their Relation to certain other Visual Phenomena. 2*19 



shown that when a bright image is suddenly formed upon the retina 

 after a period of darkness, the image generally appears for a moment 

 to be surrounded by a narrow red border. The paper referred to con- 

 tains an account of an experiment* demonstrating that when the 

 bright object producing the image was looked at through variously 

 coloured glasses, the red border did not appear unless the glass used, 

 when tested spectroscopically, transmitted red light, and it was 

 suggested that the phenomenon was due to sympathetic excitation of 

 the " red nerve fibres " lying immediately outside the portion of the 

 retina exposed to the direct action of the light. The orange and 

 yellow glasses employed in the experiment referred to of course 

 transmitted red light ; it is interesting to find that the pure orange and 

 yellow rays of the spectrum, of wave-length not necessarily exceeding 

 about X 5800, are competent to give rise to the same red borders. 



These effects can be exhibited equally well by Methods I and II, the 

 observations being rendered much easier by the aid of a device 

 described in the former paper. A darning needle, blackened with 

 camphor smoke, is cemented vertically across the opening in the 

 diaphragm F, fig. 3, dividing the bright disc which is projected upon 

 the screen or seen in the eyepiece into two equal parts. Each half 

 disc then has its red border, and, if the intervening space is sufficiently 

 narrow, the red borders along the two contiguous vertical edges meet, 

 or possibly even overlap, with the result that the focussed image of the 

 needle should appear to be red. This was the case when the slit was 

 placed in any part of the spectrum between the extreme red and the 

 greenish-yellow. With the slit in the greenish-yellow itself the image 

 of the needle appeared to be almost colourless, but as the full green 

 was approached the colour became a rather dark shade of blue-green,, 

 and. remained so until the slit reached about X 4500, near the beginning 

 of the blue-violet, when the needle again became colourless. In a 

 colour-patch formed of the pure blue rays from A 4600 to A 4725 the 

 contrasted blue-green hue assumed by the image of the needle was 

 strikingly conspicuous. The border-colour in question cannot easily be 

 observed unless the intensity of the illumination is within certain 

 limits j for, as in the case of the red borders which were discussed in a 

 former paper,! the blue-green hue becomes transformed into its comple- 

 mentary if the light is very strong, and the needle appears reddish. 

 For the green part of the spectrum it is especially necessary that the 

 illumination should be very carefully adjusted ; indeed, the phenomenon 

 would probably never have been noticed at all with green light if its 

 remarkable appearance when the light was pure blue had not first 

 attracted attention. For the more refrangible part of the spectrum it 

 is desirable to place in front of the collimator-slit a piece of blue glass 



* Experiment IV, loc. cit., p. 372. 



f ' Koy. Soc. Proe.,' 1897, vol, 61, p. 268. 



