PHYSICS. 367 



cence. This is due to differences in the rate of decay of the components 

 of a mixed luminescence, and affords another and simple criterion, i. e., 

 absence of color change during the phosphorescence period. This can 

 be tested, even for substances of exceedingly brief phosphorescence, 

 such as the uranyl salts, by the use of the disk phosphoroscope. 



The absence of color change of the fluorescent light, when lumines- 

 cence is produced at low temperatures. Such changes are marked in the 

 case of heterogeneous luminescence and afford a conclusive indication. 



By such criteria we are led to the conviction that the uranyl spec- 

 trum is a unit, appearing in its entirety, however produced, and dis- 

 appearing as a whole when excitation ceases. This is a unique charac- 

 teristic, so far as our present knowledge goes. It is certainly not true 

 of the fluorescent dyestuffs, of the phosphorescent sulphides, of the 

 rare earths, or of the luminescence of vapors. 



Another fact, even more impressive and significant, w^hich was fully 

 recognized only after the completion of our detailed studies of the 

 spectra of numerous salts under varying conditions, when we were in 

 position to view them as a whole, and to make comparisons based upon 

 adequate data, is this: 



There is but one spectrum, the uranyl spectrum, common to all uranyl 

 salts. Of a complexity which has not as yet been completely resolved 

 and analyzed, its essential structure is always the same. This state- 

 ment is based upon the following characteristics typical of all the 

 spectra thus far examined. 



(1) All uranyl spectra have the same number of equidistant fluores- 

 cence bands. 



(2) This set of bands occupies in all cases the same region of the 

 spectrum, lying, roughly, between 0.6500)u and 0.4800^. 



(3) The distribution of intensities is always the same, rising from 

 the merest visibihty in the red to a definite crest and diminishing more 

 rapidly towards the violet. 



(4) Fluorescence and absorption always overlap in the so-called 

 reversing region. 



(5) The frequency interval for absorption series is always smaller 

 than for fluorescence series, and the ratio is nearly the same for all 

 compounds. 



(6) Fluorescence and absorption bands are always complex, although 

 not generally visibly so at +20°. Resolution of the bands into groups 

 always occurs on cooling in the case of crystalline compounds. Liquid 

 or non-crystalline preparations are not resolved by cooling. 



Accompanying these universal attributes are numerous minor and 

 perfectly definite variations, which tend to obscure but never actually 

 conflict with the general uniformity of type. These enable one to 

 identify, with certainty, the spectra of the various compounds, espe- 

 cially when excitation occurs at low temperatures. 



