Evolution and the Eye: The Darwin Bicentennial and the Sesquicentennial of the Origin of Species

Evolution is an essential concept for anyone who considers science to be the best way to understand the natural world. It is as fully established as any scientific principle can be and is the great unifying theme in all of biology, as integral to understanding life-forms as gravity is to understanding the cosmos. On the bicentennial of the birth of Charles Darwin in 1809, and 150 years after the publication of On the Origin of Species by Means of Natural Selection in 1859, we should remember the main features of eye evolution and the prominent place the eye holds in the development and refinement of evolutionary theory. A few highlights include the antiquity of rhodopsin, the ready capacity of an eye to evolve, the effect of eyes on the diversification of life-forms, and the promising influence of genetics on developmental and evolutionary biology.

Charles Darwin (Figure 1) was well aware that his ideas would incite a storm of criticism from the general public, if not the scientific establishment. In a shrewd preemptive strike against “natural theology,” the creationist or “intelligent design” idea of the day, Darwin wrote the sixth chapter in the first edition of On the Origin of Species on “Difficulties on Theory,” the theory being that descent with modification, or what later became known as evolution, was produced by natural selection. One section of that chapter is titled “Organs of extreme perfection and complication.”

To suppose that the eye, with all its inimitable contrivances for adjusting the focus to different distances, for admitting different mounts of light, and for the correction of spherical and chromatic aberration, could have been formed by natural selection, seems, I freely confess, absurd in the highest degree. Yet reason tells me, that if numerous gradations from a perfect and complex eye to one very imperfect and simple, each grade being useful to its possessor, can be shown to exist; if further, the eye does vary ever so slightly, and the variations be inherited, which is certainly the case; and if any variation or modification in the organ be ever useful to an animal under changing conditions of life, then the difficulty of believing that a perfect and complex eye could be formed by natural selection, though insuperable by our imagination, can hardly be considered real. How a nerve comes to be sensitive to light, hardly concerns us more than how life itself first originated; but I may remark that several facts make me suspect that any sensitive nerve may be rendered sensitive to light, and like wise to those coarser vibrations of the air which produce sound.^1(p217)

The first sentence in this quote is often misrepresented as if Darwin were worried about explaining evolution of the eye, but in reality he was quite sure of his ground. One might wish that he had not been so cavalier about dismissing photosensitivity out of hand (“how a nerve comes to be sensitive to light, hardly concerns us”)—the evolution of rhodopsin and the architecture of rods and cones would have earned his admiration. One could also wish that his Victorian prose had been simpler and more direct—the last passages of the Origin are almost poetic in their elegance–but the thought is clear. Take variation, its “inheritability” (Mendelian genetics at that time had not yet been born), and the utility of primitive eyes, and a complex eye may be readily evolved.

Ideally, Darwin goes on, it would be best to trace the gradual development of complex eyes in the ancestors of presently known animals. But because the fossil record does not allow this, we should look to the entire spectrum of eyes in various extant species that are available for us to study, particularly invertebrates, because they show the widest variation in eyes, whereas all vertebrates have already attained the complex camera eye. If we do this:

I can see no very great difficulty (not more than in the case of many other structures) in believing that natural selection has converted the simple apparatus of an optic nerve merely coated with pigment and invested with transparent membrane, into an optical instrument as perfect as is possessed by any member of the great Articulate class [ie, vertebrates].^1(p218)

He then goes on to emphasize that these changes, although happening gradually through the immensity of geological time, could occur by the power of natural selection.

If we must compare the eye to an optical instrument, we ought in imagination to take a thick layer of transparent tissue, with a nerve sensitive to light beneath, and then suppose every part of this layer to be continually changing slowly in density, so as to separate into layers of different densities and thicknesses, placed at different distances from each other, and with the surfaces of each layer slowly changing in form. Further we must suppose that there is a power always intently watching each slight accidental alteration in the transparent layers; and carefully selecting each alteration which, under varied circumstances, may in any way, or in any degree, tend to produce a distincter image. We must suppose each new state of the instrument to be multiplied by the million; and each to be preserved till a better be produced, and then the old ones to be destroyed. In living bodies, variation will cause the slight alterations, generation will multiply them almost infinitely, and natural selection will pick out with unerring skill each improvement. Let this process go on for millions on millions of years; and during each year on millions of individuals of many kinds; and may we not believe that a living optical instrument might thus be formed as superior to one of glass, as the works of the Creator to the works of man?¹

This section on the eye is quoted here so extensively to show that within it Darwin stated many of the main propositions of the entire book. Also, Darwin implicitly was assuming the modularity of evolution, that it could change one part of an organism without interfering with the integrity and survivability of the organism as a whole. This had been the concern of Georges Cuvier, the founder of comparative anatomy and paleontology at the turn of the 19th century.²

The term rhodopsin in its widest sense (also now termed the retinylidene proteins) refers to a family of molecules that contain a chromophore segment of retinal (retinaldehyde or vitamin A aldehyde) and various types of protein segments, the opsins, all of them related to some extent. It is remarkable that this particular family of molecules—and more than 300 have been identified³—serves as the photosensitive chemical in all vision systems in creatures up and down the evolutionary tree. The exact amino acid sequences of the protein segments may differ, and the photochemical cascade differs in its details (eg, cyclic guanosine monophosphate as a second messenger in vertebrates and inositol triphosphate in invertebrates). Still, the basic vitamin A aldehyde–protein pairing is a constant feature. If it is not as pervasive as DNA, rhodopsin is ubiquitous enough that it also reflects the basic unity of life-forms.

Not only is rhodopsin found throughout the domain of the eukaryotes (true nucleated cells from which protozoa, plants, and animals evolved), it is also present in prokaryotes (microbes without nuclei or other internal organelles), the domains of bacteria and of archaea (microscopic nonnucleated cells originally found in extreme environments, such as volcanic hot springs or hypersalinated areas, that differ from bacteria in their biochemistry).

Because rhodopsin is necessary for vision, finding it in motile organisms from protozoa to humans is not surprising. These organisms find a light sense so useful because they can respond with some adaptive movement to what they sense in the environment. Even the simplest light-sensing organ would allow the detection of a shadow.

What is surprising is to find prokaryotes that also contain rhodopsin. What possible utility would vision of any sort have for such primitive organisms?

First, we tend to forget that bacteria (and archaea) often are motile, using flagella or the fiberlike structures termed pili.⁴ Movement can be observed in such pathogens as Pseudomonas aeruginosa and Neisseria gonorrhoeae and probably contributes to their pathogenicity. Any type of mobile capacity allows the organism to show phototaxis, the ability to orient itself and move toward or away from light. This ability allows for circadian rhythms and vertical migration in the ocean depths according to day-night cycles.⁵ In the broadest sense, it is a type of vision, and rhodopsin enables it.

Second, the rhodopsin-type molecule in these primitive life-forms is not necessarily used for vision—for its phenomenal capacity to capture the photons that trigger the biochemical cascade and the transduction of light into a nerve impulse. Here, rhodopsin may serve not as a triggering molecule but rather as a proton pump. This is a protein that is present within intracellular membranes and that serves to move protons across the membrane from one subcellular compartment to another. It is an energy-capturing device used to transfer energy within the cell and to power its metabolism. Such is the case with the rhodopsins found in microbes, of which bacteriorhodopsin has been most studied. In addition, the halorhodopsin found in some species serves as a chloride ion pump within the cell.⁶

Bacteria and archaea existed before eukaryotes and animals. These later life-forms took the rhodopsin that most likely was already available as a proton pump and used it as the basis for vision. Co-opting of a molecule serving one purpose and enlisting it for another purpose is a recurring theme in evolution. Molecules and metabolic pathways are subject to such co-option just as anatomical structures are. Evolution uses what is available. It is a consummate recycler.

Exactly how rhodopsin evolved originally is unclear. The protein is similar to that found in touch and chemical receptors in some primitive creatures and may have been borrowed from these other sense modalities. We can be assured that intermediate forms of the molecule had some utility for the cell, and that this adaptive utility provided for the retention and elaboration of the molecule. After all, molecular evolution obeys the same evolutionary principles that apply to anatomical structures. Such a system has been postulated in the evolution of the cytochrome c oxidase proton pump, another ancient molecule that in humans consists of 6 proteins, which is even more complex than the situation of rhodopsin.⁷

Molecular evolution is a relatively new field of study. It shows that well before the appearance of the multicellular animals and plants that we know as fossils, well before even the appearance of eukaryote microbes, more primitive life-forms were evolving the molecules and metabolic pathways that would provide for the elaboration of more complex life. Rhodopsin was there near the very beginning, doing its job so well that it has been conserved for more than a billion years of life on Earth.

Although a sensitivity to light is present in microbes, true eyes, or structures that provide information about the direction of light coming from different parts of the environment, are limited to animals. Darwin's prediction that intermediate forms of the eye would be found has been abundantly fulfilled. The transition from the photosensitive eyespots found in some protozoa and the pit eyes of worms and mollusks to the complex camera-type eyes of vertebrates has been well documented in animals living today and by extension in their ancestors as well. Several excellent surveys are available that discuss this in detail.^8-15

In general, eyes differ in at least 3 categories:

• 1. Structure type

•     a. “Simple” (a misnomer) or single-chambered eyes seen in some mollusks, spiders, and marine and land vertebrates, and in which the retina is concave and the image is inverted.

•     b. “Compound” or multichambered eyes that consist of individual pigment-covered tubules (“ommatidia”) in various arrays, as seen in crustaceans and insects, and where the equivalent of the retina is convex and the image is erect.

Each type of eye shows transitional forms. At least 11 optical systems are represented, including mirror optics as well as refractive optics and, most recently discovered, a telephoto lens⁹ (but no zoom lens and no Fresnel prism). The British Journal of Ophthalmology has featured many of the more interesting eyes on its cover with amusing explanatory essays by the ophthalmologist Ivan Schwab. Some of these eyes beggar the imagination.

(Here I must admit to a secret fondness for the patch of photoreceptors on the back of the so-called “eyeless” shrimp recovered from deep in the ocean abyss, near the hydrothermal vents in the midoceanic ridge in the central Atlantic. There, far from the reach of any trace of sunlight, the shrimp uses its light sensitivity to orient itself to the dim light coming from the vent, from which issues the sulfur-based nutrients it needs.¹⁶)

• 2. Photoreceptor type: There evolved at least 2 solutions to the challenge of arranging the rhodopsin molecules within a cell in such a way as to optimize the likelihood of an incoming photon hitting a molecule. Both involve changing the conformation of the cell's enclosing membrane to stack up many layers of photopigment against the photon's path.

•     a. The membrane of the rhabdomeric cell shows a multitude of microvilli, is present for the most part in invertebrates, and depolarizes during discharge.

•     b. The membrane of the ciliary cell, pushed inward with invaginations, is present largely in vertebrates and hyperpolarizes at discharge.

• 3. Mode of development in the embryo: This is seen most clearly when comparing the eye in vertebrates with the superficially similar camera-type eye seen in the octopus, a cephalopod mollusk.

•     a. The invertebrate octopus eye develops through a series of infoldings from an epidermal placode. The photoreceptors face toward the vitreous cavity.

•     b. The vertebrate eye forms from the neural plate and induces the overlying epidermis to form the lens. Invagination of the neural cup makes for a double-layered retina in which the photoreceptors face away from the vitreous cavity.

Lenses in different species may originate from different tissues in the embryo. But no matter what the source tissue, the substance that makes up the lens body must show a graded difference in density: greater in the center with a resulting higher index of refraction, less dense in the periphery with a lower index of refraction. This has been evolution's way of dealing with spherical aberration, a particularly pressing problem for the spherical lenses of marine vertebrates. This problem was supposedly recognized by the British scientist James Clerk Maxwell (1831-1879). One day at breakfast, he gazed at the herring he was served and decided on the spot to dissect its eye. He saw the lens was spherical and quickly recognized that spherical aberration would limit the eye's efficacy unless there was a gradation of refractive indices within the lens, and so it proved to be.¹⁷

The actual substances used within the vertebrate lens—the crystalline proteins—were once thought to be unique to the lens and to have evolved specially for this purpose. They are now recognized to be similar to proteins used for other purposes and to share their place with other proteins that seem to have been recruited from enzymes or other factors. Fernald¹⁰ terms this molecular opportunism and contrasts it with the widespread reliance on rhodopsins.

Because amino acid sequencing can now study the detailed differences in the various opsin proteins present in the photoreceptors of vertebrates, and because these proteins are closely related to each other, this technique is being used to make especially fine distinctions about the exact place of animals on the evolutionary tree.¹³ For instance, the alligator's photopigments are more closely related to those of chickens, and supposedly other birds, than to other reptiles or mammals.¹⁸

As soft parts, eyes are not generally preserved in fossils. Nevertheless, eyes composed largely of calcite lenses were fortuitously fossilized with some early trilobites, a now extinct type of arthropod that existed for almost 300 million years. They are seen to be present as already well-developed compound eyes at the very start of the Cambrian period some 540 million years ago (Figure 2). The first real eyes to develop must have occurred sometime before this and the trilobite fossils appear as the culmination of their prior evolution. But because trilobites (or most other animal fossils) do not occur before this period, these first eyes cannot be seen directly.^17,20

Another line of evidence that vision was important at this time is that externally these early creatures were colored. Parker¹⁷ has studied fossils from the famous Burgess Shale rocks in the Canadian Rockies that clearly show preserved soft parts in creatures from later in the Cambrian. Many show fossilized compound eyes, some of them bizarre in the extreme—one creature barely a few millimeters long has an eye half as big as its entire body. Electron micrographs show a layering in the hard parts preserved in these fossils. These layers may have acted as a diffraction grating to give the creature an iridescent metallic sheen in all the colors of the visible spectrum. Parker has actually visualized such color; it may have warned off predators, which presumably would also have had to be sighted themselves.

Vision provided such an asset to survival that it may have led to an “arms race” between predator and prey, with the development of shells for defense and modified appendages and biting mouths for offense. It is tempting to think of vision as the major factor among all the others acting on the survival of these early creatures and thus speeding up their evolution.¹⁷

What can be said is that soon after the appearance of eyes, animal life seems to have diversified into a multitude of different forms. (It is still a question whether such seeming diversity was really new or just could be newly appreciated because of the development of fossilizable hard parts.) Most of the 35 phyla now recognized in present-day life-forms evolved from the 3 or so phyla present at the start of the Cambrian. Of these 35 phyla, 6 have eyes: the cnidarians (jellyfish), mollusks (snails, clams, squids, and octopi), annelids (segmented worms), onychophores (velvet worms), arthropods (insects, spiders, and crustaceans), and chordates (vertebrates). These 6 phyla contain about 96% of the known species alive today. There may indeed have been an “optical liftoff” for the species with eyes.^9,10

For those dubious that natural selection could have evolved such a complex structure as an eye, the tried and true response has stressed the long period of geologic time and the short life spans of each generation of the early life-forms. But this long period of gestation may not have been necessary.

In 1994, Dan-Eric Nilsson and Susanne Pelger published a mathematical model of the transition from a pit eye to a vertebrate-type single-chambered eye.²⁰ Entitled “A pessimistic estimate of the time required for an eye to evolve,” it startled readers because it showed how rapidly a complex eye could hypothetically evolve. Their model is worth discussing for a moment because it raises some of the considerations—the anatomical constraints—involved in the origin of whole organs.

The authors made conservative, hence “pessimistic,” estimates of the successive steps required in the transformation of a flat patch of photoreceptors situated between a transparent protective layer on one side and a dark pigmented layer of tissue on the other (Figure 3). As the photoreceptor-pigment layers bend backward to form a pit eye, the capacity to merely appreciate the presence of a shadow improves to yield some sense of light direction. As this process continues, the anterior end starts to narrow to become an aperture with an iris. With the introduction of a lens, the resolution improves further, especially as a graded density of lens material is refined.

Assuming that the tissues would vary at most 1% in width, length, refractive index, etc, during each step, Nilsson and Pelger estimated that less than 2000 steps would be needed overall. (Unimpressed by a 1% change? Think of this: 2000 such changes could stretch a finger enough to reach across the Atlantic.¹²) Variation in only 1 feature could occur at one time, another “pessimistic” feature that could tend to overestimate the time needed. Another crucial point: each step would not interfere with the functionality of the intermediate form.

They further assumed a slow rate of evolution, that is, of some member of the species already having a variation itself passing a further favorable variation to a subsequent generation. This variation would improve vision a bit and thus increase survivability so that it would spread within the species population with each successive generation. Actual observations by naturalists in the field provided values for such considerations as heritability, probability of variation, and intensity of selection.²¹ It turned out that fewer than 400 000 generations would be more than adequate to develop a complex eye. Because similar creatures alive today have a short life span, one can safely assume a generation to take a year. Thus, the whole course of change could reasonably be estimated to take less than half a million years—just a passing moment on the geologic timescale.

Note that the model starts with photoreceptors already in place in a simple eyespot. The estimate applies to a single-chambered eye, but the first eye was a compound eye. Could the scheme also apply to compound eyes? Land and Nilsson¹² conceive of a loose collection of eyespots in the ancestor of trilobites changing to form their compound eyes in a similar manner. So the whole business could have occurred just before the earliest Cambrian fossils, or about 540 million years ago.

This poses a trick question, because the answer to both “once” and “many times” could be yes, sort of, in a way. A classic paper in 1977 traced the varieties of eyes in various animal lines and concluded that eyes evolved separately at least 40 times and perhaps even 60 times during the history of animal life- forms.²² This so-called “polyphyletic” concept of eye origins is now subject to the findings of experimental genetics, which cause some scientists to argue for a “monophyletic” origin under the influence of so-called master control genes. This riddle has been expanding our ideas of exactly how genetic programs influence the development of organs in general.

Experimental geneticists in the 1980s identified the homeobox, DNA sequences within genes that were particularly important in the regulation of the development of form in the embryo. These homeobox genes encode transcription factors that turn on cascades of programs from other genes.

In 1995, Halder et al reported experiments in the trusty laboratory animal Drosophila melanogaster, the fruit fly.²³ They were able to isolate its homeobox gene eyeless. (Drosophila biologists traditionally name a gene for the effect its inactivation has on the developed organism; totally deactivating this gene early in the developing fruit fly causes the birth of an individual without eyes.) But this time they deliberately activated the eyeless gene in the somatic cells of the embryo, where eyes do not normally develop. This manipulation induced the development of ectopic eyes on the legs, the wings, even the tip of the antennae. The eyes were anatomically complete and the photoreceptors could be activated by light. They then trumped their result by introducing the equivalent gene Sey (“small eye”) taken from a mouse, switching it on, and inducing the same effect, the ectopic development of a fly's compound eye with a gene from a vertebrate.²⁴

These startling experiments made scientists seem like sorcerers. The mouse and fly genes were found to have a high degree of similarity in their amino acid sequences. They were termed “master control genes” for eye development and similar genes were found across the animal spectrum in worms, squids, and humans, in whom its name is Pax-6. (A mutational change in Pax-6 is associated with various anterior segment anomalies.^25-27 Homozygous loss produces anophthalmia with brain malformation severe enough to kill the neonate, showing that Pax-6 is not only important to eye development but also to development of the brain itself.²⁸)

It was one thing to know that DNA unified all living things in a way unknown to Darwin. It was even more humbling to learn that the actual gene tool kits for the construction of eyes and other organs were so intensely conserved during the course of evolution that humans shared them with mice and, mirabile dictu, the fruit fly.

However, the concept of Pax-6 as a master control gene that evolved once, then turned on secondary programs that actually construct the wide variety of eyes, is probably too simple. Pax-6 has other important roles, particularly in formation of the head. It is present in some creatures without eyes, including some jellyfish. Other genes can also switch on eye development, and they may be upstream of Pax-6. According to Fernald:

Genetic control of eye construction is staggeringly complex and it seems more likely that an interactive ‘gene network’ regulates development of all complex organs, including the eye. . . . Because Pax-6 existed before eyes, like other developmentally important genes, it may have been recruited for the formation of eyes, just as was opsin.^10(p448)

By 1996, at least 72 genes involved in mouse eye development had been mapped, and doubtless there are more now.²⁹

In other words, Pax-6 may have been recruited at different times for different animals, so that, even if complex genetic programs were lying in wait, they were not always turned on, and the eyes themselves evolved many times. This whole question is forcing a rethinking of the classic evolutionary concept of homology, the idea that similar structures in different species could be identified as being related owing to shared ancestry.^29-32

I first encountered this mysterious phrase displayed above the blackboard in the science classroom as my introduction to high school biology. It had the solemn incantatory power of a religious dictum. At that time it had been already out of favor, if not actually totally obsolete, for 50 years.

The phrase had been coined by Ernst Haeckel (1834-1919), who wished to announce a “biogenetic law” stating that the developmental stages of the vertebrate embryo recall the adult forms of creatures that appeared earlier during the course of evolution. I expect he was hoping that such grand principles would make sense of the myriad realities of biological form and was a bit too impressed by the physicists of the time, who had inherited a stable of mathematical “laws” and were busy at the time formulating new ones. But evolution has made biology a historical science and law is the wrong term to describe the regularities with which it operates.

In the words of Steven Jay Gould, who had the same experience as I did with Haeckel's phrase, “Haeckel's biogenetic law was so extreme, and its collapse so spectacular, that the entire subject became taboo.”^33(p2) It hung on in a diluted form, which called attention, for instance, to the transient appearance of gill slits in the mammalian embryo, without being thought to explain much either in embryogenesis or in evolution. Until relatively recently, embryologists and evolutionists did not talk to each other. This was too bad because evolution can be thought of as basically the interplay of environment with development. Ignoring embryogenesis was a mistake, especially since Darwin himself had referred to embryology often in the Origin and in private letters.^34(p282)

Here, we should remember that the eye was important to early embryology, in that it helped to formulate the concept of induction, defined as a tissue affecting the subsequent development of another. It was on the developing frog eye that Hans Spemann (1869-1941) did his famous experiments in 1901. He could see that normally the lens developed from the overlying ectoderm. But if, before the lens developed, he excised the underlying eye anlage, the lens would fail to form. Transplanting the eye anlage to other parts of the embryo induced lens development in the overlying ectoderm wherever it was. Later experiments inducing development of the neural tube led to the concept of the organizer. The organizer was a breakthrough in understanding embryogenesis and caused a sensation at the time.^35(p163)

Beyond these clues, however, embryology described development but could not say why it occurred in certain ways and not others. Embryologists did not join in the modern synthesis of classic Darwinism and population genetics when it was formulated in the 1930s and 1940s. It has taken experimental genetics to show that it is not just the genes that are important, but that form is determined by when and where they are turned off and on, and for how long. The big surprise has been finding that similar approaches to patterning the embryo hold for species so far apart in the evolutionary scale. The current rapprochement between evolutionary biologists, molecular geneticists, and developmental biologists (familiarly termed “evo-devo”) is clarifying, in a way never before possible, how both small and large changes in form may occur.^34,36,37 There are evidently core processes in development whereby changes in the genotype—whether by mutation, gene recombination, or chromosomal reassortment during sexual reproduction—may have profound effects on the phenotype and speed the evolution of large changes. Genetic switches that turn genes on or off are so important in determining the phenotype that they are particularly influential sites for evolution to act upon. Understanding these may make it easier for us to bear the fact that the human genome has just 22 500 genes, only 6 times that in the genome of a bacterium.³⁷

Darwin constructed his theory on 3 main ideas: that variations occurred, that heredity passed some on to a subsequent generation, and that natural selection determined which would remain and evolve over time. He did not know the genetic and cellular mechanisms involved, and science has been actively questioning, exploring, and modifying his ideas since then. Now, 150 years after the birth of On the Origin of Species, what he termed one long argument continues, promising ever more insights into the nature of life on earth.

For those readers who have come this far and are still wondering about that elegant final passage in the Origin, here it is:

There is grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one; and that, whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved.^1(p459-460)

Correspondence: Ronald S. Fishman, MD, 47880 Cross Manor Rd, Saint Inigos, MD 20684 (rsfishman@earthlink.net).

Submitted for Publication: April 6, 2008; final revision received April 22, 2008; accepted April 22, 2008.

Financial Disclosure: None reported.