IN ITS RELATION TO THE EYE. 469 



lected, were given in total candlepower instead of candlepower per 

 square incli, they and the ratios based on them would correlate more 

 closely with the effects on the eye. That is, it is a well known law 

 in physiological optics that an increase in the area of a bright sur- 

 face functions to a certain extent as an increase in its brightness. 

 The opening of Reflector VII., for example, which gives the poorest 

 results for the eye, has a lower intrinsic brightness than has the 

 opening of Reflector VI. ; but a great deal more of it is visible to the 

 observer, also a great deal more of the lamp. When the surfaces in 

 question differ in area to any very great extent, it is obvious that a 

 measurement in candlepower per square inch is not an adequate 

 specification when effects on the eye are to be considered. The total 

 value in candlepower would be much more representative of the 

 comparative power of these surfaces to affect the eye. (b) The 

 angle of presentation to the eye. The openings of the different types 

 of reflector differed in their distances above the working plane, al- 

 though installed on the ceiling in every case. And (c) the number 

 of factors varied. Brightness is not the only one of the distribution 

 factors varied by considerable amounts by the different types of 

 reflector. Because of the irregularities shown by these curves, space 

 has not been taken to represent them here. 



For all the conditions represented in this and preceding sections 

 the work of testing was completed by determining for the different 

 types of reflectors the relative tendencies to produce ocular discom- 

 fort as was noted in our introductory chapter. Two cases were 

 made of this determination — one with the eye at rest, maintaining 

 no particular adjustment, and the other when it was at work. Space 

 has not been taken here for a statement of the results of these de- 

 terminations. They are given in tabular form,* with a comparison 

 in per cent, of the mean error of the determination and the change 

 produced by changing the conditions tested, in the references ap- 

 pended below. In these tables are included also for the sake of com- 

 parison results expressing the tendency of each type of reflector to 

 cause loss of ability to sustain clear seeing. A high correlation was 



4 Transactions of the Ilhiminating Engineering Society, 1915, X., pp. 497- 

 500; 1915, X., pp. 1114-1116; 1916, XL, pp. 1129-1131; 1917, XII., pp. 477-479. 



PROC. AMER. PHIL. SOC, VOL. LVII, FF, SEPT. 24, I918. 



