228 THE ROYAL SOCIETY OF CANADA 



differences in intensity when the light stimulus was of e(iual intensity 

 at each point of observation. For this piu'pose observations similar in 

 principle to those already described, weie made upon this "Equivalent 

 of a Spectrum of i^iual Luminosity." 



Method of procedure. 



The method of procedure was fiist to adjust the telescope of the 

 spectrometer for light as near the ultra-violet as it was possible to 

 make an observation. This was of wave length •4()U/i. The plane 

 of the polariser was adjusted to make an angle of 35° with the plane 

 of the analj'ser. This gave to the light as viewed in the eye-piece a 

 luminosity equal to the luminosity of the spectuun where the light 

 waves were -414/^— the principal sections of the niçois then being 

 parallel. This latter luminosity was taken as the standard of bright- 

 ness since all portions observed were first reduced to this luminosity. 

 After the manner of former observations the light as viewed in 

 the eye-piece was allowed, to act upon the retina, the polariser rotated 

 and the angle after rotation noted. This angle was denoted by a. 

 If /9 denote the initial angle (35°) between the planes of the niçois, 

 then since a was taken to denote the angle between the planes after 

 rotation, (Cos-a— Cos'-/?) will be proportional to the amount of light 

 by which the initial intensity was increased to effect a noticeable 

 change in the sensation. This amount of light (Cos'-a — Cos-^), luill 

 be a fraction of the total intensity at the point considered. But since the 

 luminosity of each jDart was reduced to that of the standard, it is 

 reasonable to expect that, at those points where the spectrum is 

 brighter and therefore more intense, a smaller portion of the total 

 intensity will be required to produce a noticeable change in the sen- 

 sation than will be required where the total brightness is much less. 

 That is: the brighter the spectrum at the point of observation the 

 smaller the fraction of total intensity necessary to effect a noticeable 

 change in the sensation. Hence we may say that the luminosity of 

 each part of the spectrum is inversely proportional to the portion of 

 total intensity required to effect a noticeable change in the sensation — 

 the sensation in each case being the result of a light stimulus varjàng 

 only in wave length. Therefore, plotting the reciprocals of the various 

 values of (Cos-a — Cos^/9) as functions of the coi-responding wave 

 lengths a luminosity curve for the spectrum should be obtained. This 

 is shown in Fig. 3. 



In the above case the reciprocals of the various portions of total 

 light added cannot represent sensibility because in one instance we may 

 take the reciprocal of a large quantity of light at low intensity and 



