ON ACCIDENTAL OR SUBJECTIVE COLORS. 233 



scure portion ; in other words, the crescent appears to form part of a disk sen- 

 sibly greater than that to which the rest of the planet belongs. If a decisive 

 experiment, capable of being repeated at all times, be desired, it may be made 

 in the following manner : On a rectangular blackened pasteboard. 20 centime- 

 tres in height and 15 in breadth, let two white rectangles or windows, separated 

 by a black band of a demi-cenlimetre in breadth, be left blank on the upper 

 half; let this band, prolonged in blank, divide into two black portions the lower 

 half of the pasteboard. Thus there are two bands of equal width, but the one 

 white on a black ground, the other black on a wliite ground. To make use of 

 this apparatus, it is to be placed vertically near a window, so as to be Avell 

 lighted, and observed from a distance of four or five metres. The white band 

 will now appear considerably wider than the upper and black band, and this 

 apparent difference increases with the distance. It will be perceived that the 

 apparatus is constructed in such a manner as to augment the effect resulting 

 from irradiation ; for if, on the one hand, the white band acquires apparently 

 additional width, the black band, on the other, is diminished by the irradiations 

 of the two lateral white spaces. Irradiation, therefore, with the naked eye, 

 must be regarded as one of the best established facts of vision, and one of the 

 most easily verified. Its intensity is not the same in different eyes, and even 

 varies in the eame individual, but there is no one who cannot observe it in a 

 manner more or less distinct. In telescopes, it becomes complicated ; its effect 

 is diminished ; the total error depends at once on the enlargement in itself, on 

 the lustre of the image, the nature as well as state of the eye, and the perfec- 

 tion of the instrument — qualities essentially variable. If, when micrometers of 

 double image were employed to measure the irradiation, it has not been practic- 

 able to find its sensible value ; if it has even occurred that the double image 

 micrometer has caused an irradiation to disappear which had been realized by 

 the thread micrometer, this depends on a cause revealed by experiments which 

 have rigorously demonstrated the following general principle : two irradiations 

 in presence of and sufficiently near one another, both undergo a diminution. 

 This diminution is so much the more considerable as the borders of the lumin- 

 ous spaces from which the two irradiations issue approach one another more 

 closely. 



On the cause of ocular irradiation. — The theory of irradiation now univer- 

 sally adopted consists in the admission that the impression produced on the eye 

 by a luminous object is propagated on the retina to a small distance around the 

 space directly excited by the light, so that the total sensation corresponds to an 

 image somewhat greater than the true one. The principle of continuity on which 

 this theory is founded is so simple, that, if the existence of irradiation were not 

 known, it would seem that it might be foreseen a priori. Let us suppose, in 

 efi'ect, that a luminous or illuminated body projected on a ground completely 

 black, is attentively looked at. The light emanating from this object will strike 

 a definite portion of the retina, and the rest of the organ will not be subjected 

 to any direct excitation. But is it presumable that the parts of the retina which 

 immediately environ the portioii directly excited should be in perfect repose ? 

 It will scarcely be admitted that a state of energetic excitement and a state of 

 absolute repose should be thus, in the same organ, in immediate contact. We 

 are hence led a priori to think that, around the image of the object, must be 

 manifested some appearance which constitutes the gradual transition between 

 the state of excitation of the part of the organ subjected to the direct action of 

 the light, and the state of repose of the parts more remote. Now, in whatever 

 manner this transition is effected, it must be considered as altogether probable 

 that the excitation is propagated to a greater or less distance, without a change 

 of nature, around the space struck by the light, and that hence there will result 

 the sensation of a larger image. Irradiation will be thus, in relation to space, 

 what the known phenomenon of the persistence of impressions of the retma is 



