224 THE ROYAL SOCIETY OF CANADA 



the prism might be perpeiidicuhir to its face. In this way it was thought 

 to avoid a variation in intensity of light througli reflection from the 

 surfaces of the prisms. 



The spectrometer used was of the Hilger Automatic type with four 

 prisms equivalent to three sixty degree prisms and giving a dispersion 

 slightly in excess of twelve degrees. 



I. 



Each characteristic of the visual response to radiation as outlined 

 in the commencement of this paper, was treated as a separate investi- 

 gation. Observations were made to show: 



1. Whether sensibility to slight difference in wave length has 

 two pronounced maxima, one in the yellow and one in the green, and 

 two light maxima in the extreme blue and red. 



2. AMiothcr these maxima vary with the intensity of the radiation 

 used. 



1. In the investigation concerning spectra of varying intensity, 

 other conditions remaining constant, spectra of six different intensities 

 were examined. Only observations made with my own eye (the right) 

 which seems quite normal as to color sensations, are considered in this 

 paper. Upon the spectrum of maximum intensity in all parts, which 

 for present purposes, has been represented by unity, and upon a spect- 

 rum of intensity • 25, observations were made to determine the minimum 

 decrease in intensity by which the initial intensity was changed in effect- 

 ing a change in sensation that was just perceptible. In the case of 

 spectra at all other intensities considered, the minimum increase in the 

 original intensity necessary to produce a just perceptible change in 

 the sensation was determined. 



Method of procedure. 



The principal sections of the niçois were horizontal and so gave 

 the brightest spectrum obtainable with the chosen arrangement of ap- 

 paratus. Observations were made upon thirty-two portions of this 

 spectrum; these are given in Table I. 



Light of wave length -42// was first observed. This Avas allowed 

 to act upon the retina for four seconds. Then the polariser was rotated 

 at a uniformly rapid rate; an effort was made to keep the rate at which 

 the polariser was rotated, approximately constant for all observations. 

 As soon as a change in the sensation of light was perceived the rotation 

 was stopped. The angle through which rotation had taken place was 

 noted. This was denoted by a; since unity had been taken to repre- 

 sent the amount of light passing through the prisms before the polariser 



