NO. 8 WATER-VAPOR TRANSPARENCY FOWLE 5 



aqueous vapor would be at a minimum, whereas the carbonic acid 

 gas would be practically as effective an absorber as in the summer 

 time. Finally, because of both the unexpectedly small absorption 

 of radiation at the longer observed wave-lengths and the too-small 

 quantity of aqueous vapor possible in the laboratory experiments, 

 observations were made in April, 1917, of the atmospheric absorp- 

 tion of the radiation from the sun itself. 



The necessity of the right-working of so many unruly processes, 

 the galvanometer, the bolometer, the source of radiation, a trans- 

 former working far beyond its capacity in furnishing current for the 

 lamp, the need of sunny weather, not too damp for the use of rock 

 salt nor too windy for the galvanometer or bolometer, and without 

 too man}' of the ever-prevalent summer cumuli — all these require- 

 ments tended to make the securing of good observations a tedious 

 ])rocess. Often, too, all the difficulties had to Ije attended to single- 

 handed. 



Before proceeding the writer wishes to express his gratitude to 

 Dr. C. G. Abbot, the Director of this Observatory, who not only 

 suggested the research but at all stages was ever ready to help with 

 suggestions and criticisms in the many perplexing" problems. 



WORK OF OTHER INVESTIGATORS 



Although less than i per cent of the sun's incoming energy lies 

 at a wave-length greater than 2 jx, nevertheless the intensity of the 

 sun's heat is so great that measurable energy would be expected at 

 greater wave-lengths unless lost in passage through the constituents 

 of the earth's atmosphere, or possibly in the gaseous envelopes about 

 the sun itself. Langley,^ in his solar and lunar spectrobolometric 

 researches, found indications of energy between 2.0 and 2.5 jii, 2.8 

 and 4.0 jx, and 4.5 and 5.3 /x. Beyond the maximum at 4.6^1, Langley 

 states, " lies the longest break of all, stretching from 5.0 to y.y /x." 

 BetWeen 8.4 and 9.1 /x and at lO.y fx there is practically complete 

 transmission as indicated by observations to an air mass of 3.76, and 

 the observed decrease in transparency for greater air-masses is 

 probably due to the smoke in the atmosphere. 



^ The Solar and Lunar Spectrum, Alemoirs National Academy of Sciences, 

 IV, p. 159-170, 1888. Note: The longer wave-lengths given by Langley are in 

 error and have been corrected by means of his recorded deviations and a 

 deviation-wave-length curve computed from the dispersion data for rock 

 salt given in Vol. I of the Annals of this Observatory. 



