the pencil into such a position that its double images shall exactly coincide with the centers of the two objects which you desire to combine. If you now look at the pencil, the ocular convergence will be exactly suitable for combining the objects.

In the cases thus far mentioned there is no illusion : the combined images do not produce the appearance of a real object, as in the case of combined images of the same object producing single vision; because, in the first place, the two objects are dissimilar, and therefore the combination is not perfect; and, in the second place, the illusion is destroyed by the existence of the two other uncombined images. We next try

2. Similar Objects.-If the two objects, the images of which we desire to combine, are exactly similar, then the combined image will be exactly like a natural object. For example:

Experiment 1.--Place two pieces of money of the same kind on the table, being careful that the stamped figures shall be in the same position. Now, looking down upon them, combine as before. Not only will the outlines of the two pieces combine, but the stamped figures in the minutest details, so that the middle combined binocular image will have all the appearance of a real object. This is illustrated by Figs. 37 and 38, in which the position of parts is reversed, because the eyes are supposed to be looking down. In Fig. 37 the two objects (coins), A and B, are combined by crossing the eyes, and the combined or binocular opaque image will be seen at the point of sight as a'b, while monocular shadowy images, a and b', will be seen right and left. In Fig. 38 the combination is made by looking beyond the plane of the coins, and the coins in this case must not be more than an interocular space apart. The com

bined images, like a real opaque object, will be seen at the point of sight ab', and the two shadowy monocular images right and left, as before, only they are now het eronymous.

In this case, though the combination is perfect, yet the illusion is still not complete, because of the presence of the accompanying monocular images; but the formation of these may be prevented by the use of appropriate screens.

Experiment 2.—If in the first experiment with the money, before combining, we hold two cards, sc, sc', Fig. 39, one in either hand and at about half the distance to the table (the best distance is the plane of combination or plane of sight, for then there will be no doubling of the cards), in such position that the card in the right hand, sc, will hide the right piece A from the right eye but not from the left, and the card in the

left hand, se', will hide the left piece B from the left eye but not from the right, and then make the combination by crossing the eyes, the combined binocular opaque image will be formed as before; but the monocular images will not appear, because there will be

no other retinal image formed except on the central spots. This is represented in Fig. 39. In case we combine beyond the plane of the objects, then a median screen, sc, Fig. 40, extending from the root of the nosen to the table, midway between the objects, will prevent the formation of the monocular images, as shown.

But in these cases, although the union of the two images is perfect, and although we see nothing but an apparently solid opaque object, even yet the illusion is not absolute; partly because the table is doubled and therefore unreal, and partly because the eye is adjusted to the point of sight, whereas the light comes from the

object, which is either nearer as in Fig. 40, or farther off as in Fig. 39, than that point. We will try therefore still another case.

3. Many Similar Objects regularly arranged.-The illusion is most complete when we combine the images of many similar objects regularly arranged over the whole field of view, such as the regular figures of a tessellated pavement or oilcloth, or of a regularly figured carpet of small pattern, or of a papered wall of regular pattern, or the diamond-shaped spaces of a wire grating. In such a case, when by convergence we combine two contiguous figures immediately in front, other contiguous figures all over the plane also combine. In other words, by the motion of the eyes in opposite directions in convergence, the images of the whole plane of the figured surface are slidden by one eye to the left and by the other eye to the right, until combination takes place again over the whole field. When the combination is effected, if we hold the point of sight steady, the combined images of the figures, at first a little blurred, become sharp and clear; and then the whole figured plane comes forward to the point of sight, and appears there as distinctly as a real object, but on a smaller scale in proportion to the less distance. This is represented in Fig. 41, in which the strong line PP represents the plane of the regular figures 1, 2, 3, 4, 5, etc. When contiguous figures, 6 and 7, are united by convergence at the point of sight, and seen there, then all other contiguous figures, 1 and 2, 2 and 3, etc., all over the plane, will be similarly united, and the whole plane with all its figures will advance and be distinctly seen at the distance p' p'. When by stronger convergence alternate figures, 5 and 7, are combined at a nearer point of sight 5 on the plane p" p"-or (which is the

same) when we use the plane p' p' first obtained with all its figures as a real object, and again combine contiguous figures—the whole plane advances to p" p", and is seen as a distinct object with a still smaller pattern of figures. Using the plane thus obtained again as an object, and uniting its contiguous figures, the whole

plane again advances still nearer, and the figures become still smaller at p'" p'". In this manner I have often distinctly seen a regularly figured wall or pavement on six or seven different planes coming nearer and nearer, and becoming smaller and smaller, until the nearest was within 3 inches of the eyes, and the figures