278 Dr. S. Bidwell. On Negative After-images, and 



feeble, the spot became uniformly black, remaining so until about 

 X 5000, at which point a blue nicker began to appear; at X 4700 the 

 spot had become steadily blue. Blue-violet and violet seemed to 

 illuminate the spot much more steadily than red. It was noticed that 

 as soon as the black spot became distinctly coloured the pulsative 

 image almost disappeared ; the weakness of the pulsative images 

 excited by light corresponding to the two ends of the spectrum may 

 therefore probably be accounted for by supposing that the negative 

 after-images become blended either with the primary images, or with 

 positive after-images, or perhaps with both, producing the effect of 

 white. 



It was found possible to observe these phenomena not only when the 

 zinc disc was spinning continuously, but even with a single properly 

 timed cycle of (1) darkness ; (2) colour-patch ; (3) white light; (4) dark- 

 ness. When the spectrum light was green, there appeared for a moment 

 a bright white disc with a perfectly black central spot, which was 

 surrounded by a well-defined purple annulus (as in fig. 9), the whole 

 being free from any visible trace of green. 



When a purple pulsative image excited by green rays was viewed in 

 the eyepiece by Method II, it was seen to be surrounded by a purple 

 corona, which extended considerably beyond the well-defined boundary 

 of the aperture in the diaphragm (F, fig. 3). Sometimes, indeed, when 

 the illumination was strong, the purple of the corona appeared to be 

 fuller or more saturated than that of the image itself. Moreover, a 

 purple haze of greater or less intensity always extended over the whole 

 field of the eyepiece. These phenomena are, of course, to be explained 

 by the "induced" blindness to green light which was demonstrated by 

 the black spot. 



Certain border effects of an entirely different character were also 

 observed. If the rays illuminating the circular patch seen in the eye- 

 piece were taken from the red, orange, or yellow regions of the 

 spectrum, the image appeared to be surrounded by a narrow red, or 

 rather crimson, border. Measurements of the composition of different 

 colour-patches which showed this effect include a red of A. 6420 to 

 X 6600, a reddish-orange of X 6200 to X 6280, an orange-yellow of 

 X 5890 to X 5990, and a yellow of A 5740 to X 5860 ; in the last case, 

 the border was less conspicuous, but still recognisable with certainty. 

 With a greenish-yellow patch containing rays from X 5650 to X 5750 

 no trace of the crimson border could be detected. It turned out that 

 these crimson borders could be seen when the intermittent white light 

 was screened off, though they were less easily visible against the dark 

 background than against the bright one. They evidently belong to a 

 class of phenomena discussed in a former paper,* in which it was 



* • Roy. Soc. Proc.,' vol. 60, p. 368, " On Subjective Colour-phenomena attending 

 Sudden Changes of Illumination." 



