2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



the rays of the sun as a whole and of individual wave lengths in 

 particular increase in intensity as their path in air diminishes. From 

 this we estimate the total intensity of the solar radiation outside the 

 atmosphere altogether. 



There are certain parts of the spectrum where by reason of power- 

 ful selective absorption of rays by water vapor and other terrestrial 

 atmospheric vapors and gases, sufficiently accurate atmospheric 

 transmission coefficients cannot be determined in this manner. 1 This 

 offers no great difficulty, for, with Langley, we assume that these 

 absorption bands would be absent outside the atmosphere. Hence 

 the intensity of these parts of the spectrum outside the atmosphere 

 can be determined by interpolation from the intensities found on 

 either side of them. 



Whatever the value of the atmospheric extinction of solar rays, 

 all good solar constant work depends on accurate pyrheliometry 

 expressed in standard calories. 



During the investigation we devised two forms of standard pyr- 

 heliometer on quite different principles. These instruments agree 

 with each other to within 0.5 per cent, and they yield values of the 

 solar radiation ranging from 3 to 4 per cent above those found with 

 different copies of the Angstrom pyrheliometer. This latter instru- 

 ment was adopted as the international standard for the measurement 

 of radiation by the meeting of the International Meteorological 

 Committee held at Southport in the year 1903 and by the Inter- 

 national Union for Solar Research at its meeting at Oxford in the 

 year 1905. Mr. A. K. Angstrom has, however, lately pointed out 

 that the Angstrom instrument is subject to slight errors which cause 

 it to read about 2 per cent too low, according to his opinion. If so, 

 this brings the scale of the Angstrom within less than 2 per cent of 

 the scale of the Smithsonian Institution. The latter scale is fortified 

 by the fact that in our several standard pyrheliometers it is possible 

 to introduce and determine test quantities of heat. This has been 

 repeatedly done in each of these instruments, and the test quantities 

 of heat have been recovered to within 0.5 per cent. 



1 Investigations of Fowle showed, however, that transmission coefficients 

 can be obtained even in the great infra-red bands of water vapor, whose 

 employment would practically obliterate the bands outside the atmosphere. 

 Hence we may conclude that if there are diffuse atmospheric bands not easily 

 recognizable, they will be almost exactly allowed for by ordinary transmission 

 coefficients. See Smithsonian Misc. Coll., Vol. 47. 



