The Human Eye: A design review

Published on Wednesday, July 12, 2017 By Brad Wardell In Life, the Universe and Everything

Note: This post was originally written by my friend Steven Den Beste who passed away last year. This will be a series of articles designed to preserve his writing for future readers to enjoy the way I did.

The Human Eye: A design review

Occasionally I see creationists point to the human eye as a miracle of design, as if this somehow is evidence of divine origin for the human form. Unfortunately, from an engineering perspective, the human eye is seriously suboptimal. It simply isn't that good a design.

When I was in college, I studied Computer Science. At my alma mater, that was taught by the Department of Mathematics, which of course was part of the College of Science. So while my friends who were majoring in Electrical Engineering, in the College of Engineering, were taking Statics and Dynamics, I was taking a Science curriculum, and I studied Physics and Zoology.

One term of Zoology was Vertebrate Physiology, and as a promising engineer who has always been entranced by science anyway, I found it fascinating. One of the things we studied was the vertebrate eye. And I was appalled by what I found. Frankly, not only is it not a wonderful design; on the contrary it's one of the worst designed mechanisms in the body. (It contrasts, for instance, with the sheer elegance of the design of the kidney.)

So let's do what we engineers refer to as a design review on the human eye, and discover what happens.

It may seem obvious, but we start with a formal declaration of the function the system is supposed to perform: to receive and process light of several frequencies in order to derive information about the environment, while using as little mass as possible to do so.

Because of this statement of function, it becomes clear that the critical component of the system is the retina.

The retina is a thin film of tissue a bit larger than a postage stamp which covers the inside surface of the eyeball away from the lens. It consists of light sensing cells embedded in a support matrix of epithelial cells, along with an array of blood capillaries which bring in nourishment and carry away wastes so that all these cells don't die.

And right away, the flaws begin. The vertebrate retina is a terrible design. The optic nerve comes into the eyeball at a certain point, and the nerve fibers spread out across the surface of the retina. Each individual nerve fiber reaches its assigned point, burrows down into the retina through several layers of epithelial cells, and ends with the light receptor itself pointing away from the lens of the eye, which is the direction from which the light must come. As a result, incoming light strikes the surface of the retina and must penetrate through multiple layers of inactive cells and then through the body of the nerve itself before it reaches the active point where it might be detected. This both diffuses and attenuates the light, decreasing the efficiency of the retina in accomplishing its function.

It's possible to do this better. We know this because the mollusc eye does it right. In the mollusc eye (typified by the octopus, squid and chambered nautilus, all of which have excellent vision) the optic nerve spreads out under the retina, and each nerve burrows up through the retina and ends with the light sensor on the surface of the retina, pointing towards the lens. This means that there is no attenuation of the light before it reaches the active components. (Just incidentally, this also means that molluscs have no blind spot. Vertebrates have a blind spot because there are no light receptors at the location where the nerve passes through the retina.)

The mollusc design is completely practical, but vertebrates don't use it. Our design is second rate. This alone is sufficient to demonstrate the inelegance of our eyes, but the problems don't stop there.

Some mammals have found a kludge which ameliorates this poor design to some extent. Beneath the layer where the light sensors are, there's a reflective layer. If a given photon passes through the nerve and doesn't set it off, it reflects and is given a second opportunity on the way out. That's why the eyes of deer and cats seem to glow so brightly when you hit them with a flashlight at night; the light from your flashlight is reflecting off that layer. This substantially improves the sensitivity, though not to the level of the mollusc retina -- but primates (including humans) don't have this. So even among vertebrates, humans don't have the best design available.

Continuing on: like any camera, the eye requires a focusing mechanism to sort the light. Without it, it would be like a piece of film held directly under a lamp; exposed, but showing no detail.

At this point it's necessary to make a design tradeoff. There is a certain maximum density of light receptors possible in the physical retina. To get more receptors, the retina must be made larger.

There is a physical phenomenon in focusing systems called depth of field. What it means is that for any piece of film and a lens at a given distance from that film, there's a certain range of distances on the far side of the lens which will be in focus. Anything further away or closer in will be out of focus and blurry.

Generally speaking, if the lens is closer to the film, the in-focus depth of field will be greater, but the image will be smaller. If the lens is further away from the film, the image will be larger, but the range which is in focus will be smaller. That's why a camera which uses a small film frame close to its lens doesn't require a variable focus, while a large camera such as a 35 mm SLR must have one.

If our eye was small, we could get by with a fixed focus lens. But that would limit the number of light receptors too greatly. So our eyes use a much larger retina, which means that the lens must be further away, which forces it to have a variable focusing mechanism.

As is often the case in physical systems, scaling is not a linear process. There's a scaling principle known as cube-square which means that some physical properties increase as the square of the scaling factor, while others increase as the cube. (Actually, in some physical systems there are factors which scale as the fifth power.)

By making the eye twice as large in diameter, the retina becomes four times as large because its size is proportional to the surface area of the eyeball, which rises as the square. But the clear jelly which fills the eye (known as aqueous humor) is proportional to the volume of the eyeball, which rises as the cube. What this means is that as you increase the size of the eyeball, proportionally less of the mass is active (the retina) and more is passive (everything else). Since we'd like to get as much out the mass we invest, making the eye bigger is fundamentally undesirable.

In our eyes, the variable focus is accomplished by using a flexible somewhat-rubbery lens, and using muscles to pull on it to change its shape, making it thin or fat as necessary to change the focal length. And here we have our second major poor design, because this entire approach is faulty.

Among other reasons that it is poor is that as we grow older, that lens grows more stiff and less flexible, and we loose the ability to change it. That's why nearly everyone requires bifocals or trifocals when they grow older. The external lenses provide the ability to change focal lengths which the internal mechanism has lost.

But it isn't necessary to flex a lens to change focal lengths. A camera or binocular has a focusing mechanism, and their lenses are made of glass, which is one of the least flexible substances known. So how do they do it?

They use two lenses and move one relative to the other. And indeed that solution was possible in the vertebrate eye, because we also have two lenses. The other is the cornea. Instead of changing the shape of the lens, it could have been designed to move the lens with those same muscles. Then we wouldn't have required help as we grow older, since the muscles don't stiffen up the way the lens does. But that's not what we have; again we have a second rate solution.

That decision to use a large retina, with a long focal length focusing mechanism, has another side effect. In this case, it's referred to as width of field. In SLR terms, this is typified by the difference between a telephoto lens (which has a narrow field of view) and a wide angle lens (which has a wide one). By making a wide field system, more can be seen at once, but with worse accuity for any given object. By using a narrow field, less is seen at once, but what is seen is seen better. In order to get acceptable visual accuity in our eyes, it's necessary to have a fairly narrow field of view so that enough light receptors are applied to a given item for us to tell what it is. But survival requires being able to see everything around you, which seems to require a wide field of view. How to resolve that contradiction?

There are various answers possible, but the one chosen was to put the eyeball in gimbles, and here's where the design really falls apart. The spherical eye is placed in a socket of bone covered with tissue, and muscles are attached to rotate the eye in various directions. (Because of this, the eye is forced to be spherical, irrespective of whether that's really the optimal shape, which as it happens isn't even close to optimal.) Since like any moving part it has to be lubricated, tear glands are added, and to permit regular replacement of the lubricant and to drain off excess, you need large sinus cavities (also supported by bone and covered by flesh) and all of this makes the head heavier, so the bone of the neck has to be larger and stronger and more muscles are needed in the neck to support and move it, and when the smoke clears, you've invested more than a kilogram of mass all told to support active retinal area the size of two postage stamps. The efficiency of the design is ludicrous. Far too much mass is being used to far too little effect.

Where did the design go wrong? Sometimes in engineering when we end up with a monstrosity like this, we have to retreat to the beginning and start questioning really fundamental decisions. In some cases they are so fundamental that we don't even realize that we made them. That's what we have to do here.

The critical question turns out to be this: Why are there only two eyes?

When it became necessary to increase the amount of retinal area, this design did it by scaling up two eyes, resulting in a small increase in retinal area but a huge increase in dead weight due to nonlinear scaling. That's where the design went wrong.

Look at all the inherent advantages of a small eye: High mass efficiency (high proportion of mass dedicated to retinal area relative to a larger eye) no variable focusing mechanism needed (saving weight), wide field of view (removing the need for a gimble system, leading to still more savings of mass and furtherincreases in efficiency). Against this is the single disadvantage that the absolute amount of retina is too small.

But by scaling two eyes to increase retinal area we fall into all the design traps described above.

What if instead we doubled the number of eyes? Doing so would double the amount of mass used and also double the amount of retina. It scales linearly!

If we need twenty times as much retinal area, why not use twenty times as many small eyes instead of two really huge ones?

Again, we know this is possible because it actually exists in the real world in living animals. In fact, again the molluscs are one example (where some clams have dozens of eyes). But better examples can be found among the spiders, some of which have eight eyes.

Take that big empty human forehead, which is going to waste now, and cover it with a big grid of tiny eyes. Spread some of them around on both sides on the temples, to provide peripheral vision. Not only is it more efficient, using far less mass to provide far more actual vision, but by increasing redundancy it makes it far harder to be blinded. (In some cases we invest extra mass to increase redundancy, but in this case we get increased redundancy by reducing mass. It's a double win.)

Now your first reaction is that such a person would be hideous. Well, beauty and ugliness are judgements which are genetically programmed into us. If everyone was like that, they'd think that one of us was ugly, too.

So why was it done the way it was? In a creationist scenario, the only possible answer is that God moves in mysterious ways which is simply another way of saying "Damned if I know." (Or worse, Man was created in God's Image which leads to the question of why God has such a lousy design...)

But evolution has an answer: We have two eyes because the fish from which we are descended had two eyes. Genetically speaking, it's far easier to change the shape of something than it is to create something entirely new. To take the fish eye and make it an animal eye, mutations were much more likely to alter the existing structure than to create something entirely new. That's because natural selection doesn't create perfect designs. It just creates things which are better than before, with no long range plan.

What we have makes perfect sense as the end product of a long sequence of incremental changes. However, it makes no sense at all as a unique design from scratch for this particular application.

If God designed the human eye from scratch for this application, then God is an incompetent engineer.

As I studied physiology, I found example after example of poorly designed mechanisms that didn't make sense as original designs, but which made perfect sense as modifications of previous structures which were used for different functions. That is most of what convinced me that evolution is correct and that creationism is fantasy.

Some other examples:

  • Why does the birth canal run through the middle of the pelvis?
  • Why does the backbone run down one side of the trunk instead of through the middle where it would be more balanced?
  • Why does the ankle attach at one end of the foot instead of in the middle?
  • Why are there toes?
  • Why is it that nearly every part of the brain is as far as possible from the piece of the body with which it is associated?
    • Why is the motor control center for the right side of the body on the left side of the brain, and vice versa?
    • Why is the vision center at the rear of the brain, as far from the eyes as possible -- and on the opposite sides?
  • Why is it that fully 90% of the genetic material we carry around is useless?
  • Why do we share a single canal through the neck through which we both breath and swallow?

Anyone who tries to claim that the human form is some sort of engineering marvel simply hasn't looked closely enough.