INSTRUMENT ILLUMINATION 



83 



and Williams (36) found that dial reading 

 speed increased when simulated day condi- 

 tions, i.e., white floodlighting at 45 foot 

 candles, were substituted for simulated night 

 reading conditions, i.e., ultra-violet illu- 

 mination. Loucks (50, 51, 52, 53) has data 

 which suggest that, in tachistoscopic expos- 

 ures, instruments which are more intensely 

 illuminated are read with fewer errors. 



Two sets of data have been collected for 

 the specific purpose of determining the ef- 

 fect of low illumination intensities upon in- 

 strument reading performance. One was 

 obtained in an experiment by Craik (19) in 

 which several forms of instrument illumina- 

 tion were tested. The other is reported in 

 a paper by Spragg and Rock (72) and de- 

 scribes the effect of intensity for white il- 

 lumination only. The conditions of these 

 two experiments were quite different. 

 Spragg and Rock used a serial reading task 

 wherein twelve 2.8 inch diameter dials were 

 read in succession at a 28-inch reading dis- 

 tance. Craik, on the other hand, took stop 

 watch times for single readings of a 3.75 inch 

 diameter dial. In terms of scale length, 

 Spragg and Rock's subjects were required to 

 read to the nearest 0.09 inches, Craik's to 

 the nearest 0.18 inches. The data of the 

 two experiments are summarized in Figs. 4 

 and 5. Craik's results are shown for reading 

 times only, Spragg and Rock's for reading 

 times and reading errors. 



It will be noted first that the time curves 

 for reading under white light in the two ex- 

 periments are not similarly located on the 

 brightness scale. The difference in their lo- 

 cation Tvath reference to the brightness axis 

 is to be accounted for at least in part in 

 terms of the differences in dial size and in 

 size of the figures and marks to be discrim- 

 inated. Observation distance may also have 

 been a factor. But, more significantly, the 

 curves are of different shape. Craik's 

 curves, as drawn, resemble curves obtained 

 by Richtmyer and Howes (65) and others 

 for the effect of illumination intensity upon 

 the rate of reading printed material. That 

 of Spragg and Rock, with a step and a second 



level portion, does not. At least two ques- 

 tions need answering. First, would Craik's 

 curves have had a shape similar to Spragg 

 and Rock's if he had used lower brightnesses 



& UV ON GREEN FLUO- 

 RESCENT PIGMENT 



O WHITE 



FOOT LAMBERTS 



Fig. 4. Dial-reading time as a function of bright- 

 ness for scales illuminated by white, red, and 

 ultraviolet light. Data from Craik (19) for not 

 more than six subjects reading under stop watch 

 timing. 



A READING TIME OATA 



FOCT LAMBERTS 



Fig. 5. Dial-reading performance as a function 

 of brightness for scales illuminated by white 

 light. Data from Spragg and Rock (72) for 20 

 subjects, each of whom made 100 readings at each 

 brightness level under instructions to read as 

 rapidly and as accurately as possible. 



or levels which were nearer together in the 

 region where the brightness effect begins? 

 The variability of the observed points 

 plotted for the several illuminants in Fig. 4 

 does in fact leave open the question as to 

 whether Craik drew on his graph a best 



