82 REPOKT OF THE SECRETARY. 



vening absorption wouM outweigh any number requiring correction for atmospheric 

 absorption, and the vahies observed tlirough smallest air masses are the most trust- 

 worthy. It follows that values obtained at a low altitude like that of Washington are 

 necessarily too small, owing to the difficulty of allowing with accuracy for the absorp- 

 tion of the great thickness of air above the observer. 



There has been incorporated in the act appropriating for the support of the Astro- 

 physical Observatory during the fiscal year ending June 80, 1904, a provision for 

 high-altitude observations. In anticipation it may be said that apparatus for that 

 purpose has been ordered. 



Although, therefore, uncertainty attaches to the actual values determined here 

 and to be given below, still it is probable that they are relatively comparable among 

 themselves when we consider the apparent accuracy both of the observations and 

 the exponential formula used in reducing them, as illustrated in Plate VI, and the 

 fact that the apj^lication of coefficients of atuiospheric tiansmission determined simi- 

 larly would practically obliterate the great selective absorption bands in comj)uting 

 the form of the solar energy curve outside the atmosphere. 



REDrCEI) OBSERV.ATION'S OF TOE SOLAR R.VDIATIOX CONSTANT. 



In the following table will l)e found such computations of the solar constant as 

 were made up to July 1, 1!K):>. The holographs on which they depend extend for 

 the most part from 0.375 /< to 2.5 /<, and thus include practically all the solar radia- 

 tion whi(;h reaches the earth's surface." 



A correction of about 1 per cent has been added, representing the best estimate 

 which could be made of the- excess of energy beyond these limits outside the earth's 

 atmosphere. All the results depend on the constant of the pyrheliometer, and may 

 therefore be subject to multiplication by a constant factor to be subsequently deter- 

 mined. Comparing the values obtained with those which you have given in the 

 Mount Whitney report of 3 calories, it will be seen that they are about 25 per cent 

 smaller, and that the difference does not appear to depend on the transmission 

 coefficients, but rather seems chiefly due to a difference in actinometry. 



Thus you have stated the usual actinometer reading at Allegheny, Pa., for clear 

 blue sky at 1.7 calories,'^ while the very highest value obtained here is 1.44 calories. 

 Much lower values are reported from recent observations of Mr. Kiml)all, of the 

 United States Weather Bureau, at Asheviile, N. C, and at Washington, so that on 

 the whole the question of absolute actinometry seems a very open one. 



Turning next to the relative values of the solar constant, it is seen that there is 

 generally good agreement of the results prior to March 26, 1903, and that since that 

 date there has been a decrease of about 10 per cent in the computed constant. No 

 reason for this is known, as some of the best observations were before and others 

 since March 25, those of Pebruary 19, March 25, March 26, and April 29 being con- 

 sidered to have most weight. February 25 was a most extraordinary day as regards 

 aKsence of water vapor absorption. Never since holographs have l)een taken here 

 have the great infra-red water- vapor bands <p W £1 been observed so feeble as on 

 February 19. 



«The observations of October, 1902, which reached only to 0.48«, have been cor- 

 rected l)y means of later work and are therefore of less weight. 

 ^Report of the Mount Whitney Expedition, p. 32. 



