Evolution of Vertebrate Vison

Visual systems are interesting to most people because they define so much of how we perceive the world. Humans are highly visual organisms– vision dominates most of our other systems, and has taken center stage in much of our philosophical tradition. Just think about phrases like “seeing is believing” and “the show me state.” Today we will get into the actual biological basis of vision, and look at some of the recent evidence for the evolution of color vision in vertebrates. Right now there is an incredible amount of work being done on all aspects of vision, from the mechanisms behind the divergence of photopigments to the Hox genes controlling the embryonic development of the structure of the eye. All of this work is leading to a unified view that visual systems evolved early in evolutionary history, that all organisms share the same fundamental molecular mechanisms, that there are genetic mechanisms that are uniform throughout all metazoans, and that there are a huge array of adaptive variations that demonstrate without a doubt that mutations can provide intermediary steps which confer a selective advantage to organisms– in other words, changing one part of the eye does not necessarily mess things up. 

Graphic image: http://www.psdgraphics.com/file/rainbow-colors.jpg

You’ve all heard about Darwin’s rather famous musings on vision. He began by expressing a very commonly held belief that the human eye is complete and perfect in its anatomy and physiology, and any changes to one aspect of the eye would render it useless. The idea being that something so complex and interrelated could not evolve through a series of gradual steps. The idea of gradual change was hotly debated at the time he wrote The Origin, and he poised the issue as a possible way to disprove his theory.

Science is a self-correcting knowledge system. It is possible to find and correct mistakes without having to use other ways of knowing. This often happens when there are new methodologies and instrumentation introduced, like telescopes in astronomy during Galileo’s time or microscopes leading to breakthroughs in cellular biology and embryology in  the 19th Century. It is happening today with the introduction of modern molecular genetics.

Science is also self-correcting when there are breakthroughs that lead to new facts, such as the discovery of the structure of DNA.

In the case of Darwin’s work, we had someone who didn’t introduce a new methodology or discover an exciting new fact, instead we had someone who provided a major synthesis of factual knowledge in biology that led to a re-evaluation of pretty much everything in biology. Following the publication of the first edition of The Origin there was an incredible amount of new research– biologists went nuts looking for ways to investigate whether the new theoretical framework could give them new ways to view their existing data and new ways to conduct research. I’ll talk about this more in my next lecture, but for now its important to know that between the time Darwin wrote the first addition and the time he published the sixth edition of The Origin there was a tremendous amount of new information about embryology, cell biology, physiology and anatomy that he could draw on to assess the different things that he had laid out in the first edition.

Quote: http://darwiniana.org/eyes.htm

In the first edition he had laid out a fairly thorough and lengthy description of how you could use vision as a test of natural selection, and in the first edition he had expressed doubts that biologists would find that the human eye could have evolved through evolutionary mechanisms such as natural selection of gradual steps. In the first edition he reviewed data about primitive eye spots and what was known at the time about variation in visual systems among animals and concluded that there was not enough support at that time to believe that human vision could have evolved through natural selection.

In 1859 he wrote:

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.

However, by the sixth edition in 1872 he changed this to

the difficulty of believing that a perfect and complex eye could be formed by natural selection, though insuperable by our imagination, should not be considered as subversive of the theory.

And he followed this with a lengthy discussion of new research which had provided the data that had changed his mind. What we can do now is take up his challenge to his own theory and ask whether the data from new techniques in molecular genetics supports or refutes his challenge to his own theory. We can use science to challenge science, and see if we need to make corrections to his predictions.

Quote: http://darwiniana.org/eyes.htm

We can start out looking at the complete package– the human eye is a complex and highly integrated system, as Darwin noted. It is not, however, perfect, nor is it the best. For example, the way in which vertebrate eyes develop by infolding puts the photoreceptive cells on the back side of the retina and the nervous tissue on the front side, requiring the nerves to pass through the retina forming a blind spot. For those of you who were in the behavioral ecology class last semester, I went over the differences in the way that vertebrate eyes develop compared to cephalopods like octopus and squid– in cephalopods the photoreceptive cells are on the front side of the retina and the nerves are on the backside, so there is no blind spot.

The way the vertebrate eye works is that photons enter the eye through the cornea, which does some focusing. I have an astigmatism, which means that my cornea is misshapen and it doesn’t properly focus the image, which means that my lens has to do more work to focus the image onto the  retina without the help of my cornea. Since the lens has to do additional work, the muscles working the lens do more work, and I get muscle fatigue in my eye leading to headaches. My glasses help to do the work that the cornea should be doing.

The iris expands and contracts to regulate how much light gets into the eye. Too much light can bleach all your photopigments and make it hard to see– that’s what happens when you are exposed to a quick, bright flash of light. When you look into the flash of a camera you have a hard time seeing for a few seconds until your photopigments get reset. At night you can be blinded by other cars because the bright light comes at you so quickly that your iris may not have time to contract.

 After the photons pass through the cornea and the iris and the lens, they pass through the ocular fluid in the vitreous chamber.  This ocular fluid is mostly water, and it distorts the image because water scatters short wavelength photons more (this is why the ocean is blue). This distortion by the ocular fluid is compensated for by pigments in the tissues surrounding the photoreceptors, which essentially even out the amount of scattering between the different wavelengths of photons.

We’ve already seen a couple of less-than-perfect aspects of the structure of the human eye and we haven’t even made it to the photoreceptors….

Light of various wavelengths is focused on the retina which contains the photoreceptive cells. Before going into detail on how rods and cones work to produce vision, especially color vision, I want to make a quick detour and talk about Hox genes in the development of the overall structure of the eye.


Over the past fifteen years there has been a really exciting discovery– there is a Hox gene called PAX6 that controls the development of a large bundle of anatomical traits which make up the eye. We can look at this in humans– there is a condition called aniridia (no iris). People with aniridia need to where special glasses to reduce the amount of light entering their eyes because they don’t have an iris– that means that there is no way for them to reduce the amount of light entering their eyes.

The interesting thing is that this condition appears to be linked to a genetic defect in a Hox gene that controls eye development.




And this is the same Hox gene that is involved in producing the eyeless mutant in Drosophila as well 


In experimentally producing eyes on Drosophila in places no eye was meant to be– when PAX6 is manipulated on the limbs of a Drosophila during development, complete eye structures develop from tissues that would not normally produce eyes. These eyes are nonfunctional because there are nor connections to the nervous system, but the anatomical parts are all there.

These experiments form the basis for considering what has been called the eyeless gene to be the master control gene for eye morphogenesis (Halder, 1995).


Because homologous genes are present in vertebrates, ascidians (sea squirts), insects, cephalopods (squids and octopus), and nemerteans (worms), eyeless may function as a master control gene throughout the metazoa (Halder, 1995).





Until fairly recently there was the idea (held by Ernst Myer among many others) that eyes had to develop many different times in different lineages of animals because there were really big differences in their structure and especially in their photoreceptors. However, this appears to be wrong– recent work has shown that there is a unity in the basic structure of all visual systems at the molecular, genetic, and cellular level, and that the developmental pathways have profoundly similar mechanisms



In fact it looks like animal visual systems are actually highly conservative of
basic elements like cell types and chemical consituents

When we look at the eye of an insect vs. a human, one big difference is that different cells types are used to make photoreceptors. Insects have rhabdomeres and humans have ciliary cells that form rods and cones. This is the basic pattern amongst invertebrates and vertebrates.

This difference was a major reason why biologists assumed that the compound eye of insects was so different from the vertebrate eye that they represented completely different evolutionary lines. However, with the discovery of the common use of PAX6 genes in insect and vertebrate eye development, scientists started to look for other commonalities. Just a couple of years ago we learned that an ancient lineage, dubbed a “living fossil,” has cells identical to vertebrate ciliary photoreceptor cells– but that these cells are in their brain tissue. Platynereis uses rhabdomeres in its eyes, but  in its nervous system it has both types of cells. This was an accidental discovery– a researcher was looking at recently published photographs of Platynereis nervous tissue and realized that he was looking at cells that had the same structure as vertebrate cone cells.

Now I want to concentrate on vertebrate color vision, so we will look at ciliary cells (rods and cones) and the retinal opsins

Protostomes exhibit a variety of different kinds of eyes, leading to the suggestion that eyes have evolved independently many times; in addition, eyes differ in more than just their apparent organization, and there are some significant differences at the molecular level between our photoreceptors and arthropod photoreceptors. It's all very confusing.

Research on a particular animal model, the polychaete marine worm, Platynereis dumerilii, is resolving the confusion. The short answer is that there are fundamentally two different kinds of eyes based on the biology of the cell types, and our common bilaterian ancestor had both—and the diversity arose in elaborations on those two types.

Evolution of the vertebrate eye: opsins, photoreceptors, retina and eye cup

Trevor D. Lamb, Shaun P. Collin & Edward N. Pugh, Jr

Nature Reviews Neuroscience 8, 960-976 (December 2007)


The "living fossil," Platynereis dumerilii. (Credit: Maj Britt Hansen, Photolab, EMBL heidelberg)


Restricted part of the spectrum used for vision probably due to fact that evolution initially occurred in water, so even organisms that evolved terrestrial forms were initially set up to be most sensitive to wavelengths common underwater. 2/3rds of animal phyla have some form of vision




The rod sensitivity is shifted toward shorter wavelengths compared to daylight vision, accounting for the growing apparent brightness of green leaves in twilight.

"While the visual acuity or visual resolution is much better with the cones, the rods are better motion sensors. Since the rods predominate in the peripheral vision, that peripheral vision is more light sensitive, enabling you to see dimmer objects in your peripheral vision. If you see a dim star in your peripheral vision, it may disappear when you look at it directly since you are then moving the image onto the cone-rich fovea region which is less light sensitive. You can detect motion better with your peripheral vision, since it is primarily rod vision."

The rods employ a sensitive photopigment called rhodopsin.



There is an obvious similarity between rods and cones of all vertebrates.

You can see the way in which the tops of the cells have stacks of membrane plates that contain the photopigments

Photoreceptors are light-sensitive cells that comprise the outermost layer of the retina. They are highly specialized ciliated neuroepithelial cells that convert light into chemical signals. Their distal (top) parts are specialized capture light and their proximal (bottom) parts are specialized to transmit nerve signals.

This is a rod

The outer segment of a rod is a modified cilium made up of around a thousand flat disks piled on top of each other– this provides a large surface area. The visual pigment, rhodopsin, is embedded in the membrane covering these disks.

Disks can be shed and replaced to replenish the visual pigments.




Rods have more photopigment than cones, are slower to respond, have low acuity and have only one photopigment type-- rhodopsin

Contrast to cones which are fast responding, have high acuity and different cones can have different photopigments– there are different variations on cone opsin

The total size of the retina is limited, so there are only so many total number of rods and cones that can be packed in

There are connections between individual photoreceptor cells, so vision is affected by which cells are next to each other

Rods are distributed throughout the retina, but cones are more tightly packed in the center (rods give good peripheral vision, cones give good acuity in the center of the visual field)



This figure shows how the three cone types are arranged in the fovea.

Contrast vision to smelling or hearing– you can increase the amount of area for receptor cells used for odor detection by folding more tissue into the nasal passages. Since you draw air into your nasal cavity you can make it pretty big and complex like a hounds. But you can’t do this with eyes– the eye has to be relatively exposed to the environment in order to allow light to enter. So you have a very limited amount of area available for photoreceptor cells.

The limited space on the retina means that there are trade-offs---- having more of one kind of cell means that you have fewer of another type of cell.



Human color vision


Similar to carotene, derived from Vitamin A (its about half of a vitamin A molecule). Lots of carbon bonds gives it lots of electrons which can be excited by a photon. When a photon hits it the bonds under go a trans-cis twist, changing the shape of the molecule (the photo excites electrons, making it easier for the bonds to rotate)



The chromophore is embedded in a large protein, the opsin molecule

Different cone opsins differ in the amino acids they have at key locations

These amino acid substitutions make them more sensitive to photons of different wavelengths (photons of different wavelengths have different amounts of energy– the amount of energy required to change the shape of a protein will depend on which amino acids are at key locations)



The opsin proteins are embedded in the membranes of the rods and cones

We can start with the rod photopigment, rhodopsin, and look at how many amino acid substitutions it would take to make a blue wavelength sensitive cone opsin

Then we can look at how many amino acid substitutions it would take to switch back and forth between a red (long wavelength sensitive cone opsin) and a green (medium wavelength sensitive cone opsin)

It looks like there aren’t many mutations required to go from a red to a green opsin

If you look above at the schematic diagram of the rods and cones, you will see that in the outer segments of rods the cell membrane folds in and creates disks. In the cones, the folds remain making multiple layers. The photopigment molecules reside in membranes of these disks and folds.




Two hypotheses: color vision evolved under natural selection for food choice

or color vision evolved under sexual selection for mate choice

Fruit is conspicuous in plants that want to attract frugivores

spread seeds– honest communication between plants and animals: the plant wants the animals to wait until the seed is mature before it is removed from the plant and taken to another location. The plant loads the fruit with sugars right at the best time for its seed to be transported by an animal, thus rewarding the animal for waiting until the ideal time.

Frugivorous birds have excellent color vision, and there was co-adaptation between birds and plants

Frugivorous primates could take advantage of this signaling with trichromatic color vision

Old world primates could also use their trichromatic color vision to tell the difference in leaf quality, and that may be an even bigger advantage

An alternative hypothesis about the evolution of trichromatic color vision in Old world primates is that it may not have evolved to improve foraging, but instead it may have evolved through sexual selection and then was put into use for foraging. If it evolved through sexual selection then we should see the appearance of photopigments and colorful fur or skin at about the same time. This hypothesis can be tested using molecular genetics

Trichromatic color vision evolved in Old World monkeys much earlier than red skin coloration, so it is unlikely that trichromacy evolved through sexual selection. This supports the idea that trichromacy probably evolved through natural selection for foraging on fruit and leaves.

Reprinted in Gerl and Morris 2008

Fig. 3 Ancestral State Reconstruction using maximum likelihood and the stored MK1 model (i.e., equal likelihood) implemented in Mesquite. Areas of pies indicate relative support for ancestral states. a) color vision; trichromatic color vision, indicated in black, was present at Node 2 before the evolution of b) red skin, indicated in black, and red pelage (not shown). Pie charts with asterisks indicate significant support for ancestral state reconstruction at that node.

Fernandez and Morris (2007) American Naturalist

Mandrillus sphinx (Linnaeus, 1758)

File: JPEG (29K); Photo: Carol Lofton, Distinctive Images


The Causes and Consequences of Color Vision

Ellen J. Gerl & Molly R. Morris

Published online: 2 October 2008

# Springer Science

Gerl and Morris2008

b. Sex-linked genes located on X chromosome: single plus sign LWS allele SER (557 nm), double plus sign LWS allele ALA (552 nm). The difference in the absorption frequency between the SER and ALA alleles allows females with both alleles (heterozygotes) to see more colors than males and females with trichromatic color vision (Deeb and Motulsky 1996)