Why Can't Flies See a Slowly Approaching Finger Despite Their Excellent Vision?

[fluidistic]

I've searched a bit in google about the vision of the common housefly but most studies are on the drosophilia.
First, am I right thinking that they see in 2 dimensions rather than 3? Because to see in 3 dimensions would require a somehow large distance between the extremum of their "eyes" and it's actually quite small compared to the separation between human eyes.

Second, I read that their vision is actually excellent in term of speed and precision. I guess that's basically why a fly can follow without any problem another fly that is zigzaging extremely fast for us. And apparently they can see the edges of objects really accurately, they have an extremely high resolution (can see better than a 1 mm thick object even 2 meters away from it http://phys.org/news129545593.html).

However I discovered by myself that I can actually approach a fly and even touch it without it seeing me. I would like to know what in their vision is failling in the following "test", or "set up":
let's say you are at about 40 cm away from a fly resting over a table. Point toward the fly with your index finger and go toward the fly slowly. It won't see it. You can go really close to it, and if you don't mind having a dirty finger you could just press the fly against the table. Obviously the fly doesn't see your finger getting bigger and bigger. Or should I say closer and closer. I have read nothing about this... So it doesn't seem a well known fact? I've tested this on many flies since I'm around 10 years of age.
So if they have such an amazing vision, how come they can't see a finger coming slowly to them?
I should mention that if you don't go toward the fly with your finger in a straight line, the fly will see you and get away as soon as you deviate from a straight line.

 

[256bits]

Try the same practice on yourself by focusing on the tip of the finger. Below a certain velocity threshold, you will also not be able to detect any movement of the finger towards you.

 

[fluidistic]

Try the same practice on yourself by focusing on the tip of the finger. Below a certain velocity threshold, you will also not be able to detect any movement of the finger towards you.

So basically that's because they see in 2 dimensions that they "miss" that an object is approaching them?
If I keep only 1 eye opened and I direct an index finger toward it I find it really hard to judge the distance between the approaching finger and my eye. Until my eye can't focus anymore (around 7.5 cm I think). When I lose the focus I know I'm close, but that's the only way I can really judge the distance.
I don't know what's the minimal distance a fly can focus but I think it's much smaller than 7.5 cm. That would explain why, if it really sees in 2 dimensions, it cannot really get information of how close an approaching object is getting, if it doesn't move laterally compared to the fly.

So if they see only in 2 dimensions... how can they judge distances? A human that has 1 eye blind can because he gets used to distances and used to sizes of common objects. However I don't think a fly learns the size of tables, cup of coffees, etc.
I would love to see a landing of a fly in slow motion. Does it "crash" into the object because it doesn't know it's that close to it?

Edit: I found a video (), the way they land makes me think they see in 2 dimensions because they fly in curved lines or pass over the object first, before landing. It seems their way to know the object is close. I wonder however how they would perform on a plane and clean surface like a table could be.

Edit 2: I just saw another video that reminded me they crash into windows. They probably cannot judge the distance between the window and them. And it's not because it's transparent: . If they see in 2 dimensions then I think it's normal. So am I right to think they see in 2 d?

 

[256bits]

well,I am not an expert on vision capability of organisms but I do suspect that your observation about an object moving laterally is the key. With lateral movement the image of the object and the surroundings presented to the eye has greater change than that from the object moving directly towards the eye,with the speed of the object being the same in both cases.
If you would take an object moving by distance x laterally across the eyes' viewport, the lateral movement x if above the threshold would be noticed. If on the other hand with a radial displacement x of the object towards or away from the eye, the object will have a size increase approximated by sin theta, with theta being the angle between the center of the object and its outer edge and the eye, the angle being small. ( By size I am referring to the radius of the object ) Obviously as the object gets closer the size increases at a faster rate.

at least for now you have a physical description of how lateral and radial movement of the object do differ.

Whether a 3D or 2D eye can help to discern this movement may well be part of the processing power of the brain, or of that with whatever a fly has been blessed with.

 

[Varadha]

Having got compound eye system, a common housefly should have better perception of 3-d than human beings, I suppose. As far as I've experimented I have better 3-d perception with two-eyes open than an eye closed. If that is the case, any organism with dozens of eyes should have better perceptional capabilities than binocular organism. Isn't it?

 

[fluidistic]

Having got compound eye system, a common housefly should have better perception of 3-d than human beings, I suppose. As far as I've experimented I have better 3-d perception with two-eyes open than an eye closed. If that is the case, any organism with dozens of eyes should have better perceptional capabilities than binocular organism. Isn't it?

You should not have a 3-d vision with 1 eye closed, that's why you feel you have a better "3-d perception with two-eyes open than an eye closed".
If I understood well Wikipedia (http://en.wikipedia.org/wiki/Stereopsis), no, an organism like a fly should not have a better 3d vision than us. That's because the separation between each of the eyes and ommatidia of the fly is much smaller than us, humans. To get past this, the housefly seems to need to fly in a curved way if it wants to have a better depth and thus a better estimation of distances of objects.

Edit: On the other hand, jumping spiders have a 3d vision, thanks to their frontal eyes I believe. It means they can estimate distances (close to them) without the need to move around. Just like us.

 

[DiracPool]

So if they have such an amazing vision, how come they can't see a finger coming slowly to them?
I should mention that if you don't go toward the fly with your finger in a straight line, the fly will see you and get away as soon as you deviate from a straight line.

The answer to your question is that flys do not have an apparatus to move their eyes in a saccade-like fashion like we (humans) do. Flys and many other insects compensate for this by flying back and forth in a saccade-like fashion. Check it out next time you see a fly buzzing around. Unless they are traveling somewhere, they tend to zig and zag around in a given region of space instead of just sitting there like a humming bird. The reason is that the visual-receptor neurons respond to motion and not just static images. So if they've landed on a surface and can't saccade around, it isn't surprising that they'd miss your finger.

 

[fluidistic]

Right, I've realized this during this thread, it's also showed in the video I linked. I also understand now how it's possible for jumping spiders to have a 3d vision despite lacking a "big" distance between their 2 biggest frontal eyes. I'm done with the thread, thanks guys.

 

[ChrisJA]

...
However I discovered by myself that I can actually approach a fly and even touch it without it seeing me. I would like to know what in their vision is failling in the following "test", or "set up":
let's say you are at about 40 cm away from a fly resting over a table. Point toward the fly with your index finger and go toward the fly slowly. It won't see it. You can go really close to it, and if you don't mind having a dirty finger you could just press the fly against the table. Obviously the fly doesn't see your finger getting bigger and bigger. Or should I say closer and closer. I have read nothing about this... So it doesn't seem a well known fact? I've tested this on many flies since I'm around 10 years of age...

I think they only see the the "directive" of there field of view. That is they only see changes. So in the case where they can follow another fly, all they see is the moving fly and the rest of their field of vision is blank. Flys only see change per delta t.

Humans are the same. We only see change but we continuously move our eyes, even if only small vibrations. This let's us detect edges and the brain fils in the rest. My gues is the fly's brain is not so big and never does the filling in part. So its eye aremotion detectors