Prof. S. P. Langley on Energy and Vision. 



21 



The probable error of a series of ten readings of fine print, 

 under the actual conditions of observation with a (feeble) 

 standard luminosity, is determined for two of the observers 

 as follows : — 





Violet Light. 

 \=0'^-40. 



Orange-yellow light. 



Scarlet light. 



Probable Error of one observation. 



R W. V 



Per cent. 

 5-53 



Per cent. 

 1-76 



Per cent, 

 3-14 



E. M 



7-69 



2-51 



2-86 





Probable Error of mean. 



R W. V 



1-75 0-56 



0-99 



E. M 



2-44 0-80 



0-90 









The meaBurements with violet light were made June 19, 1888, " sky hazy 

 blue, thin but uniform cirrus haze." Those at wave-lengths C^'GO and 0^-65 

 were obtained on June 20, 1888, " sky hazy blue with cumuli, haze not so dense 

 as on the 19th but possibly less uniform." 



For a large part of the spectrum the probable error of a 

 single reading does not exceed 4 per cent., but the error may 

 considerably exceed this for the violet rays, the eye req^nring 

 a much longer time to regain its sensitiveness for light of this 

 colour than for any other, so that for measurements in this 

 region an hour's stay in the darkened room is none too much 

 to develop the full power of an eye which has recently been 

 exposed to the full sunshine. 



Time required for Vision. 

 In connexion with the photometric measurements the time 



Photometry by the reading of fine print gave for the same observer 

 Violet, sensitiveness of eye =0' 104,000 

 Green „ „ =5-790,000 



Scarlet „ „ =0-036,000 



Crimson „ , „ =0-000,070 



unity being the sensitiveness for yellovf light : and the relative efi*ect by 



this method is : — 



Violet (^-40) : green ('^•55) = 1 : 56. 

 Scarlet ('^-65) : green ('^•55) = 1 : 160. 

 Crimson f-75) : green ('"•55) = 1 : 83,000. 



