Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Input/processing aspects of adaptive camouflage

  1. Oct 22, 2016 #1
    So, I was reading this comic book...

    ... not the most promising start for a science thread, I know, but I promise it's getting better from here!

    I'm going to include a little background on the work in question in an addendum, but get straight to the point for now: The book features a group of ceratosaurs, and, I presume to make them more visually interesting, the author decided to give them an adaptive camouflage capability, as found in some present-day animals, of which the chameleon is probably the best-known example. As far as I'm aware, there is no paleontological basis for this species having such an ability - but that's neither here nor there, as far as this thread is concerned.

    Below, a selection of panels showcasing the extent to which the book takes this:

    ar-123-c99.jpg ar-124-c99.jpg

    - The bottom panel shows one member of the group (dubbed "Big Nose" in the dramatis personae) in its un-camouflaged state, I think.

    - On the top left, we have the lot of them adapted to a uniformly coloured terrain. So far, so unspectacular.

    - On the top right, we have them adapting to the sauropod carcass they were approaching in the previous panel. In addition to colour matching, we now have pattern matching, and not just in general terms, but in the sense that individual features of the carcass's skin are continued across the bodies of the scavengers. That seems a lot more remarkable to me.

    - In the panoramic panel below that, we're getting into "Where's Waldo" territory (if the fifth member of the group is in it, I've not found it yet). More to the point, even the animals which are out in the open are well-camouflaged, by again continuing individual features of the backdrop across their bodies. There is an important difference from the earlier case, though: This effect now partly depends on the position of the observer. If you mentally move the "camera angle", the brown stripes cease to line up with the tree trunks they're meant to match, et cetera.

    - The next panel down illustrates this more simply and clearly: The lower half of the body matches the trunk in the foreground, the upper half the greenery in the background, so that no part breaks the outline of said trunk, as long as the observer remains in the same horizontal plane as the dinosaur.

    My question, as already implied, isn't about this species of dinosaur. It's how much of this is plausible, in a general sense, and, if plausible, how it works, based on what we know of this ability in present-day animals. By "how it works", I don't mean how an animal's skin can change colour, that part seems relatively straightforward. I mean how does the animal "decide" which colour to use where to achieve the desired effect.

    One of the "James Bond" movies features adaptively-camouflaging cars, and the idea behind that one was also relatively straightforward, IIRC: Put cameras in various spots on the surface, and display whatever those detect at a spot on the opposite side. The higher the resolution (the more cameras and displays), the better the result. The biological equivalent of cameras are eyes, though, and most animals, including those who practice adaptive camouflage, only have two of those, so that can't be how they do it. (Or can it?)

    The most remarkable example of pattern-matching I came across during my preliminary research is this picture of a peacock flounder in a tank with a checkered floor:


    It's not perfect, but clearly this goes beyond merely reproducing the pattern in a generic sense, as would usually suffice for a naturally-occurring seafloor. Some of the squares are directly continued onto the body (the middle one along the bottom-left edge is especially well-done), and the top-left-to-bottom-right diagonal going right along the long axis shows an astonishingly close match to what's underneath - the shapes themselves, the sizes, the orientation, the number, everything is roughly right. Wow. Having seen this, I have no problem accepting the comic's carcass sequence as plausible.

    So, how do you imagine it works? The wikipedia article mentions that "if one of the flounder's eyes is damaged or covered[, it has] difficulties in matching [its] surroundings", which confirms the role of vision as providing the input. How much processing is needed to translate that input into an output (a set of selections of particular colours for particular portions of the animal's body)? To put it in another way, if humans had this ability, would this be more akin to, say, moving one's hand to pick up something one sees lying on a table - something anyone can do without thinking about it - or more akin to painting a landscape - something that involves both thought and skill?

    However much processing the checkerboard example involves, the observer-specific examples depicted in the latter panels would surely involve several orders of magnitude more. Anyone know of any real-life examples there? Cephalopods show some extraordinary abilities in a different kind of camouflage, that of masquerading as a different animal, which I imagine also involves a great deal of thought and skill, so that seems like one promising place to look.

    Right, I think that's all I have so far. Looking forward to your contributions, however speculative! :)

    Addendum: The panels are from the 1996 volume "The Hunt" of the irregularly published title "Age of Reptiles". It's written and drawn by Ricardo Delgado and coloured by James Sinclair (meaning that both creators had a hand in how the camouflage ends up looking on the page), and published by Dark Horse. The stories feature nothing but non-anthropomorphic dinosaurs in their natural habitat, so there's no text and only a very rudimentary plot; most of the storytelling consists simply of portraying animals acting the way (the author images) such animals do.
  2. jcsd
  3. Oct 22, 2016 #2
  4. Oct 23, 2016 #3


    User Avatar
    Science Advisor

    I have been interested in this subject for a while.
    Your query involves several issues:

    Cloaking Device-like Camouflage
    Camouflaged item appears like what is on its opposite side (in a particular line of sight). Ideally the light from behinds bends around the object so it looks like you are seeing through to (but you are not seeing through it, the light is bending around it). Fictional examples include Romulin cloaking devices (gratuitous StarTech reference), the Predator monster, or Harry Potter's invisibility cloak. There is actually research on this (see the metamaterial research section), but as I recall its limited to small objects (mm or smaller?) and only some wavelengths. I know of no biological entities that do this.

    A similar effect could be obtained from transparent organisms. There are a lot of these. They tend to be small and aquatic (probably closer match of optical density to reduce lensing). For example, baby zebrafish (≤ 1 day post fertilization) are have not yet developed pigment cells. Larger organisms can also be transparent, either naturally or from combining several mutations. Len Zon's lab (zebrafish in Boston) made a transparent zebrafish by combining 3 or 4 mutations. You could read newsprint through it, but that font is about 1/2 the height of the fish. Biologists love this kind of stuff because they can watch what's going on in a live organism without cutting it open.

    Both of these would allow the camouflaged object to display different looks based what is behind it from differing points of view.

    Normal Biological Coloring, Active and Passive
    Passive coloring would be largely unchanging.
    Active biological camouflage usually involves pigment cells moving their pigment granules so that they either fill the large flat cell (having a large visual effect) or are concentrated into a small spot (minimizing their visual impact). This movement may be (in different cases) under neuronal and/or hormonal control. In some organisms the changes can be quite rapid.

    Different color pigments can reside in different layers and combine to produce a variety of effects. Pigments absorb or reflect different wavelengths. On the other hand, structural biological colors are based on iridescence interactions. These are not pigments which can be chemically isolated and retain the color found in the organism. Structural colors reflect light from multiple reflective layers in the camouflaged object, resulting in certain colors based on constructive and destructive interference from the reflecting layers. The reflecting layer separation is around the scale of the light's wavelength (or multiples of it). The inference patterns change based on the angles of illumination and observation because of the wavelength offset distance between the two interfering lightpaths changes. Some structural colors are based things like on repeated hairs in features (perhaps with special chemicals to make them more optically reflective). Many are based on small guanine crystals (flats sheets) usually oriented parallel to the organisms surface inside of pigment cells called iridophores. These can also be actively modulated. Feathers can be moved behind other feathers to hide them, and guanine granules can be moved like pigment granules in some iridophores, concentrating them in a point or spreading them throughout the cell. In an even more extreme case, some chameleon iridophores can adjust the spacing of their reflecting surfaces thus changing the interference patterns and the colors displayed.

    Pattern Matching
    This involves the sensory perception of the pattern and its replication on the animal's surface, as well as any size adjustments (of pattern vs. body) that might be necessary as the animal grows in size. My understanding of this is limited.

    Perceiving the pattern would seem to require visual system input, however, with the discovery extra https://www.researchgate.net/publication/280483039_The_Evolution_of_Non-visual_Photopigments_in_the_Central_Nervous_System_of_Vertebrates[/URL], sensors in the skin could conceivably obtain detailed pattern information from the surface it lies upon (given enough light). (This is a conjecture on my part.)

    Pattern matching can be exact matching to background objects or more like creating a texture (pattern) on the animal's surface with the right colors, size of color patches etc. Exact matches would require both making the pattern as well as arranging it on the body to meet up with the pattern in the environment. This seems rare to me. The pictured flatfish comes close. Creating a [URL='https://www.google.com/search?q=pattern+matching+in+biological+camouflage&safe=off&source=lnms&tbm=isch&sa=X&ved=0ahUKEwjM3t_bvvHPAhVV62MKHQVLBAMQ_AUICCgB&biw=1726&bih=950#imgrc=_']surface texture[/URL] seems to be the most common approach and works really well in some cases.
    Last edited by a moderator: May 8, 2017
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted

Similar Discussions: Input/processing aspects of adaptive camouflage