Not Exactly Rocket Science

Blogging on Peer-Reviewed ResearchJellyfish may seem like simple blobs of goo, but some are surprisingly sophisticated. The box jellyfish (Tripedelia cystophora), for example, is a fast and active hunter and stalks its prey with the aid of 24 fully functioning eyes. These are grouped into four clusters called rhopalia, which lie on each side of its cube-like body. Together, they give the box jellyfish a complete 360 degree view of its world and make it highly manoeuvrable.

i-d0fe29fb742efbb11be5c475e116043e-Boxjellyfish.jpgEach eye cluster, four eyes are merely pits containing light-sensitive pigments, but two are remarkably advanced and carry their own lenses, retinas and corneas. The lenses are good enough to produce images that are free of distortion and even though the views are blurrier than those we see, these complex ‘camera-type’ eyes are very similar to those of more advanced animals like ourselves and other vertebrates.

But these similarities extend to a more fundamental level. Even though jellyfish are the most ancient group of animals to have a well developed visual system, it turns out that their eyes are built with many of the same genetic building blocks that ours are.

The eyes have it

All animal eyes, from the familiar human version to the compound eye of insects, contain two basic components. They have a photoreceptor – a cell that converts streams of light into chemical signals – and a dark pigment that focuses said streams. The photoreceptors always work through a partnership between a protein called an opsin and a chemical called retinal. When light strikes retinal, the molecule’s shape changes and it separates from opsin. That triggers a chemical signal that ends in an electrical impulse travelling to the brain.

Vertebrates and invertebrates differ in both the pigments and the photoreceptors they use, and both groups have their own distinctive opsins and signalling cascades. Zbynek Kozmik from the Academy of Sciences of the Czech Republic found that the box jellyfish is unusual in the structure of its photoreceptors are closer to those of the back-boned vertebrates than the spineless invertebrates.

i-b513ddf0357964b4ff20964d71da2677-Jellyfisheyee.jpgWhen Kozmik looked for box jellyfish genes that are involved in sight, he found that their opsin protein is also similar to the versions found in vertebrate eyes. With further tests, Kozmik confirmed that the jellyfish’s opsin is a fully functioning visual protein. It sticks to retinal and is particularly sensitive to blue-green light.

The similarities didn’t stop there. Kozmik found that the chain of proteins that carry the message passed on by opsin, are again similar to those used by vertebrates. And just as our eyes use the dark pigment melanin, so do those of the box jellyfish. Amid its genome, Kozmik found the jellyfish versions of human gene called Oca2 and Mitf that are essential for creating melanin. The genes are switched on in a part of the jellyfish eye that’s littered with granules of pigment, which were identified as melanin through chemical tests.

Parallel or conserved?

Despite the massive evolutionary gulf that separates jellyfish and vertebrates, both groups construct their eyes using similar genetic components. It’s possible that they kept an ancient ‘eye program’ that their shared ancestor already had, but Kozmik thinks that this is unlikely. If any such program existed, it would have eventually been abandoned by many animal groups, for most sighted invertebrates, such as octopuses and insects, build their eyes with a very different set of genes. Kozmik argues that eyes provide such an important advantage that there’s no obvious reason why any group of animal should abandon one working system of building them, in favour of a completely different one.

Instead, it’s more likely that jellyfish and vertebrates evolved their eyes by independently recruiting the same genetic building blocks, in a case of parallel evolution. That’s not unfeasible – there are other examples of large networks of genes being co-opted for new purposes, and computer models have estimated that it would only take about half a million generations to evolve a sophisticated camera-type eye from a simple patch of light-sensitive cells.

In fact, it’s likely that the jellyfish’s advanced camera-type eye evolved from the primitive cup-like versions that sit next to them on the rhophalia. These simpler eyes contain proteins called crystallins, which help to build the lenses of the advanced ones. And mitf gene which helps to produce melanin in the camera-type eyes is also active in the cup-like ones.

The eyes of the box jellyfish tell us yet again that important innovations, such as eyes, evolve by changing how existing groups of genes are used, rather than adding new ones to the mix.

Reference: PNAS doi:10.1073.pnas.0800388105

Images: jellyfish by Anders Garm; eye from PNAS



  1. #1 Alex
    June 17, 2008

    Are they getting anywhere on the question of whether the thing has enough brain capacity to make use of its imaging capability?

  2. #2 John BlueFoot
    June 17, 2008

    Cool, new meaning to the term “jelly fish eyes” LOL


  3. #3 blix
    June 17, 2008

    what conclusion can be drawn here:
    That the genetic foundation for the modern eye has been in place since before the evolution of the vertibrate line?
    Or that the Box Jelly is an unusual offshoot of the vertebrates that have re-evolved a radial body plan?
    Or that complex eye genetics have evolved independently in separate lines, yet have followed a (coincidentally) same path?

  4. #4 website design
    June 17, 2008

    Parallel evolution is some pretty interesting stuff. Organisms completely isolated from one another arriving at the pretty much the same evolutionary outcome. There could conceivably be organisms very similar to us walking around on another planet somewhere.

    Come to think of it, thats a good explanation of why all aliens on star trek look like humans with playdoh stuck randomly to their facial features. Or maybe its just because of thier budgets =P.

  5. #5 Jonah Johansen
    June 17, 2008


    So tough to talk about biology without accidentally revealing the forbidden truth.

    People routinely talk about how a system was designed to do a particular task, how intricate and efficient the design is.
    As soon as someone dares to speculate about the designer, the secular ‘thought police’ arrive and haul you off for not being “scientific”.

  6. #6 StuV
    June 17, 2008

    There is no designer, because there is no design.

    Nice try, though.

  7. #7 Ed Yong
    June 17, 2008

    Alex – they have no central nervous system, but one group suggested that the eyes may be specialised for specific tasks and could filter out complex visual info much earlier than, say, our visual system can.

    Blix – the authors favour your third suggestion (see end of my post)

    Jonah – sigh.

  8. #8 Eicos
    June 17, 2008

    “Box jellies, it seems, see a world of cyan and turquoise.”

    Actually, box jellies don’t “see” anything at all, as they have no brain. In fact, diploblasts in general have no brain, in the purest sense of the word. How they’re able to process information from a camera-type eye is a fascinating question.

  9. #9 Tom
    June 17, 2008

    If there is no design, Then, I cannot describe a computer model that evolves itself into an eye, without leaving behind impossibly large trial and error data. Put another way, If evolution is not a design, but a random set of biological conditions, then it should have left behind uncountable artifacts showing divergent attempts to produce anything from a eye, to a bone joint, to a leg, to a odd set of multi jointed creatures. To date, there are no three legged animals, three armed humans, etc. Nor are there creatures that have one eye, except in Mythology.
    So, what is it exactly that exists in a cellular creature that can link itself to its surroundings such that it is able to produce an eye in the very first instance? You tell me.

  10. #10 StuV
    June 17, 2008

    Genetic information is a recipe, not a blueprint. Evolution is not a “random set of biological conditions”. Please read up on it.

    So, what is it exactly that exists in a cellular creature that can link itself to its surroundings such that it is able to produce an eye in the very first instance?

    That’s not even a sentence. What are you asking?

  11. #11 Simon Cooke
    June 17, 2008

    Tom -
    There are creatures with one eye – humans. It results in immediate death on birth though. Happens with about 1 in 180,000 fetuses IIRC.

    The reason why you have two-eyed creatures (and two/four legged animals etc) is because of a strip of genes called the HOX which lay out the shape of the creature from one end of the spine to the other. You end up with one arm/leg on each side, because the HOX tells the fetus to develop with bilateral symmetry. So you always get one on each side, because the HOX pretty much just maps out one half of the creature, and lets symmetry handle the other half (because if you think about it, we’re basically long tubes with specialized appendages).

    As for the “where are the trial and error cases”, the ones that survive are 1 out of billions of attempts. Only the strong ones survive – you can see this if you look into genetic algorithms. And we do have a few examples of errors that persist – look at human eyes; they’re “designed” all wrong! The blood supply is in front of the retina, not behind in. The octopus, however, got it right.

    The whole design vs. evolution debate comes from the fact that humans are really ultra lousy at imagining random processes on this scale, and this kind of many billions of year time frame. We ascribe fate to everything, we have parts of our brain dedicated to religion (it’s very useful for a society to stick together if it’s hardwired). Saying that something’s random is very difficult for humans to do.

  12. #12 Eicos
    June 17, 2008

    IIRC, Simon, Hox genes are primarily involved in segment identity, although Hox derivatives and homeobox-containing genes have patterning functions on finer scales. The development of left-right symmetry has a separate mechanism, which is different across phyla.

  13. #13 Science Quiz
    June 17, 2008

    Great.Nice to hear both human and jelly fish have similar genetic blocks.Interesting article.Thanks for knowledge sharing.

  14. #14 Ed Yong
    June 17, 2008

    Eicos – good point about “seeing”. I’ve removed the offending sentence for accuracy.

  15. #15 Peter E Jones
    June 18, 2008

    Prof Simon Conway Morris has commented on this convergence, eg in his 2005 Boyle Memorial lecture,

    As to whether you need a brain to work with the vision. I once built a version of Machina Speculatrix that had quite complex behaviour. It had a light sensor, a bump detector, tricycle wheels with two motors and some capacitors etc. It would be attracted to a light, could be made to “fear” the light by poking it, after a while it would forget (ie a capacitor discharged slowly). Some others had it with a light that was on when searching and off when not. It could put in front of a mirror and it sort of jiggled in front as if admiring itself. Apparently groups of this “species” would congregate. This was all from a paperback by W.G. Walter of some 50 years ago. See

  16. #16 Ian
    June 18, 2008


    So how do you account for computers which write their own programs using the principles of the Theory of Evolution?

    Go read this and get a decent education:

    Ed – great blog!

  17. #17 blix
    June 18, 2008

    Isn’t it fun how Creationists always seem to infiltrate threads about evolution and natural selection? geesh..
    You’d think they’d come up with something besides a nonsensical catchphrase like “program=programmer”.
    creationists can cling to their ‘forbidden truths’ if they must, but I sure wish they’d not keep trying to hold back real scientific progress.

    A few points:
    There are few (no?) three legged creatures because there is not much benefit in it.. I suppose its possible with a radial body plan, but no takers yet. Not sure what you think that would prove though.
    There are a few one eyed creatures, although producing another eye gives stereovision, so that’s a much effective.

  18. #18 Marshall
    June 18, 2008

    It’s almost a shame that scientists routinely anthropomorphize to make it easier to understand for the public, because it gives the Creationists an easy way to pounce and say, “See! They admitted there’s a designer!”

    When we say “Giraffes’ necks are designed to reach leaves on higher trees,” we don’t mean they were designed. We mean that giraffes who happened to have longer necks fared better and managed to reproduce at a slightly higher rate than their shorter-necked counterparts. As a result, the number of long-neck giraffes increased.

    The same applies to developing eyes. Slowly but surely, eyes evolved–not because they wanted to be able to see better or because they were designed that way, but because some organisms happened to have beneficial mutations which would allow them to see better and–subsequently–to survive better and therefore out-reproduce their wild type counterparts.

  19. #19 Nagraj
    June 19, 2008

    its an interesting read the article from Kozmik and ur interpretation on that……..sometimes it surprises me how a protein share so much similarity in hugely diverse species……pax genes ,a class of prd class are present in primitive animals like cnetophores and sponges also……..but they dont do the same function……but in cnidarians they do meaning they gained some functions additionally… course of evolution pax remained conserved in eye dev…….but many species found slightly diff ways i think to make an eye…..but as u summed up same proteins are used in various ways…….which defines evolution……

  20. #20 JC
    January 7, 2009

    I am a Creationist, I am a believer in the Alpha and Omega. And at this point in my life I am very unsure of science. In a very “evolutionary” way, I see scientific study evolving towards sciences that make money. Perhaps its better to say -sciences and specific sides of sciences that GET money.
    I have even seen magnificent science shunned and suppressed for base and often selfish reasons, very often over money.
    I do not have a formal education, but I see the wisdom in the scientific way of thinking. And so, I think thus, and consider myself a humble scientist. That said I am always learning and prone to being wrong, I hope we all share the conclusion that everyone is wrong many times in their life.
    I will thank anyone who will show me I am wrong and show me what is correct.
    So with that lengthy intro, in that spirit of science, and without further delay.
    I have a question for the evolutionists.

    Have we found evidence of one animal changing into another? for example have we any animal remains that show a sequential change from one into the other?
    If that is too tight a question then fall back to a broader inquire, evolution within one kind of animal excluded, what facts do we have of animal to animal evolution?

    Some of you seem so sure of your conviction on evolution. Please share with me if you can. I seek to understand all things that are understandable. I admit to being ignorent of the fact I am asking about. having disregarded evolution at an early stage after finding it illogical. and have yet to find anything to convince me I was wrong in my assessment.