INTIMATE STRUCTURE ON COOLING. 



71 



components characteristic of the spectrum at low temperatures and the 

 simultaneous fading away of those dominant at +20. The same 

 explanation applies equally well to the potassium and rubidium double 

 chlorides. In the case of csesium uranyl chloride the relations are 

 complicated by the further resolution of these components, so that the 

 connection with the original complexes is less easily traced. 



To indicate the general character of these resolutions and the 

 apparent temperature shift which results therefrom, the positions of 

 the bands of group 6 at 185 are plotted for all four chlorides (see 

 fig. 60). Intensities of the 185 bands are indicated roughly by the 

 height of the lines. The corresponding crests of the bands at +20 

 are represented by dotted lines. Group 6 was selected because it 

 offers better examples of the further breaking-up of the components 

 and of other phases of the process of resolution than do groups toward 

 the red in which resolution is progressively 

 less complete. 



Two questions which were left undetermined 

 in the study of the spectra at +20 may be 

 regarded as settled by these measurements of 

 the bands at 185. 



(1) That the intervals are not the same 

 for all series in a given spectrum is clearly 

 established. For example, the components 

 Ci, Cz, which take the place of the C bands in 

 all four spectra, have distinctly different in- 

 tervals, i. e., 84.00 for d and 82.75 for C 2 . It 

 is noteworthy that C 2 , which becomes the 

 crest of the group in place of C also, has the 

 small interval. It might be questioned 

 whether these so-called components are not 

 merely accidental neighbors rather than 



products of the same vibrating system, but for the fact that they 

 are present in all the spectra and have very nearly if not precisely the 

 same relative positions to each other in all. 



(2) The average interval of all series in the spectrum of the csesium 

 chloride (82.80) at +20, which causes the notable displacement of the 

 bands of that substance, becomes 83.44 when we take the average of 

 the intervals of the bands at 185. That is to say, it is, within the 

 errors of observation, the same as the general average for the other 

 salts. On the basis of the measurements at low temperatures (see 

 table 24), we must conclude that the four double chlorides have approx- 

 imately the same average frequency interval. 



The averages given in table 24 are obtained from the data of table 

 32, which contains the frequencies of all the fluorescence bands observed 

 in the spectra of the four double chlorides when excited at the tempera- 



1900 



FIG. 60. 



1950 



