Investigation of Polarized Light. 227 



attached to the plate by supports of sealing-wax t, t f , carried a 

 fine copper wire stretched across the space between t and t 1 . 

 When the sparks had passed for a certain time between S and S', 

 the ultra-violet light from it withdrew from the plate a greater 

 or smaller quantity — according to the intensity of the light — 

 of the negative charge communicated to it. The fall of 

 potential was observed with an Exner's electroscope connected 

 to it. From the values Y and Y / of the potential before and 

 after illumination, the intensity of the light was calculated by 

 means of the formula J = log V/V 7 , previously established*. 

 At the beginning of each series of experiments the brightly 

 polished zinc plate stood at right angles to the incident rays 

 (a=0 c ), then the plate was turned through 50° right and then 

 left («= + 50°), and finally the initial position was restored. 

 At the beginning the potential was always 258 volts (25 di- 

 visions of the scale of the electroscope) ; the potentials Y 7 after 

 10 seconds' exposure are given in the following table : — 



13 April, 1895. 



Angle of incidence 0° +50-50 +50-50 +50-50 



V in divisions J Series I.. ..91 7*5 7'5 8-8 7*5 7'9 90 7'9 7-8 9'1 

 of the scale. ISeries II.... 8-9 7'9 7*4 8'5 8*0 8-0 8-9 7'5 7*5 90 



The mean of the observations gives : — 

 For normal incidence Y 7 = 8'9 scale-divisions = 117 volts. 

 For oblique incidence Y / ±50 = 7*7 scale-divisions =107 volts. 



Hence for the ratio of the photo-electric action we have 

 J^so, log 258 -log 107 _ 

 J'o log 258 -log 117" n * 



Hence the photo-electric activity of obliquely incident 

 ultra-violet light in the open air is also greater than that of 

 normally incident light, although the differences are much 

 smaller than those found for visible light with alkali- 

 metal surfaces in vacuo. We suspected that this was to be 

 accounted for by the imperfection of the surface of the amal- 

 gamated zinc, and repeated the same experiment with a sur- 

 face of mercury, altering the arrangement of the apparatus 

 so as to suit the condition that the kathode-surface must 

 remain fixed and that the light- ray must now be rotated. But 

 here also the differences were found to lie within the same 

 limits. Consequently either the ratio of photo-electric activity 

 of light polarized parallel and at right angles to the plane of 

 incidence is dependent on the wave-length, so that it ap- 

 proaches to unity for small wave-lengths, or other conditions 

 * Elster and Geitel, Wied. Ann. xlviii. p. 347 (1893). 



