and Successive Contrast ivith Pure Spectral Colours. 191 



whole apparatus and head of the observer with black velvet. When these 

 most minute precautions were taken to prevent the admixture of red or 

 other light the results were exactly the same as before. The experiments 

 were conducted as follows: A region of pure violet, \ 4368-X. 4572, was 

 isolated in the Edridge-Green spectrometer, a deep blue-green glass quite 

 opaque to red being placed in front of the slit, so that no red light could 

 enter the instrument. A region of pure red. \6360-\6570, was isolated 

 in another spectrometer, deep ruby glass being placed in front of the slit so 

 that nothing but red light could enter the instrument. The eye was then 

 fatigued as before, one eye being vertically above the other, for 20 seconds, 

 and the after-image projected upon a narrow vertical band in the violet 

 region after turning the eyes round into the normal position, so that the 

 two images crossed at right angles. The result was exactly the same as 

 stated previously by us, the region of violet crossed by the after-image 

 appeared bluer and darker. 



It should be here noted that when the red band was intently regarded for 

 10 seconds and the eye then slightly moved (to another part of the same 

 telescopic field) a bright blue-green after-image was visible, although the only 

 light then being received by the eye was red light. 



The experiment with yellow light on a screen was repeated in the spectro- 

 meter with exactly the same result. Pure yellow light, A. 5820-X. 5870, was 

 isolated in one spectrometer, and red light, X 6360-X 6570, used to fatigue 

 the eye. The results were as before, the yellow appeared unchanged, or, 

 when the exciting light was comparatively intense, slightly greener and 

 darker in the region of the after-image, whilst a deep blue-green after- 

 image extended on either side. 



These experiments show that the stray light, mentioned by us in our former 

 paper, was of negligible amount ; for we have now obtained precisely the same 

 results when stray light was most rigorously excluded. Stray light, of amount 

 comparable with that in our previous experiments, is present in all spectro- 

 metric investigations unless precautions such as those described above are 

 taken. 



