Mantis shrimp are are colorful little critters. Especially in their own eyes.
Animals are able to perceive color because the eyes contain different types of light-sensitive cells, or photoreceptors, each of which is most sensitive to a different part of the visible-light spectrum. Human eyes have three such photoreceptors, with a peak sensitivity to greenish, blueish, and reddish light. (There is also a fourth type of photoreceptor, which is used mostly for peripheral vision, and vision in darkness.) In other words, humans are trichromatic. The tri in trichromatic doesn’t mean that we perceive only three colors, but that all colors that we perceive can be reduced to a mixture of three colors (see also my post on color vision).
Most other mammals, as well as colorblind humans, have only two types of photoreceptors for color vision, and are therefore bichromatic. Most birds, on the other hand, are tetrachromatic (i.e. four photoreceptors for color vision), and therefore have a slightly more colorful visual palette than we do. But the variation between species is relatively small: Most animals have two to four types of photoreceptors for color vision. And there is good reason for this evolutionary agreement: Two to four photoreceptors are all that is needed to capture the colors that are actually present in the environment. Adding a fifth photoreceptor does relatively little to improve color vision.
Source: National Geographic
But the Mantis shrimp is a remarkable exception. This coral-reef-dwelling crustacean is endowed with 12 to 21 different types of photoreceptors! And the structure of their eyes is very peculiar as well. The upper and lower parts of the eye are typical compound-eye structures, quite similar to the eyes of most insects and other crustaceans. The remarkable part is the eye’s midband: Color vision is mediated by a horizontal band of photoreceptors in the middle of the eye. You can see this midband very clearly in the photo above.
Surely then, if their eyes contain so many different types of photoreceptors, Mantis shrimp must have excellent color vision, right? In a recent issue of Science, Thoen and colleagues set out to investigate exactly this. So how do you test color vision in a shrimp? Amazingly, you can do this in pretty much the same way as you can with humans: by having the shrimp perform a color-discrimination task.
Thoen and colleagues put two optical cables in the shrimp’s aquarium. These cables served as two light sources that could show different colors. The shrimp were trained with food rewards to grab the cable that had a specific color. Sometimes the color of the other cable was very different, and the color-discrimination task was really easy. At other times the color was very similar, and the task was really difficult. This simple task allowed the researchers to determine how well the shrimp can distinguish colors. Quite surprisingly, it turns out that Mantis shrimp aren’t very good at color discrimination at all. Their performance on the task dropped to chance level when the two colors differed by about 12 nm. For comparison, humans, with only three photoreceptors for color vision, are able to distinguish color differences as small as 1 nm.
So how come that Mantis shrimp have so many different photoreceptors, but are relatively poor at discriminating colors? A trivial explanation comes to mind first, of course. Shrimp are not terribly intelligent and might simply fail to respond correctly when the task gets a bit difficult: They may get confused easily. But Thoen and colleagues did some computational modeling that does a reasonable job of arguing against this trivial explanation. (I won’t bore you with the details here.)
Another explanation might be that Mantis shrimp are not better at discriminating colors, but are better at distinguishing mixtures of colors. For human eyes, pure colors are indistinguishable from mixed colors. That’s why a computer monitor can show all the colors in the (or, rather, our) world, even though it emits only red, green, and blue light: We perceive a mixture of blue (±450 nm) and yellow light (±575 nm) as green, and cannot tell it apart from actual green light (±500 nm). Perhaps Mantis shrimp are better at this, and would in many cases be able to tell the difference between color mixtures and pure colors.
A third explanation, which is favored by Thoen and colleagues, is that shrimp don’t discriminate colors, but ‘recognize’ them. This sounds a bit esoteric, and frankly the authors don’t explain their point very clearly. But I think they mean something like the following: When Mantis shrimp look at something, they slowly moves their eyes up and down, in a scanning-like motion. This way, the eye’s horizontally oriented midband gets a good look at the entire object. Thoen and colleagues suggest that during these scans, each photoreceptor creates a temporal signature of the object. For example, if it were scanning a tree in a top-down fashion, a green-responsive photoreceptor would initially respond strongly (when scanning the green leaves), and then stop responding (when scanning the brown trunk). The shrimp could then recognize objects by the combined temporal signatures of all photoreceptors. Put more simply, Mantis shrimp might recognize an object somewhat like a blind person reads braille, by visually stroking it with 12 to 21 fingers (i.e. photoreceptor types) at the same time, each ‘feeling’ a different color.
This is an intriguing suggestion, but there are still many unanswered questions. For example, it is unclear to me how the fact that Mantis shrimp may use ‘visual stroking’ to recognize objects explains their relatively poor color discrimination. Why does the one logically follow from the other? Thoen and colleagues don’t explain this very clearly. (To be fair, they have little space to do so in a luxury journal like Science.) It would also be very interesting to see how well Mantis shrimp perform on a color discrimination task with mixed colors. Would they be able to tell a blue-yellow mixture apart from green? Fascinating stuff!