NO. 8 WATER-VAPOR TRANSPARENCY FOWLE I5 



by deflections of the galvanometer before and after the insertion of 

 the sector. It is to be noted that the vakies of intensity transmitted 

 by rotating' sectors which have been observed here in several in- 

 stances and by Coblentz are always greater than the theoretical 

 value. 



Further, not even the vertical slits of fixed widths proved satis- 

 factory and widely differing reduction values were obtained in using 

 them. The cause was obscure until Dr. Abbot suggested that the 

 wave-length of the energy might be sufficiently great to cause large 

 spreading" of the beam by diffraction ^ even with the wide slits 

 in use (e. g. 2.36 mm.). The fact that this trouble occurred only 

 when the energy passed through the water-vapor tube and an image 

 of the lamp was formed by the mirror M^ on the slit seems to sup- 

 port the explanation. For in this case the light just filled the 

 diaphragmed collimator of the spectroscope, and diffraction would, 

 as soon as the wave-length was great enough, spread the light 

 beyond both sides of the diaphragm when the narrower slit was up 

 and not when the wider slit was in use. The increasing spread of the 

 light with greater and greater wave-lengths would result in greater 

 and greater reducing ratios for the diaphragm as determined by 

 the deflections for greater and greater wave-lengths. When, how- 

 ever, the energy from the lamp flowed directly through the slit into 

 the spectroscope it so over-filled the collimator that doubtless what 

 was diffracted off from the mirror on one side was diffracted on, 

 so to speak, from the energy already oft' on the other side. Unfortu- 

 nately the observations were reduced, because of the press of other 

 work, over a year after they were made, so that there was no oppor- 

 tunity for full investigation of the supposed diffraction phenomena. 



THE BOLOMETER AND GALVANOMETER 



For the first part of the research extending to lOju. a bolometer 

 strip 0.1 mm. wide by 12 mm. tall was used. Its resistance was 

 about 4 ohms. No plate of any material whatever closed the front 

 of the bolometer case. For the latter part of the work a bolometer 

 strip 0.5 mm. wide by 12 mm. tall was used. The current of 0.04 

 ampere through the bolometer strip was furnished by two storage 

 cells in parallel. This current was increased to 0.08 and finally to 

 0.13 ampere for the 191 6- 17 work. The bolometer strip was placed 

 in an air-tight case, the front being closed with a i cm. thick rock- 



^ Annals of the Astrophysical Observatory of the Smithsonian Institution, 

 I, p. 79, 1900; Lick Observatory Bulletins, 3, 42, 1904. 



