138 



Mr. H. Hartridge. Factors Affecting the [Nov. 1, 



Fig. 6. — To show the Effect of Changes in the Threshold or Intensity of Incident Light, 

 taking place on one side of a band only, on the mean wave-length of the band. At A 

 the value of the threshold on one side has been changed ; at B' and B the normal 

 intensity of the incident light has been altered, moving with it the threshold. 



which not only varies greatly with the individual, but is also modified by 

 the quality and time of the stimuli the retina has previously experienced.* 

 In the case of the observer G. W. and myself, a difference of 10 A.U. was 

 found in the wave-length of the bands ; this would point to an abnormality 

 in the threshold values of the green or yellow regions of the spectrum for 

 one of us. To test whether this was accompanied by any change in the 

 normal intensity of sensation, tests were carried out by Eayleigh's method,f 

 a yellow being matched with a combination of red and green ; we were unable, 

 however, to demonstrate any difference in our sensation of these colours. 



The variations demonstrated above in the visual measurements of wave- 

 length of absorption bands are also to be found when a photographic process 

 has been employed ; for colour, sensitive plates have a threshold at different 

 wave-lengths which depends on the dyes used to sensitise the emulsion. 

 Further, the actinic value of the light used to illuminate the slit of the 

 spectrograph should be as far as possible adjusted to be of uniform 

 intensity throughout the spectral region under examination. For this 

 purpose the cadmium arc and similar light sources which yield a number 

 of fine lines of very varying intensity are unsuitable. When a suitable 

 source has been chosen the true density gradient may be ascertained by 

 determining experimentally the threshold values at nine or ten evenly 

 distributed positions in the desired spectral region. It thus becomes possible 

 to obtain for the infra-red and ultra-violet portions of the spectrum density 

 curves that are the equivalent of spectro-photometric measurements in the 

 visible spectrum. 



It is interesting to observe that the factors of contrast, etc., here described 

 may in their relation to the phenomena of absorption bands be conveniently 



* ' Physiology of Special Senses,' Greenwood, p. 104. 

 t Bayleigh, ' Collected Papers,' vol. 1, p. 543. 



