262 ANNUAL, EEPORT SMITHSONIAN INSTITUTION, 19 3 



pitches, frequencies or wave lengths of the sound with the source. In 

 the same way, the frequencies or wave lengths of light emitted or 

 absorbed by a molecule are characteristic of the molecule, and are 

 determined by its construction. To get the most information about 

 the characteristics of the sound and relate these characteristics to the 

 construction of the source of the sound, the ear is inadequate, and an 

 apparatus must be provided for determining just what frequencies or 

 wave lengths are present. So in the study of light we must have a 

 machine which will distinguish between the different wave lengths, 

 or colors, in a more quantitative fashion than the eye is capable of 

 doing. Such an instrument is known as a spectrometer. Looking 

 into a spectrometer we see the light spread out into its constituent 

 colors, the red on the left and the blue on the extreme right, with all 

 the intermediate wave lengths or colors arranged between them in an 

 orderly fashion. Unfortunately, the light which gives the most 

 information in regard to these large complicated molecules is of too 

 long wave lengths to be seen by the eye. It is necessary, therefore, 

 to substitute for the eye a very sensitive electrical thermometer. By 

 turning a single dial successively different wave lengths are caused to 

 fall upon this sensitive thermometer, which by its change in tempera- 

 ture, records the amount of each wave length present. 



A spectrometer is very similar in its analysis of light to a radio 

 receiving set in its analysis of the electrical oscillations which it 

 picks up. As one turns the dial slowly, the instrument responds 

 first, say, to 200 meters, then to 201, 202, and so on to the other ex- 

 treme of its range, say, 800 meters. If a station is broadcasting on 

 a wave length of 304 meters, as one comes to the dial number 302 

 he begins to hear it. It comes in strongly at 304, and fades out at 

 306. The range over which it may be picked up on the dial is asso- 

 ciated with the selectivity of the set. In the same way a spectro- 

 meter responds to the different wave lengths of light. If a single 

 strong wave length is being " broadcast " by a mercury arc, and that 

 wave length is 1/x, or one-millionth of a meter, the spectrometer will 

 pick it up weakly at a dial reading a little less than Ifi, show a 

 maximum response at 1/x, and as one continues to turn it, gradually 

 lose it as 1/a is passed. The range over which it responds depends 

 upon its selectivity, or resolving power. As a matter of fact, the 

 mercury arc broadcasts not simply on one wave length, but on sev- 

 eral, so that as one turns the dial, he picks it up at several different 

 points. The accompanying illustration. Figure 1, shows the differ- 

 ent wave lengths of radiation received from a mercury arc. The 

 widths of the bases of these peaks depend upon the selectivity, while 

 the heights indicate the intensity or strength of the signal. If the 

 selectivity were less the bases would be broader, and the signals of 

 very nearly the same wave length would be confused and not ob- 

 served as a separate effects, just as in a poor radio one gets several 



