smaller, until when the point of sight is at the screen, fig. 4, then the images will completely unite, and the screen become entirely opaque. This is shown in the visual result, fig. 5.

If next we use two fingers, one of each hand, and gaze again at the wall, we will see four images all transparent. Now approximate or separate the two fingers until the two middle images unite; we will have three images, the middle one opaque, the other two transparent. The reason is obvious. The middle one is opaque because a portion of the wall is concealed by it from both eyes. This portion of the wall is concealed from the right eye by the right finger, and from the left eye by the left finger; but it is the right-eye image of the right finger and the left-eye image of the left finger which unite to form the middle opaque image, while the right-eye image of the left finger is seen to the left, and the left-eye image of the right finger to the right, both transparent. In binocular vision, superposed images of opaque objects are always opaque, while single images are always transparent.

The principles (1, 2, 3) laid down in the early part of this paper, together with the explanation of transparent double images just given, furnish, we believe, the key to all M. Pictet's experiments. We will make the application only to those which he thinks most conclusive of the existence of illusive images. We will first give his experiments and his conclusions as fairly as we can, and then will proceed to give our own expla nation. The following experiments M. Pictet thinks conclusive:

Place an opaque screen S (fig. 6) against the nose n in the median plane of sight, in such wise that the object A may be

seen by both eyes. Now place a second screen C across the visual line A R of the right eye, so as to intercept rays from

the object A to the right eye. Nevertheless, the object A will be seen apparently through the opaque screen C, which will therefore appear transparent, and may even be drawn in outline with accuracy on the screen at b exactly where the visual line of the right eye pierces the screen-exactly where, if the screen were transparent ground glass, we would see it with the right eye, and might trace its outline. M. Pictet thinks this absolutely inexplicable, except on the assumption that an illusive image is actually seen at A by the right eye; and that it is this that we draw in outline on the screen at b, the screen being transparent because the illusive image is seen beyond.

But M. Pictet gives another experiment which he thinks still more conclusive.

On

a sheet of paper lying on the table place a piece of money; then place a screen upright on the right side of the money, and let the face of the observer be brought down upon the screen, so that the latter being in the median line shall intercept the view of the right eye. Nevertheless, says M. Pictet, "on directing the regard upon the piece of money, we see that the vertical screen appears transparent throughout, and that it permits the right eye to distinguish the piece, as if through a very diaphanous sur face." "If now we give to the optic axes a direction more parallel, we see the image of illusion of the right eye move gradually toward the right, traverse the line of intersection of the screen and the table, and come to project itself on the other side upon the paper, where we may trace its outline correctly." To represent these facts more clearly I give the diagram, fig. 7. In this figure R and L are the two eyes, pp the sheet of paper, A the money in the direction of the visual line of the left eye, SS the median screen, AR the visual line of the right eye when we look at the piece of money A and seem to see it through the screen, Rb the visual line of

he right eye when the optic axes become parallel, and b, exactly here the visual line of the right eye pierces the paper, the place here the outline of the piece may be traced. The image oves to the right or left according to the position of the optic xes, being always where the visual line pierces the paper. ut "the most advantageous position of the optic axes," says I. Pictet, "is parallelism, for it is that which removes the farthest e image of illusion from the real image." I wonder that M. ictet did not reflect that, being on corresponding points, by is own principle the image of illusion, if any, cannot be separated om the real image; and that there is in fact but one image seen. But furthermore, if a convex lens be placed across the visual ne of the right eye Rb, the image at b will not be affected, but e tracing we make of the image will be found as much smaller an the money as the lens magnifies; showing that the image not magnified but the drawing is magnified, and therefore, M. ictet thinks, that the image of the money is illusive or subjective, hile the image of the paper and of the tracing is real. If, howver, the lens be placed before the left eye, the image is magnied because, thinks M. Pictet, this image is the illusive righte fac-simile of the magnified real image of the left eye.

T

One more step of M. Pictet's proof. By keeping both eyes en, objects in the microscopic field may actually be drawn ith accuracy on a sheet of paper placed on one side of the icroscopic tube. Or, still better, if a stereoscopic card, having picture on one half and the other half blank, be placed in the ereoscope, we may trace the picture on the blank half. Acording to M. Pictet's view, the light impresses one eye, and is impression is propagated as an illusive image to the other e, and thrown on the paper just where the visual line pierces it. Such are the most important experiments upon which M. ctet bases his belief in the existence of illusive images. I ve been familiar with all these phenomena for many years; have also often used the method of tracing microscopic objects commended by him; but my explanation is wholly different. M. Pictet's view be correct; and if in the experiments deled we actually trace the outline of an illusive image belonging the right eye, then where is the true image belonging to the t eye? We see but one image. M. Pictet, in accordance th his principles (although he forgets them in the passage oted above), ought to answer, that being on corresponding ints the two images are perfectly united. Then why call the age we outline an image of illusion? The truth is, in every se, we trace the outline of the true image seen by the left eye; though by the principles laid down in the early portion of s paper, or by the properties of corresponding points, we aw the outline at a different place from the object.

[To be concluded.]

ART. XL.-On the position and height of the elevated Plateau in which the Glacier of New England, in the Glacial era, had its origin; by JAMES D. DANA.

THE existence of a region of high elevation in norther America as the source of a continental glacier in the Glacial era has often been assumed, but rather because the glacier theory was lame without it than as a result of special research with respect to the extent and location of such a region. I d not propose, at the present time, to treat the subject continertally, but simply to discuss the question as to the origin of the New England glacier. If I mistake not, however, the discus sion, so far as it goes, meets the requirements of the continental question, while it proves that the idea of one central glacier source for the continent is without foundation.

1. Position of the elevated plateau.-In my paper of last month, on the "Valley Movements of Glaciers," (p. 233) I mentioned the fact, from the observations of the Vermont Geological Survey, that the glacial scratches on Camel's Hump, Mt. Mansfield and Jay's Peak, three of the high summits of the Green Mountains in the northern two-thirds of Vermont, have a course of about S. 50° E.* In the White Mountains, 70 miles to the south of east (S. 75° E.), the direction of the scratches near the Lake of the Clouds, on the north side of Mt. Monroe, is, as I am informed by Prof. C. H. Hitchcock, at points from 5000 to at least 5200 feet above the sea, in different places, S. 34° E. and S. 54° E. (S. 20° E. and S. 40° by compass); on the north side of Mt. Clinton, 4430 feet high, (one of the White Mountain series of summits about 17 miles west of Mt. Washington), near top, S. 50° E.-S. 54° E.; and on the south peak of Mount Clinton, at a level of 4320 feet, S. 54° E. He observes further that boulders occur at a height of 5800 feet or the north side of Mt. Washington.+

These scratches, and others crossing the Green Mountains, could not have been made with the land of North America at its present level. Elevated land must have existed to the north, if they were the work of glaciers. Moreover, these scratches

* According to the Vermont Geological Report the direction of the scratches on Camel's Hump is S. 50° E. (S. 40° E., compass), on Jay's Peak, S. 50° E., and on Mt. Mansfield, situated between the two, S. 30° E., S. 55° E., S. 58° E., at different points.

+ Prof. Hitchcock observed also the course S. 24° E. near the Lake of the Clouds, but it was not common; also, in the saddle between Mt. Pleasant and Mt. Franklin, 4400 feet above the sea level, S. 34° E.; and the same in the gap between Mt. I'leasant and Mt. Clinton, 4050 feet high. It would therefore appear that the course of the general glacier over the White Mountain region was about S. 55° E, and that there were variations from this course due to the lay of the land.

n the higher summits of New England, must point approxmately toward the region of elevated land, or the great Icy 'lateau, as we may call it, on whose slopes the movement of he glacier originated.

This seat of power in the era of ice was not the Adironacks; for these stand (as will be seen on any map of North America, and even a map of the World) much too far to the outh. The scratches point over Lake Champlain and the low ills and plains beyond, and across the St. Lawrence. On the ther side of this river lies the large valley of the Ottawa, and it almost exactly in their direction; and the prevailing trend of he scratches through its lower half is about S. 45° E. Accordng to the table in Logan's Report for 1863 (p. 890), this is the ourse at Ottawa City, Hull, Rideau River at Stegman's Rapids, nd Horton near Renfrew village south of the Ottawa. llumettes Lake, the course S. 25° E. was found; but this cality is higher up the stream, and the course may well have een due to some local influence. There can be no doubt at, if the Glacial era was a glacier era, the Ottawa valley ice as a part of the same great ice-stream with that which crossed te Green Mountains, but a portion nearer the source.

At

The Ottawa valley, including that of Montreal River (which is the course of the Ottawa and is its western head-tributary), tends to a point nearly 500 miles in an air-line from Mt. ansfield, and 570 from the White Mountains, with an average end of S. 65° E. But the source of Montreal River was not e source of the glacier movement; for the course of the ratches in the Ottawa, instead of corresponding with the end of the valley, is, as stated, S. 45° E. The scratches point, erefore, to the eastward of the western source of the Ottawa, d at least as far east as the region northeast of Lake Temismang (the main source of the waters of the Ottawa), on e watershed between the St. Lawrence valley and Hudson's The scratches of the White Mountains and Green Mounns and those of the lower part of the Ottawa valley point ke toward this area.

ɩy.

It follows that the Icy Plateau, whence the great glacier took departure, must have existed either about this part of the nadian watershed, or in the same direction farther to the northst; and since the line of the scratches, if carried farther to e northwest, would strike Hudson's Bay, or its western rder, it is probable that this watershed was the actual sition.

This view accords with the great diversity of direction in scratches about Temiscamang Lake, they varying, accordto Logan, from S. 78° E. to S. 7° W., the least easting ng found on the west side of the Lake; the observations are,