28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 68 



It was next necessary to reduce the curves " with " and " without " 

 the water vapor to a common scale. It had been intended to do this 

 by comparing the heights of the branches of region a and making 

 these heights in the ratio which the higher dispersion work already 

 described on this region indicated. This method had to be aban- 

 doned not only because of the difficulty of reducing the successive 

 portions of the curve to the same scale as a, but also because of the 

 great sensitiveness of the branch a to change in the amount of vapor. 

 For the branch a, extending from say about i to 4 /x, includes several 

 important water vapor absorption bands, namely, <f), ip, fl, and X, 

 besides several smaller bands. These bands, because of the impurity 

 of the spectrum, are not indicated in the smooth contour of this 

 maximum. Their existence is felt, however, by its rapidly decreas- 

 ing height with increasing amount of absorbing vapor as may be 

 seen in figure 8. These separate bands may be seen in the result 

 with the 60° prism shown in figure 7. Only the great band between 

 5 and 8 |U, is noticeable in the impure spectrum recorded with the i^'^ 

 prism. 



The portion of the energy curve in which the greatest interest 

 centers is that part wdiere the wave-lengths are greater than 10 fi, 

 or in other words, of greater wave-length than the part treated in 

 the work with the 60° prism. Fortunately this portion was obtained 

 generally without change of slit over its whole range. After cor- 

 recting for " field light," as will be presently described, all curves 

 were put on the same scale as follows : The assumption was made 

 that there is no absorption by water vapor between the wave-lengths 

 9 and 10 fi. 



The evidence for the validity of the assumption of no absorption 

 of energy between 9 and 10 /x by aqueous vapor is cumulative and as 

 follows : The work with the 60° rock-salt prism which formed the 

 first part of this research indicated no absorption there. Rubens and 

 Aschkinass found certainly less than 5 per cent absorption with 

 steam.'' Further support comes from the work of Langley,' who 

 found at 10.7 /x, from a comparison of high and low sun (air- 

 mass =3.76) observations, complete transmission of the solar energy 

 of this wave-length. The decreased transmission which he found for 

 greater air-masses was without doubt due to the smoky atmosphere 

 of Allegheny (Pittsburgh, Pennsylvania, U. S. A.). Langley 's 

 results are confirmed by similar work on the sun done here and pre- 



^ Annalen aer Ph3sik und Cbemie, 60, p. 418, 1897. 

 'Memoirs National Academy of Sciences IV, p. 159, li 



