=E A?n, where E and E are 

 the intensities of the outgoing and entering beams, A a constant 

 expressing the fraction transmitted through unit thickness, and m 

 the thickness traversed. In the case of the atmosphere it is natural 

 to take unit thickness as that of the layer between the observer and 

 the zenith, and m as equal to the secant of the zenith distance of the 

 celestial object. This latter assumption is not strictly true, because 

 the air layer is not a plane parallel sheet, but spherical in curvature, 

 and secondly because the beam is curved by atmospheric refraction. 

 However, as the air layer of sensible density is thin compared with 

 the length of the earth's radius, and as the refraction is negligible 

 except near the horizon, the approximation is very close for zenith 

 distances less than 75°, for which m=4. Knowing m, and measuring 

 E by the pyrheliometer, two observations at different zenith dis- 

 tances fix the values of E and A. Pouillet, proceeding in some such 

 manner, made numerous determinations of these quantities, and con- 

 cluded that the value of E at mean solar distance is about 1.76 

 calories per square centimeter per minute. This, then, is Pouillet's 

 value of the " solar constant of radiation." 



For the next 40 years this result was generally adopted, although 

 the experiments of Forbes, Violle, and Crova and the theoretical 

 work of Kadau indicated that it was too low. Langley, about 1880, 

 stated Radau's argument in a highly convincing form. Briefly stated, 

 since the transmission of the atmosphere differs, depending on 

 whether we consider blue or red light, and especially on whether we 

 treat of rays which suffer only the general scattering of the molecules 

 and dust particles of the air, or take those which are selectively ab- 

 sorbed by water vapor and oxygen, and which are almost completely 

 extinguished high above the earth's surface — on account of this 

 inequality of atmospheric extinction Pouillet's method inevitably 

 yields too low results. 



Langley, by the aid of his then newly invented bolometer, meas- 

 ured at Alleghemr, and in 1881 at Lone Pine and Mount Whitney, 

 the transmission of the spectral rays separate^, conrputed how the 

 energy of the sun is distributed in its spectrum outside the atmos- 

 phere, and fixed a new value of the solar constant which has been 

 generally accepted almost until the present time. 



The method of Langley, which is that now in use, is complex, but 

 necessarily so. Imagine that you have a very intense solar spectrum 

 before you, and that it is still early morning, with the sun perhaps 

 an hour and a half high. If you had a thin, delicate, blackened 

 thermometer you could carry it along in the spectrum from the ex- 

 treme ultra-violet to far beyond the red end of the visible spectrum, 

 and detect varying degrees of temperature rise proportional to the 



