The observation also excited the curiosity of Henri Morgenroth, a consulting engineer of Santa Barbara, Calif. He wrote: "Munro's item has cost me a lot of sleep. Why should this varmint need gyroscopic eyeballs when other animals don't? Or, if others have them, which species do-and why? If Munro or your readers can shed any light on this puzzle, I wish they would speak up.

"If I remember Norbert Wiener's book Cybernetics correctly, he had a lot to say about the terrifyingly complicated apparatus that translates an image on the retina into the code of memory. Among the many transformations that signals from the retina have to undergo before they find and are matched with their twins in the memory (that is, for recognition to take place) is a step which, in effect, levels these signals regardless of the angle at which we hold our eyes.

"Since the copperhead needs gyro-controlled eyeballs, this species of snake must lack at least this one processing apparatus in the signal's path from the retina to the memory.

"It seems to me that this would represent a major design difference between the brain of the copperhead and that of animals with eyes which make no provision for perpendicularity control. Certainly you would not expect a difference of this magnitude among closely related species. Consequently other snakes must have the same eyeball control.

"Could it be that this characteristic has escaped observation because it is much harder to detect in the case of eyes with round pupils? If this suspicion makes any sense to you, why not ask one of our herpetological friends to attach or paint little markers on the pupils of other snake species and take a look?"

Munro reported that he had already examined snakes with elliptical pupils. "Without exception," he wrote, "the eye keeps its vertical position when the snake tilts its head. This observation, incidentally, casts some doubt on my original supposition that the eye accommodation evolved in pit vipers is an essential for accurate striking. Snakes that do not strike also have the gyro-control. Incidentally, you cannot attach markers to the eyeball of a snake for observing eye movements. The eyeball is protected by an immovable window, part of the snake's skin, which is cast when it sheds."

"I could not shake off the suspicion," Morgenroth went on, "that if snakes lack one or possibly more of the major complications in the picture-recognizing department of the brain, this deficiency and consequently the copperhead's 'crutch' of a gyro-controlled pupil must also exist in animals farther down the evolutionary ladder, including fish. It is just not conceivable that snakes should be distinguished from all other members of the animal kingdom by such an extraordinarily special brain design. Yet it is not easy to suppose that such a thing should have escaped notice.

"It should be said that today's snakes are a fairly recent product of evolution. Historical geologists tell us that during a long dry period some millions of years ago the snake's ancestors burrowed underground to escape extinction and that their eyes retrogressed to barely more than light-sensitive buds. When life on the surface became possible for them again, they faced the necessity of reinventing the eye. Hence in many details of design snakes' eyes are unique. Nevertheless it seems reasonable to suppose that the limited brain mechanism which requires a gyro-eye might be shared by some of the snake's evolutionary kinfolk.

"With the objective of locating one of these, I undertook a scientific expedition to the goldfish tank at a local five-and-ten-cent store. Quite a number of these little fish, I discovered, carry spots on the periphery of the pupil which are easy to watch as the fish gazes at you while swimming up and down the glass wall of the aquarium. After watching them for a while I saw that, sure enough, the fish eyeball always kept the same position relative to a vertical line regardless of whether the fish pointed its nose 60 degrees up or 60 degrees down.

"Then I invested 5O cents in a baby turtle, selecting one with two little black stripes on the tissue surrounding the pupil. I found that the turtle's eye moves in precisely the same way as the copperhead eye illustrated in your March issue! The maximum angle of turning down, as the head points up, appears to be about 70 degrees. As the head is turned down, a limit is reached at about 10 degrees. If the animal is tilted farther down, it compensates by raising its entire head. The eyes seem unable to follow a sudden tilting. The lids close for about a second, as if the unanalyzable image has to be excluded until the eyes level off.

"It seems just not possible that all this has never been discovered before. Was Dr. Munro's observation really a scoop? If it was, and if what I have just reported has not been known, then I hereby lay claim to being the first man who ever really looked a fish straight in the eye!"

Before conceding that distinction to Morgenroth, it seemed appropriate to consult a professional biologist and learn just who was scooping whom. Charles M. Bogert, curator of amphibians and reptiles at the American Museum of Natural History, decided that Munro's report on the copperhead's eye movements was in fact a scoop, though a San Diego herpetologist had once noticed, but not published, such movements in rattlesnakes.

Morgenroth, however, "is not the first man to look a fish straight in the eye," says Bogert. A number of ichthyologists have noted "wheel" movements in fishes' eyes. But they have not gone much further than merely to note the phenomenon; there has been little or no attempt to explain it.

Morgenroth, though a little disappointed to learn that the discovery is not new, makes bold to offer some conjectures as to why fishes and reptiles have gyro-controlled eyes.

"First," he writes, "let us agree on some terminology. We'll distinguish between eyeballs with two degrees of freedom (2F) and eyeballs with three degrees of freedom (3F). Offhand, it might appear that an animal with 3F eyes would require no level-correcting beyond that afforded by the eyeballs themselves, whereas animals with 2F eyes must accomplish this level-correcting somewhere in the visual cortex.

"Now let's see how this internal level-correcting works in an animal with 2F eyes, such as the human animal. When you lean your head sideways, everything seems to stay nicely vertical. The correction still works, though not perfectly, w when you look at an object upside down. But now comes the surprise: you cannot readily recognize the printed word upside down! Here is quite a contradiction. We seem to possess an internal level-correcting mechanism which works pretty well for most objects but not for written or other coded images.

"Only one explanation appears possible: we animals with 2F eyes do not, in fact, possess an internal level-correcting mechanism. Rather, what we do is make a mental correction for objects which we have seen frequently in various angular positions. We cannot recognize writing upside down because we do not ordinarily learn to read it that way.

"As Wiener has pointed out, we become so familiar with ordinary geometrical shapes in their countless sizes and perspectives that we are not normally aware of perspective distortion. Wiener believes that this explains why it took man a long time to learn how to draw a picture with the distortions as they appear on the retina. Learning to recognize perspective distortions and to draw pictures accordingly was a complex deductive process.

"We possess no skew-corrector in the circuitry linking the retina and the visual cortex. Rather, the correction of a picture seen from a skewed position is made through the cooperation of our senses of equilibrium and the 'group scanning' (discussed in Cybernetics) of our sets of experiences as stored in the memory.

"Group scanning explains why mammals and possibly many birds can do without a level-correcting mechanism. Why, then, do reptiles, fishes and possibly other animals need 3F eyes? It seems to me that the line between the two kinds of eyes is drawn exactly at the point of evolutionary demarcation which separates animals that learn by experience from those whose visual experiences are largely inborn-a subject on which the Gestalt psychologists have had so much to say. Reptiles and fishes come out of the egg as diminutive adults. A greater part of their visual experience must be born with them; without ever having learned to distinguish food or enemies by sight-experience, they must be able to interpret the picture correctly at the very beginning.

"It is hard enough to understand how evolution succeeded in storing visual experiences in the memory in a hereditary way. It would be still harder to understand if we had to assume that this hereditary storage was in the same form as the higher animals' learned angular experiences. The problem of explaining the hereditary storage of visual experience certainly is simplified if the picture has to be memorized in one position only. Gyroscopic eye-control makes this possible.

"The limited observations and experiments seem to support me pretty solidly up to this point, and I hope others who enjoy looking their fellow animals in the eye will find some interest in matching their observations against this conjecture. Pending their reports, I cannot resist the temptation of adding something highly speculative which, despite the present rudimentary state of cybernetics, may still be permissible.

"We have seen that animals with 2F eyes use group scanning of the memory storage for corrections of level and of distortions. Now it is not too farfetched to assume that reptiles and fishes, lacking the facility of group scanning, are just about blind as far as recognition of geometric shapes goes. Their seeing may be restricted to the recognition of colors and light changes-that is, movements. They may react to a mass, shapeless as it is, according to whether it grows larger or smaller, approaches or recedes.

"As Munro concluded, the always-vertical pupil of the snake's eye appears to have uses beyond that of gauging distance and direction. Certainly the snake benefits from its gyro-controlled eye steadiness in estimating a strike. On the other hand, my little turtle does not strike, yet it has wheel eyes. Why? Evidently there are answers in each case which lie in the animal's specific pattern of behavior.

"It is fun to speculate on how the 3F eye may affect the behavior of various . animals. For instance, compared to the agility of a dolphin or a seal, the average fish is a stiff, sluggish performer. The restrictions on its body movements may well be due to the fact that it cannot move its eyes to see beyond a certain angle. Or consider those strangely long and mobile necks of prehistoric 3 reptiles; it seems permissible to suspect that the main factor in the formation of those necks was the necessity for carrying the head on an even keel, rather than any advantage in reaching for food.

"A little higher up the ladder of evolution, observe the head movements of birds. I have not yet managed to look one squarely in the eye; I hope our friends in ornithology will do so. Birds in general have not yet achieved the independence of head-carrying that mammals possess. A grazing horse or cow can keep on watching its surroundings while it feeds, but a feeding pigeon has to interrupt its peckings for food every moment to resume the watching position. The obvious inability to watch and feed at once represents a major decrease in survival value. It enormously reduces the chances of catching . food and creates an interval of hazard while getting it. This does not mean that birds must possess the wheel-eye movement. It does suggest, however. that the development of the neural mechanism responsible for Gestalt perception, as discussed by the cyberneticists, must have taken a decidedly different turn in birds from that in mammals.

This department's illustrator, Roger Hayward; who on several occasions has illustrated here other people's proposals for huge telescopes, this month proposes a giant reflector that should end all giants. It has the shape of an immense eyeball with a 1,000-inch aperture [see illustration]. Within its sphere, 108 feet in diameter, are arrayed rings of circular shelves like the tiers of seats in a football bowl. On these seats are hundreds of thin mirror "tiles" (tesserae), each approximately three feet square and spherical in curvature. All have a common center of curvature at the top of the telescope. The sphere, floating in a pool of water, can be rotated in any direction. It may be clamped in declination by the electromagnet shown at its right, and thereafter a submerged truck-tire drive will move the tremendous eye in right ascension in synchronization with the turning of the earth.

The optical inspiration of Hayward's super-giant reflector is the famous lighthouse-lantern lens invented by Augustin Jean Fresnel. This kind of disk, with a sawtooth profile on the side toward the light source, is familiar today in railroad and traffic signal lamps. Hayward anticipates the objections that his "big eyeball" would have serious trouble with diffraction at the edges of its many composite mirrors and would be limited in resolving power to that of any one of its component mirror tiles. Against this he points to several compensating advantages. an immense gain in light-gathering power, speed as great as f/l, freedom from coma, and reduction of astigmatism to no more than that of a single one of its composite mirrors. Hayward writes:

"Ever since the days of Fresnel there has been intermittent interest in optical systems composed of many small elements arranged in a convenient manner. In the case of the lighthouse lantern the object is to project the most light in a horizontal direction. Resolving power is of almost no interest, while the scattering of light by the many edges in the glassware is offset by the gain in speed.

"In the recent proposals in this department to build large tessellated mirrors, the over-all figure of a paraboloid has been merely approximated. Most reflecting telescopes have their primary mirrors in the form of paraboloids of revolution. This figure is ideal for a star on the mirror axis, but stars not on the axis have their images distorted so that they appear as comets flying toward the axis. This defect, called coma, is associated in part with the fact that the images formed by different zones of the mirror are different in size. In the drawing at upper left in the panel below one can see how the images of a star which is, say, five degrees off the axis fall on the focal surface at different distances from the axis, indicating the different sizes of the images formed by the various zones of a paraboloidal mirror.

"The Schmidt system overcomes this defect by using a spherical mirror for a primary, with a correcting plate the size of the full aperture to eliminate spherical aberration. However, a Schmidt correcting plate 1,000 inches in diameter would be pretty hard to make, since it would have to piece together a number of sections of glass, which would be difficult to match optically. Yet similarly hard to take would be the aberrations of the spherical mirror if the correcting plate were omitted. The drawing at the upper right in the same panel shows how the images of a field of stars would be distributed over a multiplicity of focal surfaces with different focal lengths. A plate could be placed at only one of these focal distances, and all the others would be out of focus.

"However, each individual zone of a spherical mirror has a good focus over a considerable field. Therefore, if one were to construct a series of concentric spherical zones it would be possible to arrange them so that they would have a common center and were so graduated in radius of curvature that their images would fall on the same surface and all have the same size. The drawing at the left in the middle row shows an array of such surfaces. Of course for a large instrument such as we are considering there would have to be more than the three zones shown in the figure.

"Another approach to the problem is indicated by the drawing at the right in the middle row. The physicist Henry A. Rowland, of ruling-engine fame, pointed out that the equal conjugate foci of a small part of a spherical mirror lie in a circle whose diameter equals the radius of curvature of the mirror. He applied this principle to the spectroscope, but in the case of the telescope one focus would be at a star-that is, at infinity-and therefore the circle with which we are concerned would be half this size. If we arrange a series of mirrors around a Rowland circle so that their foci are at a point on the same circle, then their images will all have the same size and the focal surface will be concave.

"This leads to four solutions: the one at the right in the middle row and the three in the bottom row. The three at the bottom are more curious than useful. The remaining one has the most promise and is the one embodied in the big eyeball telescope.

"In these four schemes the focal surface is spherically concave;. in the one that resembles the Schmidt system, it is spherically convex. In this one a knife-edge test at the center of curvature of all the mirrors would serve to line them up and could be carried on while observations were in progress. In all five systems the mirror elements might need astigmatic surfaces to cure rather considerable astigmatism unless the individual mirrors were rather small. The problems associated with holding them in position and keeping them that way are not simple, although the weight need not be large compared with any comparable single mirror.

"If we assume that these technical details are capable of solution, it would be fun to imagine what sort of mountings would be required if the extremely radical systems were really exploited (there would be no advantage in moderately radical instruments). The ultimate speed for the first of the last four systems, which may be thought of as a Schmidt with every ring of mirrors at the neutral zone, is f/0.5. In practice, allowing for the light losses due to obstruction by the supports for the observing cage, an equivalent of f/O.6 might actually be accomplished. The other schemes, with the focus lying in the mirror circle, would be limited to f/1, which might have to be shaded to an equivalent f/1.2 to accommodate all practical considerations.

"To give an idea of the tremendous light grasp of these really fast optical systems we might take the 100-ineh telescope as a yardstick. The focal ratio is 5, and the limiting exposure is about 40 minutes, at which time the fogging of the plate due to the light of the night sky prevents further recording of faint stars. At f/1.2 an exposure of 138 seconds would accomplish the same, and at f/0.6 only 85 seconds would be required. With speeds like this, the telescope could even be held still and the film moved across a slot to keep pace with the moving image. The width of the slot would determine the exposure. For so short a time manual guiding of the plate would hardly be necessary. Photocell guiding would be quite feasible because of the extreme brilliance of comparatively faint guide stars at this focal ratio."

pitch-polishing laps for telescope mirrors can be brought into contact with the glass and kept that way by channeling the lap into facets. The channels give the pitch a space into which to flow in settling to contact. To carry the same idea further and assure better and quicker contact, each facet may be subfaceted with smaller channels by the use of onion sacking obtainable at grocery stores. However, if the sacking is cold-pressed into the pitch, the pitch will tend to flow up and around each of the meshes and lock the sacking in. A correct method, as used professionally by F. B. Ferson, is to dip the cold lap in nearly boiling water for an instant, lay the sacking, previously wetted, upon it, press it in flush by hand and strip it off before it can become locked in. Then dip the lap in hot water and apply it to the mirror long enough to flatten the tops of the subfacets without closing the little channels.

There are two kinds of onion sacking. Each is square-meshed but one is coarse, like strong wrapping twine, and the other is finer, like coarse sewing thread. The coarse kind makes deeper, longer-lasting channels. Wash it thoroughly before use.

E. B. McCartney of Minneapolis uses a woven square mesh embedded in sheets of cellulose acetate which is listed in mail-order catalogues for admitting ultraviolet radiation into hen houses. This is easier to handle than onion sacking, can be used against dry mirrors and makes a perfectly square pattern, though this has no actual importance. However, its channels are not so deep.

Bibliography

THE VERTEBRATE EYE AND ITS ADAPTIVE RADIATION. Gordon Lynn Walls. Cranbrook Institute of Science, 1942.

AMATEUR TELESCOPE MAKING. Edited by Albert G. Ingalls. Scientific American, Inc., 1952.

AMATEUR TELESCOPE MAKING-ADVANCED. Edited by Albert G. Ingalls. Scientific American, Inc., 1952.

Suppliers and Organizations

The Society for Amateur Scientists
5600 Post Road, #114-341
East Greenwich, RI 02818
Phone: 1-401-823-7800

Internet: http://www.sas.org/