NO. 8 WATER-VAPOR TRANSPARENCY — FOWLE 39 



him near wave-lengths 8.8 [x and 10.7 /x. The apparent decrease in 

 transmission for greater air-masses he attributed to the smoky atmos- 

 phere prevalent at Pittsburgh. The great increase in sensitiveness 

 and accuracy of the present day spectro-bolometric apparatus made 

 it seem worth while to repeat these observations. 



The scattered light, which in the spectrum of the Nernst lamp was 

 troublesome, becomes absolutely prohibitive here in a directly formed 

 1 5 "-prismatic rock-salt solar spectrum. At 10 /x with the large galva- 

 nometer deflection of 20 cm., practically all this deflection remained, 

 and was therefore field light, upon the insertion of the quartz plate. 

 In the case of the Nernst-lamp spectrum at this wave-length only 

 16 per cent of the deflection was found to be false. Langley elimi- 

 nated the stray light by using a sifting train, but this was not feasible 

 in the present case. Dr. Abbot suggested the use of a screen of solid 

 iodine, which Coblentz' work ^ showed to be opaque to the visible 

 radiation and increasingly transparent for the longer wave radia- 

 tion. Accordingly Dr. Abbot with Mr. Aldrich and the writer pre- 

 pared two plane parallel rock-salt plates 0.65 and 0.34 cm. thick. 

 These were uniformly heated, flakes of iodine placed on one, the 

 other plate quickly placed on top and the two plates squeezed to- 

 gether. After several attempts a thin, fairly uniform film of iodine 

 was thus obtained which micrometer measures showed to be between 

 0.005 ^^^ 0.007 cm. thick. 



TRANSPARENCY OF IODINE 



Although the direct measures on the solar spectrum were made 

 first, it is perhaps proper to discuss here the later measures on the 

 transmission of the iodine film. The measurement of this is not so 

 simple as might at first be thought, again because of the scattered 

 light. No separate allowance will be made here for the light reflected 

 from the rock-salt plates used to protect the iodine film. The efifect 

 of the insertion of the screen as a whole is wished. 



The observations consisted in making an energy curve of the 15°- 

 prismatic rock-salt spectrum of the Nernst lamps as usual, except 

 that every few seconds the iodine screen was inserted, then the iodine 

 screen with the quartz plate, then the quartz plate alone, then all 

 screens removed. The significance of the respective deflections 

 measured were as follows when the plate was inserted at wave-length 

 A greater than 4 fx : A,r is used to signify the wave-length. 



(a) Lamp alone: Energy of Aa; -|- scattered energy of A<4.iU,,-f 

 scattered energy of A>4/t. 



Physical Review 16, p. 72, 1903. 



