506 On Light- Sensations and the Theory of Forced Vibrations. 



I have, since 1910, in the practical classes which I have conducted for 

 Prof. Gotch, made a considerable number of people arrange from 7 to 10 

 strips of white paper at distances varying from \ metre to 2 metres from 

 a candle in a dark room, so that viewed from a certain point they presented 

 a series of apparently equal gradations of luminosity. The results were very 

 instructive, and in one respect unexpected. Most of the men repeated the 

 experiment two or three times in order to get what they considered a good 

 result. Almost without exception the first attempt of each person showed 

 considerably higher values for the ratios at the two ends of the series. In 

 the majority of cases the difference was less in subsequent experiments, but 

 it was evident that, to the unbiassed judgment, the eye is less sensitive to 

 differences between the brightest objects visible and also between the 

 faintest objects visible at any one time than between those that are 

 moderately illuminated. 



Although practice reduces this divergence from Weber's law it does not 

 do away with it. I still make the ratios at the two ends of the series higher 

 than those near the middle of it. But with monochromatic red light any 

 series that looks right to me under a feeble illumination looks right also in 

 light 100 times as strong, as it should do according to Weber's law. 



I am inclined therefore to accept Fechner's law with the same caution 

 that Waller expressed in reference to the relation between strength of 

 stimulus and the retinal currents of the frog's eye. " The curve plotted 

 from the data comes out concave towards the abscissae and not unlike an 

 ordinary logarithmic curve." 



If, in judging the minimum perceptible difference of brightness, we 

 instinctively make use of the shunt function only, that would lead to some- 

 thing not far removed from the logarithmic law, within the range covered by 

 the shunt. 



The greater part of the experimental work connected with this paper has 

 been done in the Physiological Laboratory, Oxford, and the expenses have 

 been defrayed out of the Government Grant Fund. 



