462 FERREE AND RAND— LIGHTING 



shown the gradation of surface brightness in the manner described 

 in the preceding section. And in order to ascertain which of the 

 brightness specifications — order of magnitude of highest brilHancy, 

 average brilHancy, ratio of highest to lowest order of brilliancy, 

 ratio of highest order of brilliancy to average brilliancy, ratio of 

 average to lowest brilliancy, ratio of highest order of brilliancy to 

 brightness at point of work (brightness of test object and reading 

 page), etc. — correlate most closely with the results for the tendency 

 to cause loss of power to sustain clear seeing, curves have been con- 

 structed in which a number of these features were plotted against 

 the results of the tests. These curves are given in Charts V. and VI. 

 In Chart V., B, per cent, loss in power to sustain clear seeing is 

 plotted against the highest order of brilliancy that varies by any 

 considerable amount from installation to installation, namely, the 

 brightness of the reflector — outer surface and opening. In Chart 

 VI. are grouped the remainder of the curves. 



Three points may be noted perhaps with reference to these 

 charts : ( i ) The prismatic reflectors, which difTer in design from the 

 rest of the series, more or less conspicuously fall out of the curve in 

 every case but two. The efifect of difference in design on the smooth- 

 ness of the curve comes out especially in the results with the opaque 

 reflectors (shown in the report of the next series of tests), in which 

 case there are marked differences in both size and design. All of the 

 curves plotted on the above bases are very irregular in case of these 

 reflectors with the exception of separate curves for three which are 

 similar in design. For a statement of the probable reasons for this 

 inequality see this paper, pp. 468-9. (2) The greater regularity of 

 the curves is rather strikingly marked in which the highest order of 

 brilliancy that varies by considerable amounts or the ratios in which 



eye. For the translucent reflectors used in these tests, the best results have 

 been obtained with the reflectors of medium density. The reverse of this 

 was true, it will be remembered, when the same reflectors were installed in- 

 verted. The highest brightnesses when these reflectors are installed pendant 

 are the filament spots on the lamps. Only very small areas of these spots 

 are visible, however, and their brightness and the brightness of the lamp 

 differ so little from installation to installation as to be, in all probability, of 

 relatively little consequence in a comparative study of effects on the eye. 

 The significant variables are thus the brightnesses of the outer surface and 

 opening of the reflector. 



