June 27, 1901] 



NA TURE 



A number of observations made with this apparatus showed 

 that the " pulsative after-images," as they may be called, 

 differed in several important respects from the ordinary negative 

 after-images seen upon a white ground after the gaze has been 

 fixed for some time upon a coloured object. Among other 

 things, they often appeared much more intense or saturated, 

 which for obvious reasons was rather surprising. And again, it 

 was found that the colours of the pulsative and of the ordinary 

 after-images were sometimes very different. The pulsative 

 after-images of red, of purple and of orange, for instance, were 

 all of nearly the same bluish-green tint, and those of yellow and 

 of blue were generally pink, showing considerable variations 

 from the true complements. These and other anomalies led me 

 to make some experiments with spectrum colours, which could 

 be blended into uniform mixtures of known composition, instead 



of with pigments, the apparatus employed for the purpose being 

 a simple form of Sir Wm. Abney's well-known " colour-patch " 

 apparatus. The experiments, of which a short account is here 

 given, are fully described in a paper recently communicated to 

 the Royal Society {Proc. Roy. Sot., May 23). 



A short pure spectrum, some three or four inches long, is 

 projected, by means of a spectroscope collimator, a prism and an 

 achromatic lens, upon an arrangement known asa *' slit-screen." 

 This consists of a wooden board having an oblong window over 

 which slide three brass slit-frames carrying adjustable slits. 

 The two outer slit-frames are attached to sliding shutters, which 

 serve to cover such portions of the window right and left of the 

 slit-frames as would otherwise be open to the light ; the spaces 

 between the middle slit-frame and the two outer ones are closed 

 by opaque black ribbons, constituting miniature spring-roller 

 blinds. Each slit-frame can be moved independently to any 

 desired position and clamped by a set-screw. When it is neces- 

 sary to use larger portions of the spectrum than can be trans- 

 mitted by the slits, the slit-frames and their appurtenances are 

 removed from the screen and the spectrum is dealt with by 

 means of one or more thin metal plates, which are inserted into 

 a second pair of guides fixed on the other side of the screen. 

 Each of these guides has three parallel saw-cut grooves in it, so 

 that plates sliding in different pairs of grooves may be made to 

 overlap one another, and thus screens or openings of almost any 

 desired width may be provided with very little trouble. The 

 selected portions of the spectrum pass through the slits to a 

 large lens, which projects upon a white screen the image of a 

 circular aperture in a diaphragm placed just in front of the 

 prism. This image, which is ij or 2 cm. in diameter, consti- 

 tutes the colour-patch ; it is illuminated by a uniform mixture of 

 the spectrum-rays transmitted by the slit-screen. When the 

 wave-lengths of the light are to be determined a screen of ground 

 glass is put in the place of the opaque white screen and the slit 

 of a spectroscope is brought near the bright image on the glass. 

 To produce a bright ground upon which to see the after-images 

 a beam of white light, derived from the same electric arc as the 

 spectrum, is directed upon the screen. The white light passes 

 through the aperture of an iris-diaphragm, and a lens is placed 

 to project an image of the aperture upon the screen. The 

 " white-light disc " so formed is concentric with the colour- 

 patch, and in most cases of slightly greater diameter. 



In the path of the two beams of light illuminating the colour- 

 patch and the white-light disc is placed a zinc disc having two 

 apertures, one near the centre to admit the spuct rum-rays, the 

 other near the circumference to admit the white light. This is 

 caused to turn about five limes in a second, and the apertures 

 are so arranged that the sequence of phenomena produced upon 

 the screen may be as follows ; — The colour-patch is projected 



NO. 1652, VOL. 64] 



for about 1/40 second, then it is extinguished and immediately 

 succeeded by the white-light disc ; this remains for 3/40 second, 

 and is followed by an interval ol darkness lasting l/io second, 

 after which the colour-patch reappears and the cycle begins 

 again. The appearance presented to the eye when the true 

 colour of the patch is green is that of a purple disc surrounded 

 by an annulus of flickering white. Other colours, of course, 

 produce pulsative images of different hues. It is sometimes 

 better to view the pulsative image directly by means of an eye- 

 piece instead of receiving it upon a screen ; two or three 

 methods by which this can be effected are described in the 

 paper. 



Perhaps the most interesting of the various effects observed 

 with this apparatus is one which appears to throw some light 

 upon the origin of the pulsative image and to show why the 

 true colour is so completely lost. It was noticed that when the 

 white-light disc was made a little smaller than the colour-patch, 

 the pulsative image, which was in this case of the same size as 

 the white-light disc, was surrounded by a dark ring. This ob- 

 servation led me to make what is called in the paper the "black 

 spot " experiment. A circular piece of tinfoil was gummed to a 

 glass plate which was placed behind the iris-diaphragm, a 

 sharply defined black spot being thus formed in the middle of 

 the white-light disc, as indicated in Fig. 2, where the outer 

 circle represents the white-light disc, the shaded circle the 

 colour-patch, and the inner one the black spot upon the white- 

 light disc. The diameter of the black spot was made, after 

 several trials, o'6 cm., or rather more than one-third of the 

 diameter of the colour-patch. Suppose the colour-patch to be 

 green. When the apparatus is worked, the patch becomes 

 purple; the site of the black spot, being strongly illuminated 

 five or six times in a second by green light, might be expected 

 to appear green ; but it remains perfectly black ; no trace of a 

 flicker of green light can be seen upon it. This induced blind- 

 ness is most conspicuous when the light is green or yellow ; it 

 does not occur at all with extreme red or with violet light ; nor 

 does it generally occur if the luminosity of the colour-patch is 

 reduced below a certain limit, that of the white-light disc remain- 

 ing constant. A colour of feeble luminosity can be seen upon 

 the black spot, while a brighter cannot, which is a paradox ; 

 and it was noticed that as soon as the spot became distinctly 

 coloured, the pulsative image almost disappeared. Absence of 

 colour from the black spot is essential for a good pulsative 

 image. 



It is clear that we have here an example of what I have called 

 in other papers a border effect ; in certain cases light has the 

 power of exciting some action just outside the boundary of the 

 image projected upon the retina. The black spot is, of course, 

 merely a device for exhibiting in a convenient manner a border 



effect which extends for about half a degree beyond the white 

 light impression. The experiment proves conclusively that, under 

 the given conditions, white light has the power of restraining the 

 visual sense-organs adjacent to those upon which the white light 

 actually falls from responding to the green stimulus. This is of 

 importance as indicating what occurs within the area illuminated 

 by white light, for it would seem to follow a fortiori that the 

 sense-organs which are directly acted upon by the white light 

 must be similarly incapacitated from evoking any green sensa- 

 tion. In the formation of the pulsative image, then, it is not the 

 fact, as generally believed, that the green sensation is produced 

 for a moment and then swamped by a more powerful white one 

 so completely as to escape notice ; it actually never comes into 

 existence at all. Nevertheless, the effects of fatigue by green are 

 exhibited, and the physically white ground is seen as purple. 

 Only one possible explanation of the phenomenon has yet 



