340 AN AMERICAN TENT-BOOK OF PHYSIOLOGY. 



feeble than with strong illuminations. This phenomenon is illustrated by the 

 disappearance of a shadow thrown by a candle in a darkened room on a sheet 

 of white paper when sunlight is allowed to fall on the paper from the opposite 

 direction. In this case the absolute difference in luminosity between the 

 shadowed and unshadowed portions of the paper remains the same, but it 

 becomes imperceptible in consequence of the increased total illumination. 



Although our power of distinguishing absolute differences in luminosity 

 diminishes as the intensity of the illumination increases, yet with regard to 

 relative differences no such dependence exists. On the contrary, it is found 

 within pretty wide limits that, whatever be the intensity of the illumination, 

 it must be increased by a certain constant fraction of its total amount in order 

 to produce a perceptible difference in sensation. This is only a special case of 

 a general law of sensation known as Weber's law, which has been formulated 

 by Foster as follows : " The smallest change in the magnitude of a stimulus 

 which we can appreciate through a change in our sensation always bears the 

 same proportion to the whole magnitude of the stimulus." 



Luminosity of Different Colors. — When two sources of light having the 

 same color are compared, it is possible to estimate their relative luminosity 

 with considerable accuracy, a difference of about 1 per cent, of the total 

 luminosity being appreciated by the eye. When the sources of light have 

 different colors, much less accuracy is attainable, but there is still a great differ- 

 ence in the intensity with which rays of light of different wave-lengths affect 

 the retina. We do not hesitate to say, for instance, that the maximum 

 intensity of the solar spectrum is found in the yellow portion, but it is import- 

 ant to observe that the position of this maximum varies with the illumina- 

 tion. In a very brilliant spectrum the maximum shifts toward the orange, 

 and in a feeble spectrum (such as may be obtained by narrowing the slit of 

 the spectroscope) it moves toward the green. Hence changes of intensity are 

 associated with changes of color, and, as Haycraft l has observed, "we cannot 

 abstract 'brightness' from our sensations of light as we can abstract 'loud- 

 ness' from our sensations of sound." The curves in Figure 152 illus- 

 trate this shifting of the maximum of luminosity of the spectrum with vary- 

 ing intensities of illumination. The abscissas represent wave-lengths in 

 millionths of a millimeter, and the ordinates the luminosity of the different 

 colors as expressed by the reciprocal values of the width of the slit necessary 

 to give to the color under observation a luminosity equal to that of an arbi- 

 trarily chosen standard. The curves from A to H represent the distribution 

 of the intensity of light in the spectrum with eight different grades of illumi- 

 nation. This shifting of the maximum of luminosity in the spectrum 

 explains the so-called " Purkinje's phenomenon " — viz. the changing rela- 

 tive values of colors in varying illumination. 'Phis can be best observed 

 at nightfall, the attention being directed to a carpet or a wall-paper 

 the pattern of which is made up of a number of different colors. As 

 the daylight fades away the red colors, which in full illumination are 



1 " Luminosity and Photometry/' by John Berry Haycraft: Journal of Physiology, xxi. 126. 



