464 FERREE AND RAND— LIGHTING 



experiments. (3) The range of brightness for this set of experi- 

 ments is quite a Httle higher in the scale than for that of the former 

 experiments with the same reflectors. This fact should be borne in 

 mind, for example, in comparing the shape of the curves for the 

 two sets of experiments in which highest order of brilliancy is 

 plotted against the results of the tests. In case of the present ex- 

 periments, for example, the curve begins at a point, 0.66 candle- 

 power per square inch, which is well above the knee of the former 

 curve. That is, the two curves are in general quite similar in shape 

 when they are compared for the same range of brightness values. 

 The former curve has, however, the greater regularity ; but in con- 

 nection with this fact it should be borne in mind that a variation of 

 the brightness factor in separation can be more nearly accomplished 

 with these reflectors when they are installed inverted than when they 

 are installed pendant. 



4. Opaque Direct Reflectors Differing in Dimensions, 

 Lining and Design. 



In this series of experiments the testing of pendant reflectors 

 was continued. Seven totally opaque reflectors differing in lining, 

 dimensions and design were used. In connection with the work of 

 this series of reflectors, two points of perhaps more than usual in- 

 terest may be noted: (a) By means of a modification of one of the 

 reflectors used. Reflector IV., made to reduce the brilliancy of the 

 opening, a field of view was given having the lowest maximum of 

 brilliancy of any that we have as yet been able to obtain, in an actual 

 lighting situation; and (&) we were able to test more eft'ectively than 

 in any lighting situation previously used, the importance of evenness 

 of surface brightness compared with evenness of illumination as a 

 factor influencing the ability of the eye to maintain its power of 

 clear and comfortable seeing. 



The opaque reflectors represent a more promiscuous selection 

 than the pendant translucent reflectors previously used. That is, in 

 case of the translucent reflectors, all but one, Reflector VI., were of 

 the same size and design, and the variations in the illumination 

 effects were obtained by varying the density of the reflector alone; 



