itself is doubled and becomes two lines which, resting upon the nose on each side of the common field of view, run out parallel to each other. Between these two lines the combined eyes look out along the combined visual lines, at the distant object. If the position of the median line of sight be occupied by a real line or rod we shall see two parallel lines or rods; if it be occupied by a plane as in the stereoscope, we shall see two parallel planes bounding the view on either side, between which we look.

Suppose now we take a slender rod AB n (fig. 8) like a straight wire, and placing one end on the root of the nose and the finger on the farther end, so that it occupies the position of the median line of sight, we look at a distant horizon. As already stated we shall see two parallel rods between which we look at the distant point, (fig. 9). If now we look at a nearer point the rods will converge, carrying all objects in the fields of view with them, and meet at the point of sight. If for example we look at the finger A, the rods will form a triangle of which the finger is the point and the distance between the two noses the base (fig. 10); if we look at a still nearer point B, the rods, still turning upon the noses as a fixed point, will cross each other at the point of sight, (fig. 11). At the same time the two eyes with their visual lines combine to form a median eye and a true median line of sight (the dotted line in the figures), passing through the point of sight and beyond to infinite distance, and this line might well be called by this name if the name had not been already appropriated for another line.

It will be observed that in the above paragraph I have used very much the same language as other writers on this subject, and used a similar mode of representation; only that I have substituted the eyes in the place of the nose and put noses in the position of the eyes. I have made median lines cross each other at the point of sight instead of visual lines, and visual lines combine in the middle as a true visual median line. In other words, I have used the true language of binocular vision.-I have expressed what we see rather than what we know, the language of simple appearance rather than that mixture of appearance and reality which form the usual language of writers on this subject.

In all the experiments which follow it will be necessary to get the interocular distance with accuracy. This may be very conveniently done by taking an ordinary pair of dividers, holding it up at arms length against the sky or a cloud, and while gazing steadily at the sky or cloud, opening the points until the double images of the points shall two of them coincide perfectly, thus: The distance between the points bc (equal to a a')

is the interocular distance.

1

I will now give a few very simple experiments and the visual results, represented in the manner already indicated. Let a smooth white plane about twenty inches long and of any convenient width, be prepared by pasting white paper on a smooth flat board, (fig. 12). Mark two points R and S with a space between them exactly equal to the interocular space. In the middle between the two points notch the end of the board so as to fit over the bridge of the nose. From the notch n draw the line nA in the position of the median line of sight when the notch is fitted on the nose, and on this line place pins at A and B. If now, the plane be fitted on the nose and held in a horizontal position a little below the visual plane, so that both the pins and the line can be distinctly seen, and the eyes be directed upon a distant object; then, as already stated the median line will be seen as two parallel lines separated by a space equal to the interocular space. If in addition to the median line the position of the visual lines be represented by dotted lines, then fig. 13 will represent the actual relation of parts, and fig. 14 the visual result. It will be observed by comparing the visual result with the drawing, that the parts of the face, the whole plane and all the objects and lines on the plane are shifted by the right eye one half the interocular distance to the left, and by the left eye the same distance to the right; the two visual lines of the drawing are combined in the result to form a true median binocular visual line EV, while the median line of the drawing is doubled, forming heteronymous images, which in the result occupy the position of the visual lines of the drawing. Besides these three lines, there are seen two other dotted lines lv and r'v'. These are not visual lines proper, i. e. lines along which vision takes place, but they are the visible representatives of the visual lines of each eye seen as objects by the other eye.

Next, having removed the dotted lines in the previous experiment, let the eyes be directed upon the pin A, and dotted lines be again drawn in the direction of the visual lines as in fig. 15. The visual result is shown in fig. 16. By comparing this visual result with the actual drawing it will be seen that the whole triangle L A R, is turned upon A as a fixed point a half interocular distance to the left by the right eye, and to the right by the left eye.

Next, having again erased the dotted lines of the previous experiment, and substituted dotted lines crossing at B as in fig 17: let the plane be again adjusted to the nose and the eyes directed upon the pin B. the pin B. The visual result is shown in fig. 18. In this visual result the whole drawing (fig. 17) is apparently rotated about the point of sight B in opposite directions by the two eyes, through a distance represented by the angle "BŘ or nBL half the visual angle. As objects on the nearer and the farther side of the point B must move in opposite directions,

the double images beyond B must be homonymous. In this experiment (as also in a less degree in the last), since the eyes are converged, the distance between the points L and R on the plane, must be a little less than the interocular distance.

In both the above cases the visual result is the drawing rotated about the point of sight in opposite directions through one-half the visual angle; and for convenience I have spoken of the eyes as accomplishing this rotation. But this is not strictly true. The fields of view of both eyes are habitually and normally shifted one-half interocular space right and left heterony. mously, when the eye is in a passive state. Under these conditions the visual results of figs. 15 and 17 would be figs. 19 and 20. But in converging the eyes upon A or upon B, the median lines n b a, n'b'a' turning upon n and n' as fixed points approach each other, meet and cross at the point of sight so as to form the result already given in figs. 16 and 18.

not

Lastly-if without using the plane, three objects A, B and C, be placed in the median line of sight, and the eyes be fixed upon the middle one B (fig. 21), then by the usual mode of representation the position of the double images of A and C will be a'a, cc'; the lines here representing of course visible lines but only lines of direction. But if using the plane we draw a visible line in the direction of the median line, and upon it place three pins at A, B and C, (fig. 22), and connect these points with R and L, also by visible lines, and then direct the eyes upon B, the falseness of the former mode of representation becomes at once evident. The visual result of this experiment is shown in fig. 23. Every line of this somewhat complex figure can be distinctly seen. By comparing the two modes of representation (figs. 21 and 23), it will be seen that the parallactic position of the double images, or their position when referred to the plane of B, is the same (the dots represent these positions in fig. 23), but only the second mode (fig. 23) represents truly their apparent distance.

Thus there are two apparent movements of the visual fields accomplished by the eye (or the mind) in binocular vision: 1st, in a passive state, a shifting of each field one-half interocular space to the opposite side, so as to bring the two visual lines together to form a true median visual line; this movement is involuntary and habitual: 2d, in ocular convergence a rotation of each field in a direction opposite to the motion of the eye, on a line passing through the eye normal to the visual plane; this takes place in every voluntary act of sight. In all other movements of the eyes, as when we look from side to side or upward or downward, objects seem stationary, and we are conscious of the movement of the visual lines over them; but when the optic axes move toward each other as in convergence,