OBJECT OF THE INVESTIGATION. 3OI 



it, was about 2 cm. in length. In this manner no radiation from the 

 electrodes could enter the spectrometer slit. An image of the arc 

 could be seen in the prism, and any slight trace of the radiation from 

 the electrodes, when visible in the prism, was sufficient to cause large 

 deflections, or even to throw the deflection entirely off the scale. The 

 spectrometer arm was rotated, and the arc was adjusted before the slit 

 by hand. The image in the rock-salt prism showed whether the adjust- 

 ment was correct in the horizontal direction and whether any part of 

 the electrodes was visible. By raising or lowering the shield or arc 

 only the radiation from the vapors could enter the spectrometer slit. 

 By placing a mirror above the scale an image of the latter, in the 

 radiometer mirror, was reflected back into the viewing telescope, which 

 was situated near the spectrometer. In this manner the observer could 

 adjust the arc, raise the shutter, and make a reading before the arc had 

 changed in intensity, which was very annoying when using metallic 

 electrodes. With the salts of the metals in the carbon arc there were 

 no serious fluctuations in intensity after starting the arc. 



OBJECT OF THE INVESTIGATION. 



Angstrom (loc. cit.) and others have found the absorption band of 

 CO2 at 4.28 fx. and CO at 4.59 fx ; Paschen (loc. cit.) found that the COj 

 emission band shifts toward the CO band with rise in temperature, 

 being at 4.4 fx. in the Bunsen flame. The emission spectra work, par- 

 ticularly that of gases in a vacuum-tube, was undertaken with the hope 

 of gaining information on the subject of the dissociation of CO^. 



As already mentioned. Snow (loc. cit.) mapped the infra-red emission 

 lines of the alkali metals to about 2 fi, and found numerous lines just 

 at the end of the red, but no strong lines were located beyond 1.5 /a. 

 The present investigation deals with the question of the distribution of 

 emission lines (bands) in the infra-red, especially with the question 

 of presence of lines beyond 1.5 /u.. All the infra-red lines predicted 

 by our spectral series formulae end in the short wave-length just beyond 

 the red. Any information as to the presence of lines beyond this point 

 will aid in establishing these formulae upon a firmer, less empirical basis 

 than they have at present. From our knowledge of the radiation from 

 the "black body," which is most intense in the region of 1.2 to 2.5 ju, at 

 high temperatures, one would expect the emission bands at 2 fi, if there 

 be any, to be just as intense as those found by Snow at i fx. Other 

 points of interest which developed as the work progressed will be noted 

 in their proper places. 



