1 6 INFRA-RED ABSORPTION SPECTRA. 



THE SPECTROMETERS. 



The general arrangement of the reflecting spectrometer is shown in 

 fig. I. The rays enter the instrument through the bilateral slit F , are 

 brought to parallelism by the concave mirror M^, traverse the prism P, 

 and after being collected by the concave mirror M^, are brought to focus 

 on the slit F^. The portion of the spectrum which passes through F^, 

 which is in the vertical focus, falls upon the exposed radiometer vane. 

 The spectrometer stood in an inner basement room with cement floor, 

 and the radiometer deflections were read through a hole in the wall. 

 As a general rule, however, the temperature of the rooms was fairly 

 constant, so that no difficulty was experienced in making the observa- 

 tions with the door open. 



The spectrometer and radiometer were mounted upon a large slab 

 of slate, which lay upon a large table. Although the latter stood upon 

 the cement floor, there was rarely any trouble with earth tremors. 



A 35 cm. focal length mirror spectrometer was used for the explora- 

 tions to 15 /x.. With this instrument the image of the collimator slit was 

 slightly curved. For the region up to 9.5 jx the spectrometer slits were 

 0.4 mm. in width, while beyond this point the collimator slit was gradu- 

 ally widened to i mm. at 15 /x. A very considerable portion of the work 

 was repeated to y jx, using a pair of i-meter focal length, 20 cm. aper- 

 ture mirrors mounted on a large spectrometer. The figuring of these 

 mirrors was excellent and the slit image was perfectly straight. These 

 mirrors gave about twice the dispersion of the smaller spectrometer, and, 

 although the concentration of the energy in the spectrum was greatly 

 reduced, which resulted in small radiometer deflections, the measure- 

 ments obtained with them were so uniform in their agreement that for 

 the region from 3.43 to 6.86 /* the results are more trustworthy than 

 those obtained with the smaller apparatus. 



With this larger apparatus the spectrometer slits were 2' of arc, 

 while with the smaller they were 4' of arc on the spectrometer circle, 

 so that for the large spectrometer the dispersion was comparable to that 

 of fluorite. With it numerous bands were resolved from 6 to 7 /a, but 

 only occasionally were small bands found in the transparent region 

 from 4 to 5 /t, to be mentioned later, while the 3 to 3.5 /x region was 

 sometimes found complex. 



THE PRISMS USED. 



For the region from 0.8 to 2.5 /la a 60 mm. quartz prism, having a 

 refracting angle of 60° i' 32", and the small spectrometer were used. 

 For the region beyond 2.5 jn to 15 ^u, a 70 mm. rock-salt prism was 



