﻿FOWLE] SOLAR RADIATION MEASURES 4O7 



air-mass the smaller the correction. However the greater the air- 

 mass, the straighter becomes the actinometer curve and therefore 

 the less the variation in the correction as a function of the air-mass. 

 On the other hand it is to be noted that the correction is not much 

 affected by quite a considerable change in the transmissibility of our 

 atmosphere. It might seem, a priori, that the amount of water- 

 vapor present in the air would be an important factor ; a small 

 amount of aqueous vapor corresponding with a smaller correction ; 

 nevertheless the value of the correction derived from February 19, 

 1903, a day of phenomenally small water-vapor absorption, does not 

 bear this out. Further study seems, however, necessary on this 

 point. 



It would be depending too far on the accuracy of the curves shown 

 to use them as a basis for the determination of secondary corrections 

 like those for changes in water-vapor absorption or general trans- 

 mission coefficients. It seems apparent from the earlier table of 

 comparisons, however, that in extrapolating from such air-masses 

 (1.5-2.5) as are generally involved, a correction of 14 percent should 

 be added to the resulting actinometer value. This corrected value 

 then would probably be of nearly the same order of accuracy as if 

 bolometrically determined. 



For other places of the same altitude as Washington, and where 

 the air-masses used would be nearly the same, it seems very probable 

 that the same correction might apply, provided the transmissibility 

 of the air was within the limits examined here. 



For places of greater altitude the correction, although in all prob- 

 ability much smaller, might possibly be more variable. 



In order to extend the scope of actinometric observations the acti- 

 nometer in use here may be compared with several of the Angstrom 

 type in use at the Weather Bureau, and from this the comparison 

 may be extended to other stations in this country and elsewhere 

 supplied with Angstrom pyrheliometers. Thus the value of the 

 radiation in units comparable with those used here, may probably be 

 determined for other and perhaps earlier days than those here in- 

 cluded. As the correction may be assumed constant for all stations 

 in or near Washington, any actinometer observations made in this 

 locality may serve to determine the correction at other places where 

 simultaneous observations have been made. 



It should be borne in mind that the determinations here are not 

 regarded as absolute but are only relatively comparable among them- 

 selves. The great variation in actinometer readings at different 

 places and by different instruments is probably often due to the 



