NICHOLS— PHOSPHORESCENCE OF SULPHIDES. 273 



The change of slope between this process and the first process 

 so called in the curves of decay for these sulphides as observed dur- 

 ing the interval from two seconds onward is very great. In a rep- 

 resentative curve obtained by Mr. Carleton E. Power^* for example 

 his first process extends for nearly 50 seconds. The slope if com- 

 puted for a time scale such as that used in our measurements where 

 i/ioo sec. may be taken as a convenient unit, is scarcely perceptible. 

 The increase in the ordinate (/"^-) in passing from time .01 sec. to 

 .02 sec. in our second process or in Zeller's process is of the order 



-'02 



T -1/2 = 1-2; 



Power's first process would give a ratio of the order of 1.008. 



In other words during the first few hundredths of a second 

 after the close of excitation the intensity of phosphorescence falls 

 in each i/ioo of a second from unity to about .70 while after sev- 

 eral seconds, it falls in i/ioo second only from unity to .99. 



It seems probable, assuming continuity in the progress of the 

 decay, that if we had a complete curve of decay for one of these 

 sulphides, the knee between our second process and the first process 

 of the long-time curves would be found to lie somewhere between 

 o.io second and i.o second. If it occurs much earlier than o.io 

 second, Zeller would have discovered it; if much beyond i.o second 

 it should appear in the long-time measurements. In fact many 

 curves for the decay of phosphorescence by the latter process do 

 show a downward trend and Lenard, among others, has disputed the 

 linear character of the curve as we approach the origin of time. 

 The existence of at least four linear processes each of longer dura- 

 tion and lesser slope than the preceding may well account for the 

 difference of opinion. An observer determining the law of decay 

 as a whole by a method not taking cognizance of time intervals of 

 less than say i/io second, would describe as a curve what under 

 much more detailed study might be revealed as a succession of linear 

 processes. 



Owing to the overlapping of the components in the spectra under 

 consideration it is difficult to determine whether the group of equi- 

 distant bands are to be regarded as a unit, as in the case of the 



1* Power, C. E., Manuscript Thesis in the Library of Cornell University. 



PROC. AMER. PHIL. SOC. , VOL. LVI, S, JUNE 21, I9I7. 



