the Light given off by a Phosphorescent Surface. 211 



from the distance of tHe burner. No absolute standard was 

 attempted, as the light of the tissue would vary with its 

 thickness, probably with the angle at which it was observed, 

 and with the exactness with which the colour of the paint was 

 imitated by means of the glass and solution. The second 

 column gives the time at which the observation was taken, 



Table. 



Observed Light. 



Observed Time. 



Calculated Time. 





m. s. 



m. s. 



10O 







0* 



56-5 



18 



19 



25-1 



45 



1 8 



141 



1 33 



2 11 



904 



2 34 



3 25 



6-10 



5 



5 0* 



2-89 



10 



9 59 



1-34 



20 



19 53 



•782 



30 



31 52 



•604 



40 



39 41 



•433 



50 



53 



•376 



60 



60 0* 



•285 

 1 



80 



76 14 



I also made an attempt to calculate a curve which should 

 as nearly as possible coincide with these results. The form 



A 



of the curve wdiich naturally suggested itself was ^ = l c ; 



t + B 



t = time, 1 = light; A, B, and C being constants. A curve 

 of this form to pass through three points (*) was cal- 

 culated : A = 26*1, B = -498 minute, C = -86. That is to say, 

 the rate of decrease of the light varies as the light to the 

 power of 1*86 ; and at that rate the light would have 

 been infinitely great 30 seconds before the first observation. 

 Taking the illuminations as obtained from the observations, 

 the time of the observations was calculated according to the 

 above formula ; and the result is given in the third column. 

 Three of the observed times differ from the calculated times 

 more than can be accounted for by errors of observation; and 

 it appears that the equation does not give quite the correct 

 curve. From this and from another series of observations, it 

 appears that in its lower parts the curve is very nearly of 



A 



the form 7375 = ^ ; that is to say, that the rate of decrease of 



the light varies as the square of the light. 



By comparing several series of observations it appears to 



