218 



STUDIES IN LUMINESCENCE. 



where 



T = 



Vt + 



-mt 



and 7 = A 'B 



The first factor alone plots as a straight line. The deviation from 

 linearity in the curve for I~^ is therefore determined by the second factor 

 T. Putting m = i we have computed T as 

 a function of / for several values of 7, the 

 results being plotted in Fig. 190. For 

 values of 7 ranging from o. 1 to 0.5 the 

 curves are of such a character as to cor- 

 respond with the experimental curves for 

 /~*. But for smaller values of 7 a double 

 curvature is shown which is not found in 

 any of the curves that we have deter- 

 mined. There is, however, an indication 

 of such a double curvature in some of the 

 results of Pierce. It is clear that disturb- 

 ances, due, for example, to irregularities 

 in the distribution of the active material, 

 might so alter the early part of the curve 

 as to eliminate any peculiarities of this 

 nature. In fact, there is at present so 

 much uncertainty regarding the relative 

 importance of the different factors that 

 influence the form of the curve for small 

 values of / that it is difficult to reach any 

 definite conclusions concerning this part 



of the curve. Experiments bearing upon My + e- 



the distribution of the active material, the rate of diffusion of the ions, and 

 related matters are greatly needed. 



ao 



60 



4-0 



Seconds 

 Fig. 190. 



Curves showing relation between T and I for 

 different values of 7 where 



I 



T= 



J- 



HYSTERESIS, TEMPERATURE EFFECTS, ETC., EXPLAINED BY IONIC GROUPING. 



To account for the hysteresis exhibited by phosphorescent substances, 

 in other words, the effect of a previous exposure upon the phosphorescence 

 produced by a given excitation, it is necessary to consider the essential 

 difference that probably exists between the groups formed by the union of 

 positive ions with neutral molecules of the active substance and those 

 formed by the attachment of positive ions with molecules of the solvent. 

 For the sake of brevity, as well as for the reasons that will appear shortly, 

 we shall refer to the former as favorable groups and to the latter as unfavor- 

 able groups. 



We have already introduced the assumption that light is produced by 

 the recombination of a negative ion with a favorable group, while the re- 

 combination of the unfavorable groups, at least in some cases, gives out no 

 light. A difference is to be expected also in the behavior of the neutral 

 molecules that result from recombination in the two cases. A favorable 

 group consists in a positive ion attached to at least one neutral molecule 

 of the same sort. It seems natural to expect that forces similar to those 

 that hold together the molecules of a crystal may cause this grouping to 



