﻿620 Profs. 0. W. Eichardson and F. J. Rogers on 



Curve 5. This was taken somewhat later than curve 4, 

 with the same flat strip set so that the light was incident at 

 approximately 15°. Owing to the inclination of the strip 

 the effective width of the beam of radiation falling on it was 

 less than in the previous cases. Allowing for this, the unit 

 of N is 1*98 X 10" 6 coul./cal. and the maximum value of N 

 (atz> = l'37x lO^sec." 1 ) 



= 9-7 x 10 ~ 5 coul./cal. 



Sodium floe. cit. p. 563). 



Curve 1. The observations used in constructing this curve 

 were commenced about one hour after distilling the sodium 

 on to the strip and took over an hour to complete. There 

 was a rapid photoelectric fatigue. This was corrected for by 

 extrapolation ; so that all the observations were reduced to 

 the values corresponding to the instant of the first observa- 

 tion. This extrapolation cannot be made quite exact since 

 different parts of the curve decay at different rates. This 

 difference was allowed for in so far as it could be ascertained ; 

 so that Curve 1 may be taken as representing the emission 

 from a surface of distilled sodium one hour after distillation. 

 The unit of N for this curve is 2*56 x 10~ 5 coul./cal. 



Curve 2. This curve represents the actual condition of 

 affairs 2-1 hours after distillation. There is some doubt, 

 arising from a possible inaccuracy in one of the data used 

 in correcting for fatigue, about the position of the curve for 

 values of v X 10 ~ 13 greater than 125. Relative to the rest of 

 the curve on the left-hand part of the diagram, the point at 

 v X 10~ 13 = 125 is correct ; but the true position of the maxi- 

 mum at vX 10~ 13 = 133 may be 10 per cent, below the value 

 on the curve as drawn, and the end point at z>xl0~ 13 =150 

 as much as 30 per cent, below the value shown, intermediate 

 points dropping by regularly increasing percentages. The 

 unit of N for this curve is 1*31 X 10~ 5 coul./cal. 



Curve 3. This represents the state of affairs 18 hours 

 after preparation, when the rate of fatigue was extremely 

 slow. The unit of N for this curve is 6*25 x 10~ 6 coul./cal. 



The values of the quantity of electricity emitted per unit 

 energy of incident (isotropic) radiation are collected in the 

 following table. The values at t- have been derived on 

 the assumption that the decay of the photoelectric effect is 

 exponential for the first 2£ hours. As there is a certain 

 amount of experimental evidence indicating that under 

 certain circumstances, not yet properly understood, the 

 emission may show an initial increase followed by decay 

 according to an exponential law — as well as other types of 

 behaviour under different conditions — these numbers may be 



