"Let there be in a given plane two reflecting points very near each other, and let the plane be so situated that the reflected image of a luminous object seen in it may appear to coincide with the points; then it is obvious that the length of the incident and reflected ray, taken together, is equal with respect to both points, considering them as capable of reflecting in all directions. Let one of the points be now depressed below the given plane; then the whole path of the light reflected from it will be lengthened by a line which is to the depression of the point as twice the cosine of incidence to the radius. "If, therefore, equal undulations of given dimensions be reflected from two points, situated near enough to appear to the eye but as one, whenever this line is equal to half the breadth of a whole undulation the reflection from the depressed point will so interfere with the reflection from the fixed point that the progressive motion of the one will coincide with the retrograde motion of the other, and they will both be destroyed; but when this line is equal to the whole breadth of an undulation, the effect will be doubled, and when to a breadth and a half, again destroyed; and thus for a considerable number of alternations, and if the reflected undulations be of a different kind, they will be variously affected, according to their proportions to the various length of the line which is the difference between the lengths of their two paths, and which may be denominated the interval of a retardation. "In order that the effect may be the more perceptible, a number of pairs of points must be united into two parallel lines; and if several such pairs of lines be placed near each other, they will facilitate the observation. If one of the lines be made to revolve round the other as an axis, the depression below the given plane will be as the sine of the inclination; and while the eye and the luminous object remain fixed the difference of the length of the paths will vary as this sine. "The best subjects for the experiment are Mr. Coventry's exquisite micrometers; such of them as consist of parallel lines drawn on glass, at a distance of one-five-hundredth of an inch, are the most convenient. Each of these lines appears under a microscope to consist of two or more finer lines, exactly parallel, and at a distance of somewhat more than a twentieth more than the adjacent lines. I placed one of these so as to reflect the sun's light at an angle of forty-five degrees, and fixed it in such a manner that while it revolved round one of the lines as an axis, I could measure its angular motion; I found that the longest red color occurred at the inclination 10 1/4 degrees, 20 3/4 degrees, 32 degrees, and 45 degrees; of which the sines are as the numbers 1, 2, 3, and 4. At all other angles also, when the sun's light was reflected from the surface, the color vanished with the inclination, and was equal at equal inclinations on either side.