How far can ants see?

Cecil's Columns/Staff Reports
1

Oh, come on, the answer of “ants see as well as digital cameras” is not believable. Where is the evidence that ants see one “pixel” per eye? Where is the evidence that ants and other crustaceans have “only a few thousand pixels resolution”? Have no studies been done about this - oh, I don’t know, look at the flight of bees. We know bees see patterns in flowers that we do not (ultraviolet or infrared - I’ve forgotten which) - so see how far away they can distinguish different flowers. Give us something other than guesswork involving digital cameras.

Andrius Tamulis

2

How far can ants see?

3

Welcome to the Straight Dope Message Board. It’s customary to post a link to the Staff Report you are commenting on, so that others can be on the same page, so to speak.

The column you’re working from can be found here: How far can ants see?

I think Doug gave a fairly good answer to a fairly weak question. It seems reasonable to me that each element of a compound eye resolves only one point of light or dark (or maybe color). In this sense, there is no such thing as focus in such an eye.

Bees, I would think, fall into a visual category with the dragonflies and horse flies that Doug brought up, with a much higher resolution than ants, due to a higher number of elements in the compound eye.

I’m sure Doug will show up to defend his conclusions on his own, but I trust what he said. If I recall correctly, he is an entomologist by profession.

4

don’t know much about ant physiology (hey that would make a good song lyric). But we can make some guesses and see where the physics takes us. For example, I conclude that for an ant to be able to “see” another ant that tant would could be no more than about 1.5 inches away; otherwise it would be a shapeless blur.
I would guess an ant eye cluster is probably about less than a millimeter in diameter, probably more like a tenth of a millimeter on those tini-tiny ants. And if we figure this is subdivided into a thousand independent eye-lets then each of these would be about 3 to 10 microns in diameter. Now as you may know the bigger the telescope or eye the smaller a spot it can resolve at a distance. Also the smaller an “eye” is the easier it is to build a nearly perfect one in practice, so ant eyes may be nearly perfect. We can put some numbers to this. Assuming ant vision works in the visible like most eyes then the average wavelength of light is 0.5 microns. For a perfect lens with 1 diameter of 5 microns then the smallest resolvable spot at 1 a distance of a distance of 1 centimeter is:
1.2*(10000/5)*0.5 = 1200 microns = 1.2 mililmeter.

In otherwords one ant-eyelet if equiped with the best possible lenslet would see anything smaller than 1.2 milimeters at a distance of 1 cm from the ant to be a blur. this ratio (1.2: 10) is maintained at further distances. so the blur size at 1 meter (39 inches) would be 12 centimeters (anout 3 inches). another ant 0.5 cm would be blurred at anything beyond about 4 cm distance (1.5 inches).
so at a distance of 1 meter a tasty maple leaf would be totally blurred out.
as you can see from this analyis there would be no need for each eyelet to see more than one pixels worth of an image since the blur size is so large. Now there are some tricks you can pull to enhance the resolution of a blurred image but these rarely gain you a lot, rarely more than a factor of two and certainly for this compound ant-eyy no more than a factor of ten would enven be theoretically possible.

well hows that for an better answer.

5

What is a pixel? An individual picture image, basically a digital point of light or dark, the equivalent of on/off in visual pickup.

Eace facet of an insect’s eye functions to pick up light or dark, on or off. No focus, no lens. (Sorry, I cannot “prove” that. No, I don’t have any evidence.)

Stack a handful together and you get an image composed of dots. Add more dots for a given area, better resolution.

Ants don’t have many dots compared to, say, bees. Ant’s can’t discern as much detail as bees.

Perhaps certain facets pick up different wavelengths. Thus variations in color. At sacrifice of detail.

For a given image made of a given number of pixels and a given area, the farther the object viewed, the more likely details disappear into the graininess of the image. This is the effect seen in digital images from spacecraft. It’s called resolution.

6

Mor of a pedant than an ant ranter. I am just a simple caveman. I know little of your formic formulae or optic equations. But I do know one thing. Ants have been around a long time and thus have had ample time to evolve a form of vision which likely satisfies what they need to do. Sure those chemicals help when Cecil is picnicking nearby, but I would bet Candian dollars to your American ones they could see a lot longer thsn an inch and a half on flat sand.

7

As others have pointed out, one dot of color per ommatidium (the elements of a compound eye) is what we’re dealing with, so the analogy to a digital camera image and pixels is quite adequate. Each ommatidium is a long cylinder with pigment cells of various types along its length (including, in some insects, pigments that respond to UV), so some of them do have color vision - but there’s no retina, so no projected image, no focal distance. What we don’t and can NEVER know is how an insect’s brain integrates that information. All we have to work on is analogies, since our own visual system and optics are so utterly different.

8

Really, now much resolution to ants need? They get around the world largely by going up and over obstacles rather than avoiding them. They’re so light that a fall doesn’t really hurt them. They just get up and start running again. They are guided toward desirable things (food, other ants, home) much more by their sense of smell than by sight.

But before we pooh-pooh ant vision too much, remember that some ants can perceive the polarization of light–something we can’t do–so they can navigate by using the position of the Sun in the sky.

9

The camera analogy is slightly flawed. Becaue an ant can move his head the size of the scene he can see (or the detail therein) is not limted to a single frame image with a certain number of pixels. To make this clear, imagine the ant had only a single eye-let instead of a bundle. He could simply scan that one eyelet around and look with it in each direction to build up the image one pixel at a time, just as for example, a person with a telescope or binoculars swivels them across the horizon looking at one area at a time.

I would speculate that the reason ants have fewer eyelets than say bees is that bees generally dont fly after dark whereas ants I beleive do operate at night. Perhaps to aid light collection ant-eyelets must be larger and hence fewer in number for the same size eye.

Furthermore, since an ant has the luxury of standing still and turning its head to effectively increace the number of pixels it can take in, it may not need to have as many in the first place.

10

Turning the “camera” does not increase resolution. It increases the area being inspected, but the detail level it can pick up is the same.

Offset views may provide a different vantage, and thus allow better resolution - if combined properly. But the offset would need to be more significant than the distance between the ant’s eyes. Say a couple feet. (That last is a guess on my part.)

Having larger facets may allow the ant to have more light sensitivity. Then again, do bees use their eyes in their hives? (I don’t know. I assume so, but we all know what assumptions do.)

11

Right, so were agreed then, having more eyelets or pixels does not mean you “see better” it simply means you see more at one glance. In fact due to diffraction limited resolution the more eyelets you have per eye the worse your far field vision. Hence if bee and ant eyes are the same size then bees ought to see worse at distance since each eyelet is smaller. thier bigger eyes merely let them see more at a glance. Ants can make up for their smaller number of pixels by scanning their heads.

on the otherhand, for near field vision (objects close on the scale of the size of your eyes) then scanning your head actually can increase the resolution to sizes finer than a pixel size. Indeed this is the way the scanner hooked to your computer functions.

12

bigkahuna said: "Right, so were agreed then, having more eyelets or pixels does not mean you “see better”

NO. We most definitely are NOT agreed here. The insects that see best have by far the greatest numbers of smaller ommatidia, without exception. Day-flying predatory insects (such as dragonflies, mantises, robber flies, hunting wasps, etc.), whose lives are utterly dependent upon their vision, have the most and smallest ommatidia. Bees are also totally dependent on their vision, and have well-developed eyes, and can certainly see better than any ants at ANY distance (we’ve done more than enough experiments with bee vision to know this, trust me). There are only two or three ant genera in existence that have eyes as complex as even the simplest bee eyes, and those ants are day-active predators. Male insects that locate females visually instead of by smell or guard territories almost invariably have hyperdeveloped eyes, sometimes making up almost the entire head (as in honeybee drones). Show an entomologist the head of an insect, and we can tell almost without fail whether it is active at day or night, airborne or ground-dwelling (or aquatic), predatory or herbivorous, male or female - just from eye structure. Millions of years of evolution tend to fine-tune things.

13

Okay so doug Y says that there is an observable phenomenological correlation between eye characterisitics and some sort of unspecified vision testing done on insects. Fine. What I am saying is what the physics of insect eyes tells us. The physics gives us upper bounds on how good an insect could possibly see. Not how good it actually sees. That is, given a certain size telesope I can tell you which planets you could inprinicple see but you might have a crappy telescope.

so to reiterate:
the diameter of you eye determines

  1. the smallest size object you can see at a given distance. the larger the eye the smaller the object.
  2. the larger the eye the more light it gathers and thus the better potential night vision.

the more pixels in your eye the larger the scene one can see at one glance. But you can always turn your head to see a larger scene. But then you have the problem of having to “store” the parts of the scene you aren’t looking at in memory.

If you disallow advanced signal processing then for far field viewing the smaller the pixels the worse the vision. Thus the more pixels for a given size eye the worse the far field vision without advanced signal processing. With advanced signal processing this may not matter–were just back to the total size of the eye.

14

So, bigk, you’re sayin’ it’s unlikely that an ant colony would ever develop an astronomy department?

15

After that digression I return to my point.
It would seem to me that if my estimates of eye-sizes aren’t too far off that bugs dont see very well. and that an ant is basically blind beyond a few inches to anything smaller than the size of another ant. These are upper bounds: the ant might in fact see worse than this but it could not see better.

thoughts on ant-stronomy. Like ants people, even ones with big telescopes, dont resolve very distand stars. they are just points of light to us and to ants. Hence I suppose Antsronomy is not ruled out. Dont moths use the moon?

16

It is an established fact that humans have some kind of “advanced signal processing.” We demonstrably “see” more pixels than we have.

It’s hard to tell with bugs. They can’t even tell you which way the elephant is standing.

17

bigkahuna, I don’t think I agree. The size of the aperature is the light gathering capability. A bigger telescope will pick up more light, thus fainter objects.

The number of pixels determine resolution, the size distinction that can be made of that object. Of course distance will affect size distinction. Turning the telescope lets you see around more, get a larger field of view, but it does not let you get better resolution. I would think this would mean for bug eyes that big facets (ommatidia?) would have more light gathering ability, and thus would be better for seeing at night or in tunnels, but having large eyes per se does not mean you have better resolution. For telescope work, it means you can see farther into space because you can pick up fainter objects, but for more detail, you need magnification to get more resolution.

I think what you mean by signal processing is like the VLA telescope that uses an array of smaller telescopes and processes the signals to make a larger image/better resolution. I’m not sure how bug eyes work relating to this - I don’t really know how human eyes work on this, either.

For bugs, lots of facets, even tiny facets that pack a whole bunch more into a smaller area, mean better resolution, better ability to define shapes and identify objects at a distance and while moving. Big facets - even a relatively few number - means better light sensitivity.

This is true if the pixels are all the same size, so you are increasing the area of the image reception. However, if the eye size is constant, than the more pixels would be smaller pixels, they would pick up better resolution but less sensitivity to faint objects.

Sort of like those new computer generated pictures that use small pictures to build a large composite. If you stand up close, you see all the small images and the overall image is indistinct. Stand back a bit, the big image becomes clear but the small images are not identifiable. Of course if those small images were not images in themselves, but rather a single homogeneous tone, then you get the idea.

18

Irishman,
if it weren’t for diffraction you would be right. But in this case its dffraction that is OH SO IMPORTANT!

its counter intuitive but the smaller an eye or lens gets the poorer its theoretical resolution. This is not to say that bigger is always better. A poorly constructed big telescope might not do as well as a well constructed little telescope. And for example, a hawks eye is proably better built than an elephant’s eye. but the fact is that if both are working at their theoretical best then the digger the diameter the better the resolution.

(this turns out to be exactly analogous to the uncertainty principle. larger aperature lens => tighter focus.)

resolution is thus determined by aperature size and lens quality (or what passes for a lens on an ant).

you could build the worlds best telescope using a single pixel as follows: first find the biggest best lens or mirror, then place a single pixel at its focus (image plane). Now scan the telescope through the heavens to build up as large an image as you want, noting the light intensity as you scan the telescope. If you had more pixels available you could not improve the image quality. you could however collect the image more quickly since you could look in multiple directions at once. And like you said, gather more light in a shorter amount of time. On the other hand if time is not an issue I can gather as much light as I want, regardless of the telescope size simply by waiting longer (scanning more slowly).

I think what is confusing people here is the notion that there are two sorts of pixels were dealing with here. One is the fundamental blur spot size of the imaging system. this is what I’m talking about with regard to aprerture sizes limiting resolution. for a given aperature size and focal length there is a minumum blur spot size that cannot be resolved any finer. that is even if your detector had many actual pixels covering this blur region you could not resolve it any finer. The number of pixels in the detector does not set the limit, the limit is set independent of the detection method.

19

for a discussion of how aperature size is related to resolution see:
http://astrosun.tn.cornell.edu/courses/astro201/diff_limit.htm
or
http://www.faqs.org/faqs/astronomy/faq/part2/section-6.html
or
http://128.252.223.112:80/cgi-bin/circR?/posts/archives/may97/864446241.Ph.r.html
or
http://www.astro.soton.ac.uk/~trm/PH112/notes/notes/node15.html
or
http://gulliver.gps.caltech.edu/Thesis_Chapter_2/Introduction.html

note: you can beat the diffraction limit in the near field. for practical purposes you can consider the near filed to be any object closer to you than the diameter of your eye. In the far field resolution is set by diffraction.

the formula for far filed resolution is:
approximately:
B = W*L/D
where
D = diameter of Eye
L = distance to object
B = smallest size of object that can be resolved
W=wavelength of light (approx 0.5 microns)

so for if an ant eye is 100 microns in diameter ( tenth of a millimeter) then at distance of 2 inches the blur-spot is:
B= L/200 = 1/100th inch
but if an ant eyelet is only 3 microns (assuming 1000 eyelets per eye) the blur spor increaces to:
B = L/6 = 1/3th inch or about the si ze of an ant.

at a range of 20 inches these numbers scale to:
1/10th inch and 3 inches respectively.

so depnding on which eye size you want to go with, those are the smallest possible object sizes an ant can see at 2 and 20 inches.

My belief is that the ant is limited by the eyelet size.

20

bigkahuna wrote:

“the formula for far filed resolution is:
approximately:
B = W*L/D where
D = diameter of Eye
L = distance to object
B = smallest size of object that can be resolved
W=wavelength of light (approx 0.5 microns)”

This is why you’re wrong about ant eyes. THEY DON’T OPERATE USING THESE PRINCIPLES!! There is no aperture, no focus, no lens. You CANNOT apply optical physics to ant eyes! For the best analogy, just imagine removing your lens and iris, and placing your retina on the outer surface of your cornea - then give each single cell in your retina an individual transparent hexagonal covering, slightly convex. THAT is how an insect eye is constructed. Each ommatidium has ONE RECEPTOR CELL for each color: no focus, no aperture. Distance from the object viewed is irrelevant when there is no aperture. Like Irishman said:

“For bugs, lots of facets, even tiny facets that pack a whole bunch more into a smaller area, mean better resolution, better ability to define shapes and identify objects at a distance and while moving. Big facets - even a relatively few number - means better light sensitivity.”

The eye area and visual field of a nocturnal longhorn beetle and a honeybee may be the same, but the honeybee has over 100 times as many ommatidia, meaning over 100 times as many cells in its “retina”, meaning 100 times as many “pixels” for the same exact visual field. This is the very definition of higher resolution. If my computer screen had only 102 x 77 pixels, instead of 1024 x 768, that would be pretty crappy resolution, and 10 x 7 would be even worse.