September 6, 1894] 



NA 7 URE 



467 



centre, there appears upon the while sector, near to its leading 

 edge, a well-defined dark band, which is separated from the 

 black ground of the disk by a similar white band. The angular 

 extension of the dark band increase-; with the speed of rota- 

 tion, so that it always takes the same time to pass over a fixed 

 point on the retina ; it begins about one-sixty-fifth or one- 

 seventieth of a second after the first passage of the white, and 

 lasts sensibly the same time. He goes on: — " The dark band 

 is in fact only a kind of reaction of the retina after the luminous 

 excitation, a reaction which can be demonstrated in a totally 

 different manner. I have found that if an instantaneous lumi- 

 nous excitation is produced in complete darkness the sensation 

 appears to be reduplicated ; shortly after its first generation it 

 seems to disappear, and then manifest itself again. This is the 

 case, for exam(ile, when a single discharge from a Ruhmkorfif 

 coil is passed through a Crookes or Geissler vacuum tube, or 



simply, but less obviously, through the air There is, 



then, in this last experiment, as in the first, a negative reaction 



of the retina under the influence of excitation It 



would be difficult, and in any case premature, to indicate the 

 cause of this phenomenon, but it may fairly be characterised as 

 the result of a retinal oscillation set up under the influence of 

 the beginning of the luminous excitation." 



The present paper deals partly with the colours of recurrent 

 images under different conditions, and partly with the reaction 

 attending the early stages of a luminous impression as noticed 

 by Charpentier. 



In the observation of the recurrent images set up by the action 

 of light of different colours, the author began, like Mr. Davis, 

 by using coloured glasses. 



A metal disk, about 8 cm. in diameter, was arranged so as 

 to rotate slowly and steadily about its centre in front of the con- 

 denser of a projection lantern. Near the edge of the disk was 

 a circular aperture about 0-5 cm. in diameter, the image of 

 which was focussed upon a distant screen. A plate of coloured 

 glass w.as placed before the projecting lens, and thus was 

 obtained a small coloured disk of light, which described a 

 circular path upon the screen. The coloured disk was, in 

 most cases, seen to be followed at an interval of a few de- 

 grees by a ghost oft he same size and shape, but of feebler lumi- 

 nosity, and of a hue which varied more or less with the colour of 

 the glass employed. With white electric light the colour of the 

 ghost was violet. 



This method of experimenting was, however, found to be un- 

 suited for the purpose in view, and it is mentioned only on 

 account of the facility which it affords for exhibiting the 

 phenomenon to a large number of persons. To obtain results 

 of any value, it was necessary to employ the simple colours of 

 the spectrum. 



In the arrangement finally adopted, light from a selected 

 portion of a spectrum was projected upon a small mirror, to 

 the back of which was attached a horizontal arm, not quite 

 perpendicular to the mirror : the arm was rotated by clockwork, 

 and the reflected beam of light was received upon a white 

 screen, fornung a coloured disk about I '5 cm. in diameter, 

 which revolved in a circular path having a diameter of 30 cm. 



When the mirror turned once in I J seconds the ghost or re- 

 current image appeared about 50° behind the coloured disk, the 

 corresponding time interval being one-fifth of a second. The 

 ghost appeared to be circular in form, its diameter being gener- 

 ally rather less than that of the ori-ginal. The colours of the re- 

 current images, as specified below, have all been observed by 

 several persons, and, except as to those at the extreme limits of 

 visibility, all the observations were in agreement. 



Spcclnimcolou 

 lixlreme violet.. 

 Middle violet .. 



Daik blue 



Light blue 



Middle green . . 



Greenish-yellow 

 Urangeycllow .. 



Orange 



Orange- red 

 Red 



Recurrent colours. 



No perceptible image. 



A pale image, variously described as 

 grey, yellow, and greenish-yellow. 



Feeble violet. 



Brighter violet. 



Bright violet. The image is more con- 

 spicuous with green light than with 

 any other. 



Blue. 



Bluish-green. 



D.iik bluish-green. 



Very dark bluish-green. 



No image at all, however bright the red 

 was made. 



A complete small spectrum, revolving parallel to itself in a 

 circle about i metre in diameter, was followed by a ghost 

 which corresponded to the portion of the spectrum comprised 

 between the orange and the beginning of the violet. The 

 -vhoUoi this ghost was of a violet hue; no trace whatever of 

 yellow or greenish-yellow could be detected at the more 

 refrangible end, nor of blue or bluish-green at the other. 



From other experiments it appeared probable that the blue 

 and bluish-green recurrent colours apparently observed when 

 the yellow and orange portions of the spectrum are tested 

 separately are due merely to an effect of mental judgment, and 

 not to any cause of a physiological nature. 



Four independent facts are consistent with the conclusion 

 that luminous recurrent images are due to a reaction of the 

 violet nerve fibres only. 



(a) With white light the recurrent colour is violet. 

 {b) In the recurrent image of the complete spectrum no colour 

 but violet can be detected. 



(c) A pure red light, however intense, gives no recurrent 

 image. {It is generally supposed by the supporters of the 

 Voung-Helmholiz theory that red light has no action upon the 

 violet nerve-fibres.) 



{d) The apparently blue colour of the ghost of simple 

 spectrum yellow is just as well produced by a compound yellow 

 consisting of green and red, the latter of which is inert when 

 tested separately. 



The path of the revolving spot of light is generally marked 

 by a phosphorescent track, which, when the rate of revolution 

 is not less than one turn in \\ seconds, often forms a complete 

 circle. The trail is due to the usually feeble continuation of the 

 after-image, of which the bright initial stage constitutes the 

 recurrent image. 



In the experiment next to be described, the Charpentier effect 

 and the recurrent image are made to exhibit themselves 

 simultaneously. 



Two blackened zinc disks, 15 cm. in diameter, from each of 

 which two opposite quadrants were cut out, were mounted in 

 contact with each other on a horizontal axis, driven by clock- 

 work, and making one turn in \\ seconds. By slipping the 

 disks over one another round their centres, opposite open 

 sectors might be obtained, of any aperture from 0° to 90'. 

 The apparatus was set up opposite a box containing a 32-candIe 

 power incandescent lamp, with a variable resistance in the 

 circuit, the side of the box between the lamp and the disks 

 being covered with a sheet of ground glass. 



The sectors being in the first place opened as widely as 

 possible, Charpentier's dark band was easily seen upon the 

 illuminated background. 



The sectors were then gradually closed up, until the posterior 

 edge of the dark band approximately coincided with that of the 

 sector. When this was accomplished, it was found that the arc 

 of the open sector was equal to about ,.'.; part of the whole cir- 

 cumference. The dark reaction, therefore, ceased in (5V of '4 

 seconds =) ^V second after the first impact of the light upon 

 the eye. 



For more readily demonstrating the succeeding phenomena, 

 it was found convenient to again open the sectors a little, so 

 that they covered an angle of about lo' or 12'. Resuming 

 the observation, it was seen that the posterior edge of the open 

 sector was bordered by a luminous fringe due to persistence. .\ 

 little beyond the termination of the fringe there appeared an 

 intensely black radial band, estimated to cover a space of from 

 3° to 4°, and distinguishable even upon the black ground of the 

 metal disk, though it is shown far more conspicuously upon a 

 translucent disk made of stout writing-paper with a sector cut 

 out. Lastly, after another interval of perhaps 35"' or 40^ came 

 the luminous recurrent image, which, with the yellowish light 

 of the incandescent lamp, appeared to be of a blue colour. 



This method of observation revealed one other point of in- 

 terest, which seems hitherto to have esca[4ed no'ice, though it 

 is evident enough with a Charpentier disk, when once attention 

 has been directed to it. The average illumination of the bright 

 band intervening between the dark band and the leading edge 

 of the sector is much more intense than thai of the other portion 

 of the sector. Moreover, it is not uniform, but increases, 

 gradually at first, and very rapidly at last, from the leading 

 edge up to the daik band. In fact when the light used is not 

 strong, the luminous margin of the bright band is a far more 

 conspicuous object than the dark band itself : it appears to 

 glow almost like a white-hot wire. 



NO. T 397, VOL. 50] 



