212 ON ACCIDENTAL OE SUBJECTIVE COLOES. 



cease, but persists during a very sensible interval of time. Let us pause for 

 a few instants to consider with M. Plateau this capital phenomenon, on which 

 depend, in part at least, the subjective colors. 



Persistence of images. — Everyone knows .that when a burning coal is rapidly 

 whirled in the dark, a luminous curve will be seen, as if the coal left on the eye 

 the trace of its passage. This appearance evidently results from the fact that 

 the impression produced by the object at any point of its course still subsists for 

 some time after it has left that point, so that the successive positions of the 

 luminous object must appear simultaneous. To the same cause are to be referred 

 numerous appearances which exhibit themselves whenever we observe an ani- 

 mated object in rapid motion. To this fireworks owe a part of their brilliant 

 effect; rain and hail, for this reason, instead of presenting in their fall the ap- 

 pearance of rounded bodies, offer that of parallel lines; the teeth of a wheel 

 which turns rapidly disappear and seem to be replaced by a semitransparent 

 gauze, through which objects maybe distinguished; a vibrating cord presents an 

 analogous effect, &c. We may also directly observe the phenomenon of the 

 persistence of impressions if, after having looked at a luminous object of suffi- 

 cient lustre, a window for instance, we suddenly close and cover the eyes, when 

 the image of tlie object will be seen to continue for some time. 



Measure of the chiration of impressions. — Although this persistence of impres- 

 sions is constantly reproduced and under a multitude of circumstances, and has 

 received many useful and curious applications, it has been very little studied. 

 D'Arcy first conceived the idea of measuring the duration of the phenomenon 

 by causing a burning coal to revolve in darkness so as to produce ihe appear- 

 ance of a luminous ring. And if, in effect, we succeed in giving to the object 

 such a velocity that it shall repass at each 'point of its course, precisely at the 

 instant when the impression produced by its passage is at the point of vanishing, 

 the duration of the impression will be measured by that of a revolution. D'Arcy 

 concluded from his experiments that the duration of the impression produced on 

 his eye by a burning coal was equal to 8'" or 0".13. But this method is de- 

 fective; it gives not the complete duration of the impression, but only the time 

 during which this impression is maintained without sensible loss. To obtain 

 with exactness the total duration of the impression it would be necessary that 

 the velocity of the object should be such that it shall find at each point in its 

 course the preceding impression so much weakened as to be just at the point of 

 extinction ; which is almost impossible. M. Plateau has, however, succeeded in 

 surmounting, in part, these difficulties, and has determined the total duration 

 of the impressions produced on his eye by white, yellow, red, or blue objects, 

 observed by daylight. The values for the different colors employed are jDer- 

 ceptibly equal: their mean is 0". 34, instead of 0" 13 found by D'Arcy; still 

 this result is but a first approximation. 



Liength of time necessary for the production of the impression. — Experiments 

 of this kind plainly establish the existence of another fact too little remarked: 

 namely, that for the complete formation of an impression on the retina some lapse 

 of time is requisite. This might have been inferred from the well-known fact 

 that an object which passes very rapidly before the eye is scarcely distinguishable, 

 or even is not perceived at all. The following experiment places this beyond 

 doubt : Let a small white object, a piece of paper for instance, be made to move 

 circularly before a dark back-ground with such a velocity that the apparent ring 

 shall present a tint perfectly uniform and constant ; this ring will not appear white, 

 but gray. Now, it is apparent from the uniformity of the tint that, during the 

 duration of a revolution, the impression produced is maintained without sensible 

 loss ; if this impression were white, it is evident that the entire ring ought to 

 appear decideflly white, not gray; it follows that by reason of the small space 

 of time during which the object acts on each point of the retina, it produces only 

 a grayish impression — that is to say, an impression imperfectly white. No 



