PSYCHO-PHYSICAL METHOD 127 



second a very considerable decline takes place ; and this 

 decline affects one's judgement of the brightness of the 

 total impression, although one may be unable to distin- 

 guish the phases of declining brightness. Now in Martius' 

 experiments one of the two impressions to be compared 

 was in each case one of considerable duration, i.e. it 

 appeared less bright than it should have done, with the 

 result that a duration was assigned to the action-time 

 much briefer than its true duration. 



We have, then, to devise a procedure which shall 

 avoid all these three sources of error, and this seems to 

 be accomplished by the aid of the simple apparatus that 

 we have here a rough imitation of the laboratory 

 apparatus. A large disc, with two windows of variable 

 angular width at opposite ends of one diameter, rotates 

 at a constant rate in the focus of a ray of light from the 

 lantern, and in the plane perpendicular to the path of 

 the ray. 



As each window passes across the ray the light passes 

 through it, falls upon the screen, and is reflected to your 

 eyes during a fraction of a second, whose duration is 

 easily calculated when we know the rate of rotation of the 

 disc and the angular width of the window ; e. g. if the 

 disc makes i revolution in 3 seconds, and the angular 

 width of the window is 10, then the duration of ex- 

 posure and of the action of the light on your eyes is 



= sec. By altering the width of the windows 



360 12 



and reading off their width on a scale of degrees, the 

 durations of the alternate flashes may be modified to 

 any required extent. 



The first point I want to demonstrate is that a flash 

 of very brief duration has not time to develop its full 

 effect in the retina of the eye. We make the windows 



