Fiddler crabs may not see very well, but they make up for that lack with an incredible ability to statically determine whether a swooping creature is a predator or a harmless bug. The process these crabs use to decide if something is a threat is so complex that scientists are even thinking of using it to improve robotic vision systems. Sounds like these crabs are smarter than robots! — Global Animal
Discovery News, Anna Salleh
Fiddler crabs can’t see very well, so they have to use statistical calculations to distinguish between swooping predators and harmless passing insects, say researchers.
Jan Hemmi from the ARC Centre of Excellence for Vision Science at the Australian National University and colleagues report their findings online in the Proceedings of the Royal Society B.
“To our surprise we found that the crabs are actually able to distinguish between predators that are real and those that are not,” says Hemmi. “By using the statistics of how objects move, they seem to be able to discriminate.”
Fiddler crabs live in very open intertidal environments of northern Australia.
“They are very exposed to predation,” says Hemmi. “In any case of danger they retreat to their burrows and disappear under ground.”
To keep a look out for predators, the crabs use two eyes that sit two centimeters (less than an inch) above ground at the end of long stalks.
Each eye is capable of seeing virtually the entire hemisphere above the crab, but is limited in terms of the detail of what it sees.
For example, says Hemmi, the crab’s stereoscopic vision is limited to within 14 centimeters (5.5 inches), so they can’t judge longer distances.
The crab’s eyes also have very low resolution — the equivalent of 8,000 pixels in each eye, and an approaching bird is likely to be detected by just a single ‘pixel,’ he says.
While the crabs can distinguish brightness they are not good at distinguishing the shape of flying objects.
“Basically they see a spot flying in the sky,” says Hemmi.
Hemmi says his previous research, which used model predators, suggests the crabs find it very difficult to tell between moving things that are a threat and those that aren’t.
The crabs were thought to live in something akin to a permanent state of paranoia, darting underground at the fluttering of a harmless butterfly or a close-buzzing fly.
In their latest research, Hemmi and colleagues collected hours of real-time field footage of crabs interacting with actual predators and found this was not the case.
They found the crabs were more likely to run away from predatory terns, than they were from non-predatory birds or harmless dragonflies.
The researchers analyzed the movement of predators and other flying animals, such as their apparent height, size, speed and their contrast against the sky, to identify the visual cues that crabs could use.
Hemmi says while individual visual clues might not on their own signal alarm, the crabs appeared to take into account the statistical likelihood of predators moving in a certain way — with a particular combination of visual cues.
“They must use a combination of those cues and say okay, if that happens and that happens at the same time then I’ll run away,” he says. “Something that moves fast, flickers a lot and is low on the horizon — that’s what they took as dangerous.”
A “flicker” occurs when a bird flaps its wings and the crab sees alternately the shaded underside of the bird and then its more brightly lit wings.
This is significant because a bird flapping its wings is more likely to be actively maneuvering than just gliding by on its way somewhere else.
Hemmi is not sure why the crabs would respond more to something that appears low on the horizon, but he says it could help them distinguish between harmless insects and potentially harmful birds.
Because insects are so small, crabs would most likely not be able to see them until they fly right over them, whereas they are more likely to spot a bird coming for further away, appearing lower on the horizon.
The researchers now intend to study crab sensitivity to different colors and light polarization to get a better model for their visual abilities.
They say research of this kind could be used to help improve robotic vision systems.