all objects or images in the fields of view seem to move while the visual lines seem stationary. As the visual line of the right eye moves to the left all the images in its field seem to move to the right, and as the visual line of the left eye moves to the right all its images move to the left. Thus as the visual lines of the two eyes pass successively over objects in the external world right and left of the median line, the images of these objects move forward, and successively take position directly in front; so that under all circumstances the two visual lines combine to form a median visual line passing through the point of sight and onward to infinity. In case the images thus brought together are similar, as in my previous experiments, they combine, and are seen single at the point of sight; but otherwise they maintain their relative distances. This is especially true when we look at continuous lines on a plane parallel to the visual plane, as in the experiments detailed above. The fields rotate toward each other and images come together on the median visual line, but maintain their relative distances from the observer.

I might multiply experiments of this kind without number, but I wish now to show into what singular confusion some of the most careful observers have fallen by adopting the usual mode of representation. I will give first an experiment of Claparède which, according to him, is only a confirmation of what had been previously observed in the last century by Robert Smith, again in 1818 by Vieth, and again in 1838 by Wheatstone. It is therefore well fortified by great names. I translate:*"Take a compass widely opened (fig. 24), apply its head to the root of the nose, holding the branches in the plane of vision. The screen S is so disposed that the left eye L sees only the branch a, and the right eye R only the branch b. If now we press the branches until the points are brought on the optic axes, (the point of sight being at A) as indicated in the figure (fig. 24); we will perceive a closed compass between the images ss s's' of the screen (fig. 25), and this compass stretches itself even to the point of view A." Now I have repeatedly tried this experiment and I find the visual result entirely differ

ent.

This result is shown by my mode of representation in fig. 26. It is two half compasses united at the point and separated at the head, by the whole common field of view. It cannot be otherwise, since we must see two noses bounding the common field of view on each side, and the head of the compass resting on the nose must be similarly doubled and the images similarly separated. As to length, the compass may be imagined of any extent short of the point of sight A. If the compass is placed exactly in the plane of vision, the experiment succeeds equally Bib. Univ. Archives des Scien., II, Tome 3, p. 254.

well without the screen, for in that case only the right branch can be seen by the right eye, (the left branch being covered by it), and the left branch by the left eye. If the compass be placed a little below the plane of vision, so that both branches may be seen by each eye, or better if it be laid on the plane before described as in fig. 27, then the visual result will be represented by fig. 28.

The error in the interpretation of this very simple experiment, is due to the difficulty which most persons, even good observers, experience in separating what they know from what they see. In this experiment we know that the nose is in the middle between the eyes; we know also that the head of the compass is on the nose; and we see that the two points of the compass are together, therefore we imagine that we see a closed compass stretching from the nose to this point. But this is not only different from what we really see when our visual impressions are properly analysed, but is contrary to the first principles of binocular vision. The head of the compass impresses the extreme temporal limits of the retina of the two eyes, and must therefore not only be seen double, but the double images are separated by the whole common field of view. It seems almost incredible that images so widely separated should be imagined single.

Again: in Guy's Hospital reports is found an elaborate series of papers on the "Physiology of Vision," by Joseph Townes, extending through 8 vols., viz: 3d ser., viii and xv inclusive. Mr. Townes' experiments are performed with great care, many of them with the apparatus, and in the manner, already described. He has also in most cases truly described and truly drawn his visual impressions, as far as he has distinctly perceived them. But on account of the faulty mode of representation he seems to have neglected one half of the visual impressions which he ought to have perceived. His papers also show misconception on some fundamental points in binocular vision which I cannot now notice. I wish now only to compare his visual results with my own, and thus to show the superiority of my mode of representation. Among his many figures I will select only two, as the principles involved are precisely the same in all.

In the accompanying figure (fig. 29) I reproduce his drawing fig. 1. Pl. I, only changing the lettering so as to bring it into accord with the previous figures of this paper. The visual result which he obtains when this drawing is viewed on the plane already described, the point of sight being at A, is given in fig. 30. In this result V is the combined visual lines vr and vl, b is the median line B of the drawing, as seen by the right eye, and b' the same line as seen by the left eye, but the right eye image of vl and the left eye image of vr are entirely overlooked.

30.

31.

33.

The result which I obtain is given, represented by my method, in fig. 31.

In fig. 32, I reproduce one of his more complex drawings, viz: fig. 1 of his Pl. II, lettering of course as before. The visual result which he obtains by viewing this figure with the point of sight at A is given in fig. 33, where An is the combined lines AR and A L of the drawing, bede. The corresponding lines of the drawing, as seen by the right eye alone, and f'g'h'i corresponding lines of the drawing as seen by the left eye alone. My result is given in fig. 34. By careful comparison of these two results, remembering that capitals represent combined images, small italics right-eye images, and dashed italics left-eye images, the nature of their differences will be readily understood. Mr. Townes in his result makes the right eve shift to the left only the right portion, and the left shift to the right only the left portion, of the drawing. My result shows the whole drawing shifted by the right eye to the left and the left eye to the right. Mr. Townes has neglected entirely the left eye images of b c d e, and the right eye images of fghi. Every one of his visual results of figures drawn on the plane described, except a few in which he used a median septum, are faulty in the same way, and may be reconstructed on the same principle, with the utmost certainty, even before performing the experiment. Oakland, Cal., Sept. 27, 1870.

ART. IX.-Brief Contributions to Zoology from the Museum of Yale College. No. IX.-Notice of a Fossil Insect from the Carboniferous formation of Indiana; by SIDNEY I. SMITH.

Paolia vetusta, gen. et sp. nov.

THE fossil insect wing, represented in the accompanying figure, was sent for examination and description, from the Museum of Hanover College, Indiana, by Prof. E. Thompson Nelson. It occurred in the "grit" quarry near Paoli, Orange county, Indiana, and was discovered by the workmen while sawing the stone into small pieces for whetstones. The fossil is remarkably perfect, nearly the entire wing being beautifully preserved, and of peculiar interest for showing the complete wing-skeleton of an insect so ancient as the coal period. It is apparently the anterior wing of the left side, and measures 254 in. in length and about 85 in. in breadth, the posterior margin being slightly cut away by the saw.

Along the anterior border, the space between the second and third nervures is bent abruptly downward in the specimen, so as to be somewhat foreshortened as seen in the figure. This

is probably a natural fold such as is found in many Neuropterous insects.

The neuration is remarkable for the number of slender branchlets which the nervures throw off toward the posterior border and the tip of the wing. The first, or marginal, nervure* is continuous to the tip of the wing. The second is simple and straight nearly to the extremity, where it throws off several very minute branchlets and itself impinges upon the margin, about three-fourths of an inch from the tip of the wing, as a very delicate nervure. The third divides a short distance from its origin; the anterior branch runs parallel to the second nervure, is separated from it nearly as far as the second is from the first, and is simple to near its extremity, where it is twice forked, the minute branchlets reaching the margin a little way from the tip of the wing; the lower branch is broken and wanting just at its origin; it forks about the middle of its course, the anterior of these secondary branches again subdividing like the anterior primary branch, while the posterior secondary branch divides. about the middle of its course, the anterior of these tertiary branches again dividing near its base, and the anterior of the branchlets reaching the margin at the apex of the wing. The fourth nervure divides a third of the way from the base to the tip of the wing; both these branches fork or branch several times, so that there are twelve branchlets extending from them to the margin, and one from the posterior, or inner, side which is lost in the space between the fourth and fifth nervures and does not reach the margin. The fifth divides near the base of the wing; the inner branch runs an almost straight course to near the margin where it divides; the outer primary branch divides about the middle of its course, both its branches again dividing or branching several times, and the first branchlet on the outer side not reaching the margin. The basal portion of Figure 1, is once and a half natural size lineally, and was drawn on wood from a photograph.

*For convenience of comparison, I follow Heer's notation of the main nervures of the Neuropterous wing, as adopted by Mr. Scudder, regarding what are usually considered the primary branches of the median as distinct nervures, making six in all.