let the line A be a string of six beads, each of which is a minimum visibile, or smallest point from which a cone of rays can come. Now, the ordinary optical doctrine, as we understand it, is, that this string of beads A falls upon the retina in an image in the form of the row of figures B; that is to say, in an image in which the bead 1 is thrown with its head downwards on the retina, and all the other beads in the same way with their heads downwards. Now, on the contrary, it appears to us demonstrable, that the beads A must fall upon the retina in an image in the form of the row of figures C; that is to say, in an image in which each particular bead or minimum lies with its head upwards upon the retina. In the annexed scheme our meaning, and the difference between the two views, are made perfectly plain; and it is evident, that if the object were reduced to only one minimum—the bead 2, for instance—there would be no inversion, but a perfectly erect image of it thrown upon the retina.
Now, there are just five different ways in which the fact we have now stated may be viewed. It is either a fact notoriously announced in all or in most optical works; and if it is so, we are surprised (though our reading has not been very extensive in that way) that we should never have come across it. Or else it is a fact so familiar to all optical writers, and so obvious and commonplace in itself, that they never have thought it necessary or worth their while to announce it. But if this be the case,
