# Why can't you flatten a sphere?

It's a well-known fact that you can't flatten a sphere without tearing or deforming it. How can I explain why this is so to a 10 year old?

As soon as an explanation starts using terms like "Gaussian curvature", it's going too far for the audience in question.

A great explanation would work just as well for a hemisphere as for a sphere, as both have positive curvature.

-
This is a balloon. Suppose it is a sphere and that you can flatten it... – Git Gud Mar 29 '13 at 18:16
You need to be clear about what the rules are. It is obvious to a child that if the air goes out of a balloon then you have a flattened sphere. No corners seems a bit artificial to a child, I imagine. Maybe appeal to the notion of invertibility as in if you really flattened the balloon, then the top and bottom would be indistinguishable and you could not reverse the process? – copper.hat Mar 29 '13 at 18:31
Because you are not a Decepticons drone. – Asaf Karagila Mar 29 '13 at 18:38
What is meant here by "flatten a sphere"? E.g., you can flatten a hemisphere in the topological, but not in a metrical sense. On the other hand, you cannot map a full sphere on a piece of the plane without folds and overlaps. – Christian Blatter Mar 29 '13 at 19:23
Because if you could, all the flowers would die. – Will Jagy Mar 29 '13 at 23:20

I take it that you will be explaining this to a group of kids. Why don't you try several approaches? Some will probably appreciate the orange peel explanation simply because it gives them something to do. Others may appreciate the following. Mathematically it is based on the fact that the circumference of a circle on a spherical surface is shorter than the circumference of a circle (with same radius) on a flat plane (see also rschwieb's suggestion of calculating the ratio of the length of the equator to the distance to a pole). So changing the radius by a fixed amount will not result in a fixed change in the circumferemce.

I suspect some of the kids in your audience will have seen just that while crocheting. If you are crocheting a circular piece (like this), then you have to be careful to add just the right number of links per round. Add too few, and the end result will have a positive gaussian curvature (a bag-like shape) - add too many and you get a warped surface of a pseudosphere (was still good enough, when yours truly handed such a potholder to his dear Grandma 40+ years ago).

Undoubtedly you will have shown how to flatten a cylindrical surface prior to this (also suggested by rschwieb).

Also make them gift wrap both a cylindrical tube and an inflated basketball.

-
+1: Love the gift wrapping idea! – Peter Majeed Mar 29 '13 at 23:23

Experimentation is fun!

Give him a globe, some string, a ruler, a compass, some paper, and a pencil. Then have him try to map 4 different cities from the globe onto the paper, keeping ALL the distances between them the same as on the globe.

-
This is a great idea for an experiment! – Joe Mar 29 '13 at 21:57

Show him a triangle on the sphere with 3 right angles. He knows that it is impossible in a plane right ?

-
I'm not sure I'm convinced by this answer. You can "partially flatten" the sphere so that the triangle sits flat on the table, but some of the sphere is still above the table. Of course the triangle will no longer have right angles, but you can expect that from the change of geometry anyway. While the flattening of the sphere is "global", this only demonstrates something "local". (It's possible I didn't understand what's meant by "without deforming it". Is this allowed?) – Yoni Rozenshein Mar 29 '13 at 18:44
I liked this. It starts with an intuitive fact, the 90 degree corners when drawn on the sphere, but to flatten, they must deform. +1 – JoeTaxpayer Mar 30 '13 at 1:56

One simple way is to think of an orange peel. Have the 10-year old peel an orange nicely and tell him to try to flatten the peel pieces completely without breaking or streching any part of the peel.

-
This shows him that the sphere can't be flattened. Now he wants to know why. – Joe Mar 29 '13 at 18:54
@Joe: that's more about psychology, I guess - but why not? – Ilya Mar 29 '13 at 18:59
Moralists will be moralists... – Git Gud Mar 29 '13 at 19:01
@Ilya Because it's a sphere, and that is a property of a sphere. Wonder at it. Marvel at it. But I certainly can't tell you why. – bobobobo Mar 30 '13 at 0:30
@ilya: Psychology? Math is -all- about psychology. – Mitch Mar 30 '13 at 1:36

Being a cartographer, it becomes second nature to think that you simply can't do that without introducing distortion of any kind, yet it's always a struggle to explain.

B.J.S. Cahill's 1913 rubber-ball globe, might be even better than peeling an orange. Sadly they are currently not made or sold anywhere to my knowledge; desperately hoping for a kickstarter.

I think the most hands on explanation/visualization could be the act of trying to flatten a hemisphere-shaped object like:

You might notice, that you can't change the form of the dome without breaking links‡. With the beanie or elastic cup it's notable that the mesh/rubber changed its form, and any pattern drawn on it has been distorted, after it has been pressed into a flatter state.

Simply put, there is no projection without distortion, since by definition all points on a sphere have the same distance to its center, the closest you get to this in flat space is a circle yet that's about it.

† a nice little hands on project by itself

‡ plus this might explain how Buckminster Fuller came up with his Dymaxion map

-
+1 for the last sentence. – Joe Mar 30 '13 at 0:52

I've got another idea akin to the orange peel experiement mentioned by Ovi. At its heart, it is talking about curvature.

Look up with the student, online or otherwise, what the distance R from the geographic north pole to the equator is. Convince your student that the equater is "a circle centered around the north pole," since all the points on the equator are equidistant from the north pole.

Now look up the length of the equator, and compute for them the circumference of a circle radius R in a plane. Point out that although the radii of the circles are the same, the circumference isn't the same. Argue that the circumferences should be the same, if you are going to flatten without stretching things.

To provide a constrasting example, you can draw a flat circle on a piece of paper, and then go between rolling the paper into a cylinder and flatting it out to show that that circle can be flattened, and that the radius and circumference are indeed unchanged.

-
+1 Sorry about not noticing that you had already brought up comparison with a cylinder. – Jyrki Lahtonen Mar 30 '13 at 7:52
@JyrkiLahtonen Thanks for the courtesy! Really, this sort of thing is bound to happen... – rschwieb Mar 30 '13 at 13:28

How about using a counter example: Take a flat sheet of paper and form it into a cylinder and a cone and demonstrate that you can unroll it to make it flat. Now if you do the same with a ball of some sort, it should be obvious to the child that it cannot be unrolled in the same way. This will help them understand the difference between intrinsic and extrinsic properties. Although the cylinder and cone are curved, their curvature is different. You can then explain that on a cylinder and cone you can always draw a line that stays straight when the paper is rolled up. In contrast, on a ball the lines are always curved.

-

Stand with your right arm pointing forwards and the thumb pointing up.
Swing your arm upwards till it is vertical, then to the right until it is again horizontal and swing it back so it is pointing forwards.

Now your thumb is pointing to the right, but relative to the surface of the sphere on which you have been moving your hand, these moves all kept the orientation of the thumb.

Now consider what would happen if we flattened the sphere without tearing it, and tried to do the same sequence of moves there.

-
I feel like you're getting at something important here, but I can't quite follow it. Why does the orientation of the thumb matter? – Joe Mar 29 '13 at 18:21
@Joe: He's parallel transporting the tangent thumb. – Henry T. Horton Mar 29 '13 at 18:23
@HenryT.Horton, I get that he's parallel transporting the thumb, but why does that matter? – Joe Mar 29 '13 at 18:44
@Joe Because on a sphere you can parallel transport a vector from point $p$ to point $q$ in two different ways, obtaining two different vectors, but you cannot do this on a plane. Hence you cannot flatten a sphere. – A.P. Mar 29 '13 at 21:32
A good idea. I'm just wondering whether it would be easier to follow an imaginary walk from North Pole to the equator along the zero meridian, then one quarter of the equator, and then back to the pole along the meridian 90 degrees East. – Jyrki Lahtonen Mar 31 '13 at 6:11

Which kind of flattening do you mean? If it's about bounded flat surfaces, then imagine a fly captured inside Git's balloon. Would it be possible to flatten the balloon without deforming it, the fly would do this an escape - but it can't. However, it does not help in case you allow the flat surface to be unbounded

-

For a 10 years old :

Take a fix sphere (half of it, a quarter) and a sheet of paper. Then ask the child to perfectly "flatten[1]" the paper on the sphere => it is impossible.

Of course because, the sphere has a curvature and the sheet does not have.so they are not locally isometric

That is actually the theorem (for any portion of the sphere and arbitrary small sheet), because the sheets of paper can only be deformed isometrically unless you start tearing it of. you can make them try the same with a cylinder to see the difference.

[1]: It was the translation I found for "plaquer" hope it is not misleading.

-
I missed the post of Ovi, otherwise I should not have post mine.That is also the best in my opinion. – aximab Mar 30 '13 at 2:41

You can discuss curvature with a child. You do it by taking a flat piece of paper, the previously mentioned orange peel and a mustard leaf. From each you cut an annulus and make a transverse slit in it. You show that the flat piece of paper's annulus can be laid flat with no adjustment. You see that the orange peel's annulus opens up when you lay it flat. You explain that that angle is the curvature. The mustard leaf will be more in interesting. When you lay it flat you will find there is excess annulus. You can then explain negative curvature.

-
This is an excellent idea! It simplifies the notion of Gaussian curvature to an easily-seen measure. – Joe Apr 5 '13 at 2:17
I wish I could say I invented it but this is due to Bill Thurston. I had him for Geometry and the Imagination when I was a graduate student and this is how we did it. – John Douma Apr 5 '13 at 2:34

(This is a stab in the dark, but let me know if you try it!)

I'm tempted to say "Read Flatland by Edwin Abbott", but that will take some extra explaining.

Have them draw an outline of their hand on a piece of paper, cut the drawing out of the paper.

Now, ask them to compare the the width of their hand to the "width" in the drawing. (the axis from their thumb to their pinky). Then have them compare the "height" of their hand to the one in the drawing (wrist to the tip of their middle finger).

Hopefully, they will notice that they are about the same.

Now, what about the "thickness" of their hand? Is it the same as the paper? Probably (hopefully) not.

How would we make it so that the "thickness" was the same?

Well, one way would be to use a rolling pin and squish their hand as flat as the paper.

Spheres work the same way, when you look at them from one direction, they have a width and a height, but when you look at them from another angle, they have a thickness, like a person's hand.

You might consider repeating by outlining the hand clenched in a fist, or get a ball of dough and use a rolling pin. Look at the mess it will make, the dough must go somewhere when you reduce the thickness (a dryer dough that will crack will demonstrate this better).

-
It seems to me that what you are talking about is a ball ( filled-in sphere), whereas what the question is almost certainly about is what in maths is called a sphere, which is only the outside of the ball (hollow inside). – Tara B Mar 29 '13 at 23:41
@TaraB - I'm not a mathematician. But my explanation is also not for a mathematician. This answer could be interpreted as dealing with a solid, but it doesn't have to. The point was to show that when you reduce the number of dimensions there's a need to compensate, and using elements that someone is familiar with, even if it's not a exactly accurate is a useful way to make the problem more tangible. – Andrew Theken Mar 30 '13 at 12:38
But the thing is that a sphere is actually a two-dimensional surface, the same as the plane. So in the actual question being asked, what you are talking about doesn't come up. – Tara B Mar 30 '13 at 15:00

Get a basketball and cut it open (make sure it stays as one piece) and try to flatten it, it will not flatten ... there will always be some part of the basketball that is bulging.

-

Imagine that you have a sphere and a plane (a ball and a floor). Draw an equilateral triangle on both of them (same side length for both).

The area of the triangle on the sphere is greater (this is apparent if the triangle is big enough). If you wanted to flatten out the sphere, you'd need to make the bigger triangle fit inside the smaller one, which you can't do without squashing (projecting).

Alternatively, think about the film of soap when you blow soap bubbles (a plane is deformed). The ring on the outside never grows, but the surface area of the film evidently does. The message is that "bulges" have greater surface area than proper, law-abiding euclidean planes. Again, to flatten out a "bulging" (curved) shape, you'd need to fit a big area into a little one, which you can't do. (In this spirit, of course, a sphere is just a big bulge.)

(these do require him to intuitively grasp that his "flattening out" requires isometry, and that any kind of "squashing" breaks the isometry.)

If he wants to know why the triangle on the sphere (or bulge) is bigger, first make a note about the shortest distance between two points being a straight line. You can go on to observe that all lines on a sphere surface, when viewed in 3D space, are actually not straight, while those on a plane are. So because all lines on a sphere are curved, they are also "longer than they need to be", and since areas are products of lengths, it follows that the areas on a sphere are also larger.

-
+1 for the soap bubbles. – Fred Kline Apr 4 '13 at 19:01

You can, just imagine a perfect sphere, being put into a piece of paper that breaks away as it enters, then take a picture of it. Perfect sphere, the reason you can't physically flatten a sphere, is because all of that third dimension of space has to go SOMEWHERE, so it's either going to fall apart, and split the third dimension perfectly (EXTREMELY unlikely) or, it'll stay together, and become deformed.

A perfect sphere is merely a circle anyway, it's just a filled in surface of the spheres diameter.

This is a complete shot in the dark, and I hope my answer helps, but it's obvious so it probably won't :/

-
Could you explain what you mean by a "perfect sphere being put into a piece of paper that breaks away as it enters"? – Vincent Tjeng Mar 30 '13 at 3:08
+1 for the different point of view. Draw an Antarctic circle sized circle on a flat paper, place the South Pole on the center and push the earth down until both circles match. The paper circle's area must tear because it is less than the corresponding area on the earth. – Fred Kline Apr 4 '13 at 19:11
^Thanks man, Sorry for not clarifying guys, I couldn't find a way to describe it properly. – MarcusJ Apr 5 '13 at 11:17

## protected by mixedmath♦Mar 30 '13 at 20:13

Thank you for your interest in this question. Because it has attracted low-quality answers, posting an answer now requires 10 reputation on this site.